MOSS voice-acting v2 — round-2 SFT evaluation

One prompt per emotion, in English and German · 80 prompts · SFT2 vs SFT2 + DPO2 · 640 generated clips · band intense · LAION

What this measures

Every prompt is rendered in the v2 timed script format: each sentence carries its own [D seconds duration], every silence of 0.2 s or more is an explicit [G seconds pause], and each vocal burst appears in round brackets with its length — (contented sigh, 0.3 seconds). The question is whether the model obeys those numbers, and whether it reaches the requested emotional intensity.

Intensity is asked for as a percentile band on the corpus distribution, not as a raw score. The 40 Empathic-Insight heads are on incomparable scales — Interest has a median of 2.082 and no zeros, Infatuation −0.017 and zeros on 87.7 % of clips — so a raw target would train an Interest detector. The intense band is percentile 0.90–0.98.

Model against model

metricSFT2SFT2 + DPO2change
emotion percentile reached0.4900.508+0.018
quality (blend + genuineness)0.9150.928+0.013
burst hit rate0.6660.650-0.016
word error rate0.1120.118+0.006
duration error, median abs (s)0.1000.080-0.020
within 0.5 s0.9941.000+0.006
overall score0.4880.494+0.006

Same 80 prompts, same source rows, same sampling settings for every model — only the weights differ. "change" is the last model minus the first; green means the change is in the desirable direction for that metric.

The timing is obeyed. The median clip lands 0.10 s from the length its own script adds up to, and 99.7 % land within a second. The mean signed error is -0.011 s, so there is no systematic drift toward longer or shorter.
The emotion is not. These prompts ask for percentile 0.90–0.98 of each emotion and reach 0.490 — the median of the corpus. Asking for intense anger moves the model about as much as not asking. This is the gap the reinforcement-learning stage exists to close, and it is the reason this page exists: it is the baseline that stage has to beat.

English against German

metricende
emotion percentile reached0.4920.489
quality0.9470.883
burst realisation0.3840.366
word error rate0.0800.144
duration error (s, signed)-0.0310.009

German transcribes twice as badly as English. The duration control holds equally in both.

Every prompt

languagesort
Affectiondescripted 8.3 s · 1 burst(s)
SFT2
emotion 0.042duration err 0.20 soverall 0.495 ± 0.029
SFT2 + DPO2
emotion 0.042duration err 0.18 soverall 0.479 ± 0.031
script and prompt
(contented sigh, 0.6 seconds) [0.5 seconds pause] [3.8 seconds duration] Ich versuche immer noch herauszufinden, wie diese Angriffe genau funktionieren. [0.9 seconds pause] [2.5 seconds duration] Das ganze Bild ist ehrlich gesagt noch nicht ganz klar.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Billie
- Instruction:
GENERAL: A voice powerfully expressing confusion, a baffled, disoriented voice, struggling to make sense of things, hesitant and lost, pouring out; reads as intense affection.
SCRIPT:
(contented sigh, 0.6 seconds) [0.5 seconds pause] [3.8 seconds duration] Ich versuche immer noch herauszufinden, wie diese Angriffe genau funktionieren. [0.9 seconds pause] [2.5 seconds duration] Das ganze Bild ist ehrlich gesagt noch nicht ganz klar.
- Tokens:
106
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(contented sigh, 0.6 seconds) [0.5 seconds pause] [3.8 seconds duration] Ich versuche immer noch herauszufinden, wie diese Angriffe genau funktionieren. [0.9 seconds pause] [2.5 seconds duration] Das ganze Bild ist ehrlich gesagt noch nicht ganz klar.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0421.0000.3830.0008.56 s+0.26 s
SFT220.0421.0000.4150.0008.48 s+0.18 s
SFT230.0421.0000.3830.0008.48 s+0.18 s
SFT240.0421.0000.7260.0008.48 s+0.18 s
SFT2 + DPO210.0421.0000.3000.0008.48 s+0.18 s
SFT2 + DPO220.0421.0000.4490.0008.48 s+0.18 s
SFT2 + DPO230.0421.0000.4320.1008.48 s+0.18 s
SFT2 + DPO240.0421.0000.6190.0008.48 s+0.18 s
Affectionenscripted 10.2 s · 1 burst(s)
SFT2
emotion 0.042duration err 0.04 soverall 0.526 ± 0.014
SFT2 + DPO2
emotion 0.042duration err 0.04 soverall 0.526 ± 0.080
script and prompt
[0.6 seconds pause] [4.0 seconds duration] It's too much to the point that I still have both feet on the bar of the stool. [2.7 seconds duration] Like I'm like not foot planted, I'm like, I'm sitting uh, (low mumble, 0.4 seconds) [0.2 seconds pause] [2.0 seconds duration] like perched like a fucking dumb fat gargoyle. [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is energised, brisk, neutral tension, moderately variable; timbre is neutral-toned, slightly bright, very rough, thin; somewhat unclear, some disfluency, wide pitch range, normal breath; affect is elated, slightly dominant, neutral openness; reads as intense affection; style: casual, conversational; below-average recording, some background noise; mildly explicit content; 10.2s, EN.
SCRIPT:
[0.6 seconds pause] [4.0 seconds duration] It's too much to the point that I still have both feet on the bar of the stool. [2.7 seconds duration] Like I'm like not foot planted, I'm like, I'm sitting uh, (low mumble, 0.4 seconds) [0.2 seconds pause] [2.0 seconds duration] like perched like a fucking dumb fat gargoyle. [0.3 seconds pause]
- Tokens:
127
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.6 seconds pause] [4.0 seconds duration] It's too much to the point that I still have both feet on the bar of the stool. [2.7 seconds duration] Like I'm like not foot planted, I'm like, I'm sitting uh, (low mumble, 0.4 seconds) [0.2 seconds pause] [2.0 seconds duration] like perched like a fucking dumb fat gargoyle. [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0421.0000.6410.02710.16 s-0.04 s
SFT220.0421.0000.6410.02710.16 s-0.04 s
SFT230.0421.0000.6980.02710.16 s-0.04 s
SFT240.0421.0000.8190.02710.16 s-0.04 s
SFT2 + DPO210.0421.0000.8520.02710.16 s-0.04 s
SFT2 + DPO220.0421.0000.9610.02710.16 s-0.04 s
SFT2 + DPO230.0421.0000.0000.02710.16 s-0.04 s
SFT2 + DPO240.0421.0001.0000.02710.16 s-0.04 s
Amusementdescripted 17.9 s · 3 burst(s)
SFT2
emotion 0.630duration err 0.06 soverall 0.502 ± 0.034
SFT2 + DPO2
emotion 0.652duration err 0.06 soverall 0.547 ± 0.065
script and prompt
[0.2 seconds duration] ähm, (low mumble, 0.1 seconds) [0.4 seconds pause] [0.8 seconds duration] genau, wie ich schon sagte, (low mumble, 0.1 seconds) [11.8 seconds duration] ähm, spanische Mode, also, man muss dazu sagen, jetzt in dem Film, das ist jetzt eigentlich so die Frühform, also, das fing ja 1550 an, so, und eigentlich kam das dann mit Elisabeth ja erst auf, die Mode, aber in dem Film wird das schon wesentlich früher angesetzt. [3.4 seconds duration] Sie ist da einfach eben aufgrund des Oberteils, wo es ja vorne spitz zuläuft, die Schneppe, [0.4 seconds pause] (low mumble, 0.1 seconds) [0.6 seconds duration] ähm,
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult feminine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense amusement; style: conversational, casual; good recording, quiet background; 17.8s, DE.
SCRIPT:
[0.2 seconds duration] ähm, (low mumble, 0.1 seconds) [0.4 seconds pause] [0.8 seconds duration] genau, wie ich schon sagte, (low mumble, 0.1 seconds) [11.8 seconds duration] ähm, spanische Mode, also, man muss dazu sagen, jetzt in dem Film, das ist jetzt eigentlich so die Frühform, also, das fing ja 1550 an, so, und eigentlich kam das dann mit Elisabeth ja erst auf, die Mode, aber in dem Film wird das schon wesentlich früher angesetzt. [3.4 seconds duration] Sie ist da einfach eben aufgrund des Oberteils, wo es ja vorne spitz zuläuft, die Schneppe, [0.4 seconds pause] (low mumble, 0.1 seconds) [0.6 seconds duration] ähm,
- Tokens:
222
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.2 seconds duration] ähm, (low mumble, 0.1 seconds) [0.4 seconds pause] [0.8 seconds duration] genau, wie ich schon sagte, (low mumble, 0.1 seconds) [11.8 seconds duration] ähm, spanische Mode, also, man muss dazu sagen, jetzt in dem Film, das ist jetzt eigentlich so die Frühform, also, das fing ja 1550 an, so, und eigentlich kam das dann mit Elisabeth ja erst auf, die Mode, aber in dem Film wird das schon wesentlich früher angesetzt. [3.4 seconds duration] Sie ist da einfach eben aufgrund des Oberteils, wo es ja vorne spitz zuläuft, die Schneppe, [0.4 seconds pause] (low mumble, 0.1 seconds) [0.6 seconds duration] ähm,
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5711.0000.5520.05617.84 s-0.06 s
SFT220.6261.0000.5240.05617.84 s-0.06 s
SFT230.6511.0000.6280.08517.84 s-0.06 s
SFT240.6711.0000.6360.00017.84 s-0.06 s
SFT2 + DPO210.6721.0000.6100.05617.84 s-0.06 s
SFT2 + DPO220.6281.0000.5770.08517.84 s-0.06 s
SFT2 + DPO230.5641.0000.6900.01417.84 s-0.06 s
SFT2 + DPO240.7441.0000.6040.02817.84 s-0.06 s
Amusementenscripted 13.3 s · 1 burst(s)
SFT2
emotion 0.027duration err 0.16 soverall 0.416 ± 0.027
SFT2 + DPO2
emotion 0.027duration err 0.14 soverall 0.400 ± 0.000
script and prompt
[0.7 seconds pause] [5.4 seconds duration] If they refuse, the grief we face will be unbearable, a sorrow we surely merit. [0.7 seconds pause] (contented sigh, 0.3 seconds) [0.7 seconds pause] [5.5 seconds duration] Such a fate for all humankind weighs so heavily upon my heart.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Katrin
- Instruction:
GENERAL: A voice at the extreme edge of sad crying, entirely inside the moment, involuntary rather than performed; reads as intense amusement.
SCRIPT:
[0.7 seconds pause] [5.4 seconds duration] If they refuse, the grief we face will be unbearable, a sorrow we surely merit. [0.7 seconds pause] (contented sigh, 0.3 seconds) [0.7 seconds pause] [5.5 seconds duration] Such a fate for all humankind weighs so heavily upon my heart.
- Tokens:
168
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.7 seconds pause] [5.4 seconds duration] If they refuse, the grief we face will be unbearable, a sorrow we surely merit. [0.7 seconds pause] (contented sigh, 0.3 seconds) [0.7 seconds pause] [5.5 seconds duration] Such a fate for all humankind weighs so heavily upon my heart.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0271.0000.0000.00013.44 s+0.14 s
SFT220.0271.0000.0000.00013.44 s+0.14 s
SFT230.0271.0000.0000.00013.44 s+0.14 s
SFT240.0271.0000.3100.00013.52 s+0.22 s
SFT2 + DPO210.0271.0000.0000.00013.44 s+0.14 s
SFT2 + DPO220.0271.0000.0000.00013.44 s+0.14 s
SFT2 + DPO230.0271.0000.0000.00013.44 s+0.14 s
SFT2 + DPO240.0271.0000.0000.00013.44 s+0.14 s
Angerdescripted 11.4 s · 0 burst(s)
SFT2
emotion 0.518duration err 0.12 soverall 0.460 ± 0.136
SFT2 + DPO2
emotion 0.629duration err 0.10 soverall 0.483 ± 0.044
script and prompt
(ahem) [0.3 seconds duration] Hast [1.0 seconds duration] du schon mal Nimik gehört? [2.1 seconds duration] Ich wollte grad sagen, von Nimik hab ich ja noch nie gehört. [0.6 seconds pause] [0.8 seconds duration] Ich auch nicht. [2.5 seconds duration] Naja, dann wird sich das ja nächste Woche erinnern. [1.0 seconds pause] [0.9 seconds duration] Nimik.com. [0.8 seconds pause] [0.5 seconds duration] Unleashing. [0.9 seconds duration] Was ich aber gehört.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is very low-energy, measured, relaxed, fairly steady; timbre is neutral-toned, slightly dark, slightly rough, balanced body; somewhat unclear, frequent disfluency, fairly narrow pitch, normal breath; affect is mildly negative, neutral stance, neutral openness; reads as intense anger; style: casual, conversational; below-average recording, quiet background; 11.6s, DE.
SCRIPT:
(ahem) [0.3 seconds duration] Hast [1.0 seconds duration] du schon mal Nimik gehört? [2.1 seconds duration] Ich wollte grad sagen, von Nimik hab ich ja noch nie gehört. [0.6 seconds pause] [0.8 seconds duration] Ich auch nicht. [2.5 seconds duration] Naja, dann wird sich das ja nächste Woche erinnern. [1.0 seconds pause] [0.9 seconds duration] Nimik.com. [0.8 seconds pause] [0.5 seconds duration] Unleashing. [0.9 seconds duration] Was ich aber gehört.
- Tokens:
144
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(ahem) [0.3 seconds duration] Hast [1.0 seconds duration] du schon mal Nimik gehört? [2.1 seconds duration] Ich wollte grad sagen, von Nimik hab ich ja noch nie gehört. [0.6 seconds pause] [0.8 seconds duration] Ich auch nicht. [2.5 seconds duration] Naja, dann wird sich das ja nächste Woche erinnern. [1.0 seconds pause] [0.9 seconds duration] Nimik.com. [0.8 seconds pause] [0.5 seconds duration] Unleashing. [0.9 seconds duration] Was ich aber gehört.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.4971.0001.0000.10811.52 s+0.12 s
SFT220.6391.0001.0000.08111.52 s+0.12 s
SFT230.4510.6540.0000.16211.52 s+0.12 s
SFT240.4841.0001.0000.16211.52 s+0.12 s
SFT2 + DPO210.6521.0001.0000.13511.52 s+0.12 s
SFT2 + DPO220.5060.6931.0000.13511.52 s+0.12 s
SFT2 + DPO230.7490.5781.0000.27011.36 s-0.04 s
SFT2 + DPO240.6091.0001.0000.16211.52 s+0.12 s
Angerenscripted 15.4 s · 0 burst(s)
SFT2
emotion 0.231duration err 0.12 soverall 0.385 ± 0.006
SFT2 + DPO2
emotion 0.231duration err 0.12 soverall 0.386 ± 0.005
script and prompt
[1.0 seconds duration] You perfect the logo. [0.9 seconds duration] Uhm, (low mumble) [0.6 seconds pause] [2.4 seconds duration] I actually am not very good with math and this is really funny. [2.9 seconds duration] I found the email I sent to myself from a restaurant. [0.3 seconds pause] [2.0 seconds duration] I couldn't figure out the math of that little flip. [0.4 seconds pause] [4.9 seconds duration] So I actually took a napkin and I flipped it and I traced it in Photoshop and
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense anger; style: casual, conversational; average recording, quiet background; 15.4s, EN.
SCRIPT:
[1.0 seconds duration] You perfect the logo. [0.9 seconds duration] Uhm, (low mumble) [0.6 seconds pause] [2.4 seconds duration] I actually am not very good with math and this is really funny. [2.9 seconds duration] I found the email I sent to myself from a restaurant. [0.3 seconds pause] [2.0 seconds duration] I couldn't figure out the math of that little flip. [0.4 seconds pause] [4.9 seconds duration] So I actually took a napkin and I flipped it and I traced it in Photoshop and
- Tokens:
192
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[1.0 seconds duration] You perfect the logo. [0.9 seconds duration] Uhm, (low mumble) [0.6 seconds pause] [2.4 seconds duration] I actually am not very good with math and this is really funny. [2.9 seconds duration] I found the email I sent to myself from a restaurant. [0.3 seconds pause] [2.0 seconds duration] I couldn't figure out the math of that little flip. [0.4 seconds pause] [4.9 seconds duration] So I actually took a napkin and I flipped it and I traced it in Photoshop and
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.2801.0000.0000.05415.28 s-0.12 s
SFT220.0841.0000.0000.01815.28 s-0.12 s
SFT230.2801.0000.0000.05415.28 s-0.12 s
SFT240.2801.0000.0000.03615.28 s-0.12 s
SFT2 + DPO210.0841.0000.0000.03615.28 s-0.12 s
SFT2 + DPO220.2801.0000.0000.05415.28 s-0.12 s
SFT2 + DPO230.2801.0000.0000.01815.28 s-0.12 s
SFT2 + DPO240.2801.0000.0000.03615.28 s-0.12 s
Astonishment Surprisedescripted 5.2 s · 1 burst(s)
SFT2
emotion 0.352duration err 0.00 soverall 0.297 ± 0.095
SFT2 + DPO2
emotion 0.454duration err 0.00 soverall 0.396 ± 0.045
script and prompt
(ahem, 0.1 seconds) [0.1 seconds duration] Ja, [1.9 seconds duration] das ist auch viel zu spät. [3.1 seconds duration] Wir nehmen um kurz vor eins in der Nacht auf, ist es irgendwie nicht.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is positive, slightly submissive, neutral openness; reads as intense astonishment surprise; style: conversational, casual; average recording, some background noise; 5.3s, DE.
SCRIPT:
(ahem, 0.1 seconds) [0.1 seconds duration] Ja, [1.9 seconds duration] das ist auch viel zu spät. [3.1 seconds duration] Wir nehmen um kurz vor eins in der Nacht auf, ist es irgendwie nicht.
- Tokens:
65
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(ahem, 0.1 seconds) [0.1 seconds duration] Ja, [1.9 seconds duration] das ist auch viel zu spät. [3.1 seconds duration] Wir nehmen um kurz vor eins in der Nacht auf, ist es irgendwie nicht.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1420.5030.2080.1435.20 s+0.00 s
SFT220.5941.0000.0000.0485.20 s+0.00 s
SFT230.2560.5500.0000.1435.20 s+0.00 s
SFT240.4161.0000.0000.0005.20 s+0.00 s
SFT2 + DPO210.4161.0000.0000.0005.20 s+0.00 s
SFT2 + DPO220.1421.0000.0000.2385.20 s+0.00 s
SFT2 + DPO230.4161.0000.1920.0485.20 s+0.00 s
SFT2 + DPO240.8420.5000.0000.1435.20 s+0.00 s
Astonishment Surpriseenscripted 7.4 s · 1 burst(s)
SFT2
emotion 0.539duration err 0.04 soverall 0.303 ± 0.207
SFT2 + DPO2
emotion 0.670duration err 0.04 soverall 0.541 ± 0.200
script and prompt
[3.0 seconds duration] That whole fish and chips dream just means I finally feel settled. [0.5 seconds pause] (breathy giggle, 0.4 seconds) [0.3 seconds pause] [3.2 seconds duration] The nagging issue felt so big, but it's completely over now.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Linda
- Instruction:
GENERAL: A voice slightly expressing relief, a voice flooding with relief, exhaling the tension, grateful it is finally over, in every breath. The feeling is present but dialled down, not performed at full size. The feeling is fought down rather than shown: it is held in and only leaks through at the edges of phrases; reads as intense astonishment surprise.
SCRIPT:
[3.0 seconds duration] That whole fish and chips dream just means I finally feel settled. [0.5 seconds pause] (breathy giggle, 0.4 seconds) [0.3 seconds pause] [3.2 seconds duration] The nagging issue felt so big, but it's completely over now.
- Tokens:
92
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[3.0 seconds duration] That whole fish and chips dream just means I finally feel settled. [0.5 seconds pause] (breathy giggle, 0.4 seconds) [0.3 seconds pause] [3.2 seconds duration] The nagging issue felt so big, but it's completely over now.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.4160.5100.0000.0007.36 s-0.04 s
SFT220.6041.0000.7560.0007.36 s-0.04 s
SFT230.5590.0440.0000.0437.36 s-0.04 s
SFT240.5761.0000.2000.0007.36 s-0.04 s
SFT2 + DPO210.8301.0000.2350.0007.36 s-0.04 s
SFT2 + DPO220.7951.0000.2170.0007.36 s-0.04 s
SFT2 + DPO230.5010.5440.0000.0007.36 s-0.04 s
SFT2 + DPO240.5551.0000.6700.0007.36 s-0.04 s
Awedescripted 14.4 s · 0 burst(s)
SFT2
emotion 0.822duration err 0.02 soverall 0.821 ± 0.081
SFT2 + DPO2
emotion 0.822duration err 0.06 soverall 0.854 ± 0.014
script and prompt
(ahem) [0.2 seconds duration] Ich [4.8 seconds duration] schäme mich so, darüber reden zu müssen, aber nach der Reha muss ich wirklich auf mich aufpassen. [0.6 seconds pause] [7.0 seconds duration] Ich meine, auf meine Hygiene achten, auf meine Ernährung schauen und meinen Spezialisten kontaktieren, wenn irgendetwas komisch ist... [0.4 seconds pause] [1.4 seconds duration] das ist alles so privat.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Susannah
- Instruction:
GENERAL: A voice deeply expressing shame, a small, shame-filled voice, cringing and self-reproaching, unable to look up, building and building; reads as intense awe.
SCRIPT:
(ahem) [0.2 seconds duration] Ich [4.8 seconds duration] schäme mich so, darüber reden zu müssen, aber nach der Reha muss ich wirklich auf mich aufpassen. [0.6 seconds pause] [7.0 seconds duration] Ich meine, auf meine Hygiene achten, auf meine Ernährung schauen und meinen Spezialisten kontaktieren, wenn irgendetwas komisch ist... [0.4 seconds pause] [1.4 seconds duration] das ist alles so privat.
- Tokens:
180
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(ahem) [0.2 seconds duration] Ich [4.8 seconds duration] schäme mich so, darüber reden zu müssen, aber nach der Reha muss ich wirklich auf mich aufpassen. [0.6 seconds pause] [7.0 seconds duration] Ich meine, auf meine Hygiene achten, auf meine Ernährung schauen und meinen Spezialisten kontaktieren, wenn irgendetwas komisch ist... [0.4 seconds pause] [1.4 seconds duration] das ist alles so privat.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.8221.0001.0000.00014.40 s+0.00 s
SFT220.8221.0001.0000.00014.48 s+0.08 s
SFT230.8220.5611.0000.02414.40 s+0.00 s
SFT240.8220.9441.0000.00014.40 s+0.00 s
SFT2 + DPO210.8221.0001.0000.00014.40 s+0.00 s
SFT2 + DPO220.8221.0001.0000.02414.48 s+0.08 s
SFT2 + DPO230.8221.0001.0000.02414.48 s+0.08 s
SFT2 + DPO240.8221.0001.0000.04914.48 s+0.08 s
Aweenscripted 20.9 s · 2 burst(s)
SFT2
emotion 0.822duration err 0.10 soverall 0.836 ± 0.026
SFT2 + DPO2
emotion 0.822duration err 0.10 soverall 0.863 ± 0.003
script and prompt
[3.4 seconds duration] We got a nice little break with the Dominic Mysterio storyline, which I really enjoyed. [1.4 seconds duration] We put that over a lot. [0.4 seconds pause] (low mumble, 0.3 seconds) [0.2 seconds pause] [4.6 seconds duration] I actually did like the Backlash match, and I liked a lot of the angles that they were doing in the lead up to that. [0.5 seconds pause] [4.3 seconds duration] I think this was always what was going to be, and I think the fact that we've known, this has been signposted, that we were getting [0.5 seconds pause] (low mumble, 0.3 seconds) [0.5 seconds pause] [4.2 seconds duration] a three match series that would culminate in a big match for SummerSlam has kind of... [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense awe; style: casual, conversational; good recording, quiet background; 20.9s, EN.
SCRIPT:
[3.4 seconds duration] We got a nice little break with the Dominic Mysterio storyline, which I really enjoyed. [1.4 seconds duration] We put that over a lot. [0.4 seconds pause] (low mumble, 0.3 seconds) [0.2 seconds pause] [4.6 seconds duration] I actually did like the Backlash match, and I liked a lot of the angles that they were doing in the lead up to that. [0.5 seconds pause] [4.3 seconds duration] I think this was always what was going to be, and I think the fact that we've known, this has been signposted, that we were getting [0.5 seconds pause] (low mumble, 0.3 seconds) [0.5 seconds pause] [4.2 seconds duration] a three match series that would culminate in a big match for SummerSlam has kind of... [0.3 seconds pause]
- Tokens:
260
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[3.4 seconds duration] We got a nice little break with the Dominic Mysterio storyline, which I really enjoyed. [1.4 seconds duration] We put that over a lot. [0.4 seconds pause] (low mumble, 0.3 seconds) [0.2 seconds pause] [4.6 seconds duration] I actually did like the Backlash match, and I liked a lot of the angles that they were doing in the lead up to that. [0.5 seconds pause] [4.3 seconds duration] I think this was always what was going to be, and I think the fact that we've known, this has been signposted, that we were getting [0.5 seconds pause] (low mumble, 0.3 seconds) [0.5 seconds pause] [4.2 seconds duration] a three match series that would culminate in a big match for SummerSlam has kind of... [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.8221.0000.8870.01120.80 s-0.10 s
SFT220.8221.0000.9230.00020.80 s-0.10 s
SFT230.8221.0000.8900.00020.80 s-0.10 s
SFT240.8221.0000.7240.03420.80 s-0.10 s
SFT2 + DPO210.8221.0000.9240.00020.80 s-0.10 s
SFT2 + DPO220.8221.0000.9420.00020.80 s-0.10 s
SFT2 + DPO230.8221.0000.9420.00020.80 s-0.10 s
SFT2 + DPO240.8221.0000.9620.00020.80 s-0.10 s
Bitternessdescripted 15.4 s · 2 burst(s)
SFT2
emotion 0.291duration err 0.04 soverall 0.559 ± 0.017
SFT2 + DPO2
emotion 0.546duration err 0.04 soverall 0.554 ± 0.016
script and prompt
(low mumble, 0.4 seconds) [8.7 seconds duration] ich hab dich jetzt so bei den eiern du musst jetzt auch ein bisschen mitspielen und wenn du gut mitspielst dann verspreche ich dir auch dann bist du in 18 monaten hier irgendwie detective und dann (low mumble, 0.2 seconds) [6.1 seconds duration] ist das ganze thema auch komplett durch für dich so und das ist glaube ich so eine gewisse ernüchterung auch bei jake dass er sagt okay
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense bitterness; style: casual, conversational; good recording, quiet background; 15.4s, DE.
SCRIPT:
(low mumble, 0.4 seconds) [8.7 seconds duration] ich hab dich jetzt so bei den eiern du musst jetzt auch ein bisschen mitspielen und wenn du gut mitspielst dann verspreche ich dir auch dann bist du in 18 monaten hier irgendwie detective und dann (low mumble, 0.2 seconds) [6.1 seconds duration] ist das ganze thema auch komplett durch für dich so und das ist glaube ich so eine gewisse ernüchterung auch bei jake dass er sagt okay
- Tokens:
192
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(low mumble, 0.4 seconds) [8.7 seconds duration] ich hab dich jetzt so bei den eiern du musst jetzt auch ein bisschen mitspielen und wenn du gut mitspielst dann verspreche ich dir auch dann bist du in 18 monaten hier irgendwie detective und dann (low mumble, 0.2 seconds) [6.1 seconds duration] ist das ganze thema auch komplett durch für dich so und das ist glaube ich so eine gewisse ernüchterung auch bei jake dass er sagt okay
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0181.0000.8550.03215.36 s-0.04 s
SFT220.4641.0000.9430.01615.36 s-0.04 s
SFT230.4081.0000.7100.01615.36 s-0.04 s
SFT240.2731.0000.8430.00015.36 s-0.04 s
SFT2 + DPO210.6761.0000.7920.03215.36 s-0.04 s
SFT2 + DPO220.5411.0000.7970.01615.36 s-0.04 s
SFT2 + DPO230.4641.0000.7410.01615.36 s-0.04 s
SFT2 + DPO240.5021.0000.7710.01615.36 s-0.04 s
Bitternessenscripted 10.2 s · 1 burst(s)
SFT2
emotion 0.835duration err 0.12 soverall 0.780 ± 0.130
SFT2 + DPO2
emotion 0.833duration err 0.12 soverall 0.799 ± 0.072
script and prompt
[2.1 seconds duration] When they brought out that Midas touch shit, (ahem, 0.2 seconds) [0.3 seconds pause] [1.0 seconds duration] Peterson got hit with that. [0.5 seconds pause] [6.1 seconds duration] He got hit with like a, like a mechanized media machine that's beholden to the democratic party.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is mildly negative, slightly dominant, neutral openness; reads as intense bitterness; style: casual, conversational; average recording, quiet background; mildly explicit content; 10.1s, EN.
SCRIPT:
[2.1 seconds duration] When they brought out that Midas touch shit, (ahem, 0.2 seconds) [0.3 seconds pause] [1.0 seconds duration] Peterson got hit with that. [0.5 seconds pause] [6.1 seconds duration] He got hit with like a, like a mechanized media machine that's beholden to the democratic party.
- Tokens:
126
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[2.1 seconds duration] When they brought out that Midas touch shit, (ahem, 0.2 seconds) [0.3 seconds pause] [1.0 seconds duration] Peterson got hit with that. [0.5 seconds pause] [6.1 seconds duration] He got hit with like a, like a mechanized media machine that's beholden to the democratic party.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7431.0000.2750.06710.08 s-0.12 s
SFT220.8421.0000.8870.00010.08 s-0.12 s
SFT230.8811.0000.3230.06710.08 s-0.12 s
SFT240.8731.0000.6440.00010.08 s-0.12 s
SFT2 + DPO210.8861.0000.3500.00010.08 s-0.12 s
SFT2 + DPO220.7861.0000.9230.06710.08 s-0.12 s
SFT2 + DPO230.9111.0000.8870.10010.08 s-0.12 s
SFT2 + DPO240.7511.0000.8520.03310.08 s-0.12 s
Concentrationdescripted 17.0 s · 0 burst(s)
SFT2
emotion 0.605duration err 0.12 soverall 0.356 ± 0.012
SFT2 + DPO2
emotion 0.577duration err 0.12 soverall 0.352 ± 0.010
script and prompt
[1.4 seconds duration] Ich komm mit dem Ding gar nix anfangen. [0.3 seconds pause] [0.6 seconds duration] Also wirklich nix. [0.2 seconds pause] [2.0 seconds duration] Also, äh, (low mumble) [4.6 seconds duration] ich bin jetzt nicht der größte Rick and Morty Fan, von daher bin ich jetzt nicht irgendwie sauer, was daraus gemacht wurde. [1.4 seconds pause] [4.1 seconds duration] Ich hab einfach im Ohr sehr viele Fragezeichen im Kopf und weiß nicht, warum das existiert. [0.8 seconds duration] Ich geb ne Eins von Zehn. [0.8 seconds duration] Das war für mich nichts. [0.6 seconds pause] [0.2 seconds duration] Wird schon.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense concentration; style: casual, monologue; average recording, quiet background; 17.2s, DE.
SCRIPT:
[1.4 seconds duration] Ich komm mit dem Ding gar nix anfangen. [0.3 seconds pause] [0.6 seconds duration] Also wirklich nix. [0.2 seconds pause] [2.0 seconds duration] Also, äh, (low mumble) [4.6 seconds duration] ich bin jetzt nicht der größte Rick and Morty Fan, von daher bin ich jetzt nicht irgendwie sauer, was daraus gemacht wurde. [1.4 seconds pause] [4.1 seconds duration] Ich hab einfach im Ohr sehr viele Fragezeichen im Kopf und weiß nicht, warum das existiert. [0.8 seconds duration] Ich geb ne Eins von Zehn. [0.8 seconds duration] Das war für mich nichts. [0.6 seconds pause] [0.2 seconds duration] Wird schon.
- Tokens:
214
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[1.4 seconds duration] Ich komm mit dem Ding gar nix anfangen. [0.3 seconds pause] [0.6 seconds duration] Also wirklich nix. [0.2 seconds pause] [2.0 seconds duration] Also, äh, (low mumble) [4.6 seconds duration] ich bin jetzt nicht der größte Rick and Morty Fan, von daher bin ich jetzt nicht irgendwie sauer, was daraus gemacht wurde. [1.4 seconds pause] [4.1 seconds duration] Ich hab einfach im Ohr sehr viele Fragezeichen im Kopf und weiß nicht, warum das existiert. [0.8 seconds duration] Ich geb ne Eins von Zehn. [0.8 seconds duration] Das war für mich nichts. [0.6 seconds pause] [0.2 seconds duration] Wird schon.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6271.0000.0000.14117.12 s+0.12 s
SFT220.6201.0000.0000.15617.12 s+0.12 s
SFT230.6021.0000.0000.06217.12 s+0.12 s
SFT240.5701.0000.0000.14117.12 s+0.12 s
SFT2 + DPO210.5591.0000.0000.18817.12 s+0.12 s
SFT2 + DPO220.6581.0000.0000.14117.12 s+0.12 s
SFT2 + DPO230.6601.0000.0000.15617.12 s+0.12 s
SFT2 + DPO240.4321.0000.0000.14117.12 s+0.12 s
Concentrationenscripted 13.7 s · 2 burst(s)
SFT2
emotion 0.569duration err 0.10 soverall 0.460 ± 0.009
SFT2 + DPO2
emotion 0.594duration err 0.10 soverall 0.464 ± 0.025
script and prompt
[0.2 seconds pause] (low mumble, 0.1 seconds) [0.4 seconds duration] Uhm, [0.4 seconds pause] [2.5 seconds duration] you know, we both had Red and Blue and, you know, played a ton of it. [0.4 seconds pause] [8.1 seconds duration] And, you know, like, if those titles, you know, Final Fantasy VII, you know, if those titles, those specific titles didn't happen, you know, where would we be now? [1.3 seconds duration] It's an interesting idea, but, (ahem, 0.1 seconds) [0.2 seconds duration] uh...
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly negative, neutral stance, slightly guarded; reads as intense concentration; style: casual, conversational; average recording, quiet background; 13.6s, EN.
SCRIPT:
[0.2 seconds pause] (low mumble, 0.1 seconds) [0.4 seconds duration] Uhm, [0.4 seconds pause] [2.5 seconds duration] you know, we both had Red and Blue and, you know, played a ton of it. [0.4 seconds pause] [8.1 seconds duration] And, you know, like, if those titles, you know, Final Fantasy VII, you know, if those titles, those specific titles didn't happen, you know, where would we be now? [1.3 seconds duration] It's an interesting idea, but, (ahem, 0.1 seconds) [0.2 seconds duration] uh...
- Tokens:
170
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.2 seconds pause] (low mumble, 0.1 seconds) [0.4 seconds duration] Uhm, [0.4 seconds pause] [2.5 seconds duration] you know, we both had Red and Blue and, you know, played a ton of it. [0.4 seconds pause] [8.1 seconds duration] And, you know, like, if those titles, you know, Final Fantasy VII, you know, if those titles, those specific titles didn't happen, you know, where would we be now? [1.3 seconds duration] It's an interesting idea, but, (ahem, 0.1 seconds) [0.2 seconds duration] uh...
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6361.0000.4180.03813.60 s-0.10 s
SFT220.6351.0000.3920.05813.60 s-0.10 s
SFT230.5101.0000.4540.03813.60 s-0.10 s
SFT240.4941.0000.3870.05813.60 s-0.10 s
SFT2 + DPO210.6301.0000.6190.03813.60 s-0.10 s
SFT2 + DPO220.5461.0000.3700.05813.60 s-0.10 s
SFT2 + DPO230.6271.0000.4470.05813.60 s-0.10 s
SFT2 + DPO240.5751.0000.3210.05813.60 s-0.10 s
Confusiondescripted 9.0 s · 1 burst(s)
SFT2
emotion 0.214duration err 0.28 soverall 0.536 ± 0.147
SFT2 + DPO2
emotion 0.474duration err 0.28 soverall 0.439 ± 0.067
script and prompt
(contented sigh, 0.2 seconds) [0.3 seconds pause] [3.3 seconds duration] Mein Gott, dieses Frühstück ist ja unglaublich reichhaltig. [0.9 seconds pause] [3.9 seconds duration] Echtes englisches Essen, weißt du, einfach randvoll mit allem, was es sein sollte. [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Vali
- Instruction:
GENERAL: A voice very high in mask resonance. Mask resonance is maximally developed and overwhelmingly dominant, completely saturating the voice with brilliant, forward mask cavity resonance. The acoustic output is almost entirely derived from facial mask vibration, creating an impossibly bright, forward-projected vocal quality; reads as intense confusion.
SCRIPT:
(contented sigh, 0.2 seconds) [0.3 seconds pause] [3.3 seconds duration] Mein Gott, dieses Frühstück ist ja unglaublich reichhaltig. [0.9 seconds pause] [3.9 seconds duration] Echtes englisches Essen, weißt du, einfach randvoll mit allem, was es sein sollte. [0.4 seconds pause]
- Tokens:
116
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(contented sigh, 0.2 seconds) [0.3 seconds pause] [3.3 seconds duration] Mein Gott, dieses Frühstück ist ja unglaublich reichhaltig. [0.9 seconds pause] [3.9 seconds duration] Echtes englisches Essen, weißt du, einfach randvoll mit allem, was es sein sollte. [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7581.0000.8870.0009.28 s+0.28 s
SFT220.0461.0000.0000.0009.28 s+0.28 s
SFT230.0461.0000.9610.0009.28 s+0.28 s
SFT240.0081.0000.0000.0009.28 s+0.28 s
SFT2 + DPO210.5721.0000.0000.0009.28 s+0.28 s
SFT2 + DPO220.0461.0000.0000.0009.28 s+0.28 s
SFT2 + DPO230.5791.0000.0000.0009.28 s+0.28 s
SFT2 + DPO240.6991.0000.6410.0959.28 s+0.28 s
Confusionenscripted 13.8 s · 2 burst(s)
SFT2
emotion 0.211duration err 0.18 soverall 0.510 ± 0.043
SFT2 + DPO2
emotion 0.178duration err 0.16 soverall 0.490 ± 0.025
script and prompt
[2.2 seconds duration] You know, commercial systems are coming out that can do hundreds. [0.4 seconds pause] [4.1 seconds duration] And so this is like a major area of improvement that will be very enabling for content mix. [0.4 seconds pause] (ahem, 0.1 seconds) [1.7 seconds duration] And the third component is barcoding, right? [1.3 seconds duration] So I started out by saying that (low mumble, 0.3 seconds) [1.0 seconds pause] [2.3 seconds duration] interactions in the brain are extremely highly non-local.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, brisk, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, slightly dominant, slightly guarded; reads as intense confusion; style: casual, conversational; average recording, quiet background; 14.0s, EN.
SCRIPT:
[2.2 seconds duration] You know, commercial systems are coming out that can do hundreds. [0.4 seconds pause] [4.1 seconds duration] And so this is like a major area of improvement that will be very enabling for content mix. [0.4 seconds pause] (ahem, 0.1 seconds) [1.7 seconds duration] And the third component is barcoding, right? [1.3 seconds duration] So I started out by saying that (low mumble, 0.3 seconds) [1.0 seconds pause] [2.3 seconds duration] interactions in the brain are extremely highly non-local.
- Tokens:
175
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[2.2 seconds duration] You know, commercial systems are coming out that can do hundreds. [0.4 seconds pause] [4.1 seconds duration] And so this is like a major area of improvement that will be very enabling for content mix. [0.4 seconds pause] (ahem, 0.1 seconds) [1.7 seconds duration] And the third component is barcoding, right? [1.3 seconds duration] So I started out by saying that (low mumble, 0.3 seconds) [1.0 seconds pause] [2.3 seconds duration] interactions in the brain are extremely highly non-local.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1641.0000.6670.07714.00 s+0.20 s
SFT220.4691.0000.2990.00014.00 s+0.20 s
SFT230.0461.0000.8900.00014.00 s+0.20 s
SFT240.1641.0000.5910.01913.92 s+0.12 s
SFT2 + DPO210.1641.0000.6240.00014.00 s+0.20 s
SFT2 + DPO220.1641.0000.2660.00014.00 s+0.20 s
SFT2 + DPO230.0461.0000.5690.03813.92 s+0.12 s
SFT2 + DPO240.3371.0000.5680.05813.92 s+0.12 s
Contemplationdescripted 17.7 s · 2 burst(s)
SFT2
emotion 0.429duration err 0.23 soverall 0.311 ± 0.073
SFT2 + DPO2
emotion 0.493duration err 0.08 soverall 0.317 ± 0.072
script and prompt
[0.7 seconds duration] Und, (low mumble, 0.1 seconds) [4.4 seconds duration] äh, da ist verloren gegangen, Namen, E-Mail, Adressen, Login-IDs, äh, [0.6 seconds pause] (low mumble, 0.3 seconds) [4.1 seconds duration] auch gehashte Passwörter, aber hier weiß man, dass es Char1 mit Salz ist. [0.6 seconds duration] ouch. [1.1 seconds pause] [4.9 seconds duration] Ja, ich glaube, Char1 gesalzen ist auch nicht so viel besser als Char1 ohne was. [0.3 seconds pause] [0.6 seconds duration] Ist nix wert.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is very low-energy, measured, relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; somewhat unclear, frequent disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense contemplation; style: monologue, casual; average recording, quiet background; 17.9s, DE.
SCRIPT:
[0.7 seconds duration] Und, (low mumble, 0.1 seconds) [4.4 seconds duration] äh, da ist verloren gegangen, Namen, E-Mail, Adressen, Login-IDs, äh, [0.6 seconds pause] (low mumble, 0.3 seconds) [4.1 seconds duration] auch gehashte Passwörter, aber hier weiß man, dass es Char1 mit Salz ist. [0.6 seconds duration] ouch. [1.1 seconds pause] [4.9 seconds duration] Ja, ich glaube, Char1 gesalzen ist auch nicht so viel besser als Char1 ohne was. [0.3 seconds pause] [0.6 seconds duration] Ist nix wert.
- Tokens:
223
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.7 seconds duration] Und, (low mumble, 0.1 seconds) [4.4 seconds duration] äh, da ist verloren gegangen, Namen, E-Mail, Adressen, Login-IDs, äh, [0.6 seconds pause] (low mumble, 0.3 seconds) [4.1 seconds duration] auch gehashte Passwörter, aber hier weiß man, dass es Char1 mit Salz ist. [0.6 seconds duration] ouch. [1.1 seconds pause] [4.9 seconds duration] Ja, ich glaube, Char1 gesalzen ist auch nicht so viel besser als Char1 ohne was. [0.3 seconds pause] [0.6 seconds duration] Ist nix wert.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.4070.5230.7120.24417.76 s+0.06 s
SFT220.4840.5000.5030.22217.76 s+0.06 s
SFT230.3861.0000.5660.17816.96 s-0.74 s
SFT240.4370.5120.6640.20017.76 s+0.06 s
SFT2 + DPO210.3670.5000.4480.24417.76 s+0.06 s
SFT2 + DPO220.6830.5630.7250.17817.76 s+0.06 s
SFT2 + DPO230.5171.0000.6020.26717.76 s+0.06 s
SFT2 + DPO240.4050.5090.5660.20017.84 s+0.14 s
Contemplationenscripted 14.2 s · 2 burst(s)
SFT2
emotion 0.439duration err 0.04 soverall 0.413 ± 0.026
SFT2 + DPO2
emotion 0.399duration err 0.04 soverall 0.445 ± 0.031
script and prompt
[2.7 seconds duration] We missed it the first time to record it, so it happened again. [0.2 seconds pause] [1.4 seconds duration] He was able to put up two again. [0.6 seconds pause] [1.7 seconds duration] So we tried to attack both at one time. [0.2 seconds pause] (ahem, 0.2 seconds) [0.5 seconds pause] [1.4 seconds duration] So here I am with Little Easy. [0.6 seconds pause] [1.8 seconds duration] So this is gonna be, this is not gonna be pretty. [0.4 seconds pause] (exhausted groan, 0.4 seconds) [0.6 seconds pause] [1.2 seconds duration] Little Easy's in no, no... [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, very dark, slightly rough, balanced body; somewhat unclear, frequent disfluency, fairly narrow pitch, normal breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense contemplation; style: casual, monologue; below-average recording, quiet background; 14.2s, EN.
SCRIPT:
[2.7 seconds duration] We missed it the first time to record it, so it happened again. [0.2 seconds pause] [1.4 seconds duration] He was able to put up two again. [0.6 seconds pause] [1.7 seconds duration] So we tried to attack both at one time. [0.2 seconds pause] (ahem, 0.2 seconds) [0.5 seconds pause] [1.4 seconds duration] So here I am with Little Easy. [0.6 seconds pause] [1.8 seconds duration] So this is gonna be, this is not gonna be pretty. [0.4 seconds pause] (exhausted groan, 0.4 seconds) [0.6 seconds pause] [1.2 seconds duration] Little Easy's in no, no... [0.3 seconds pause]
- Tokens:
177
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[2.7 seconds duration] We missed it the first time to record it, so it happened again. [0.2 seconds pause] [1.4 seconds duration] He was able to put up two again. [0.6 seconds pause] [1.7 seconds duration] So we tried to attack both at one time. [0.2 seconds pause] (ahem, 0.2 seconds) [0.5 seconds pause] [1.4 seconds duration] So here I am with Little Easy. [0.6 seconds pause] [1.8 seconds duration] So this is gonna be, this is not gonna be pretty. [0.4 seconds pause] (exhausted groan, 0.4 seconds) [0.6 seconds pause] [1.2 seconds duration] Little Easy's in no, no... [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5371.0000.5930.13214.16 s-0.04 s
SFT220.4641.0000.3300.18914.16 s-0.04 s
SFT230.3711.0000.3300.17014.16 s-0.04 s
SFT240.3861.0000.2850.13214.16 s-0.04 s
SFT2 + DPO210.4501.0000.5980.11314.16 s-0.04 s
SFT2 + DPO220.3171.0000.3170.17014.24 s+0.04 s
SFT2 + DPO230.3961.0000.6430.11314.16 s-0.04 s
SFT2 + DPO240.4351.0000.5840.17014.16 s-0.04 s
Contemptdescripted 13.2 s · 2 burst(s)
SFT2
emotion 0.875duration err 0.10 soverall 0.655 ± 0.180
SFT2 + DPO2
emotion 0.948duration err 0.14 soverall 0.783 ± 0.060
script and prompt
[10.3 seconds duration] Alles Wichtige aus der Linguistik findest du im Lexikonabschnitt, von A für Abaza, eine westkaukasische Sprache, bis Z für das Zyklusprinzip in einer bestimmten Syntax. [1.4 seconds pause] (contented sigh, 0.2 seconds) (exhausted groan, 0.9 seconds) [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Fiona
- Instruction:
GENERAL: A voice extremely low in whisper-talk style. The delivery is intensely loud and completely non-whispered, entirely rejecting any whispered or breath-based element. The acoustic profile shows zero whisper characteristics or breathy quality, sounding harsh and utterly non-whispered; reads as intense contempt.
SCRIPT:
[10.3 seconds duration] Alles Wichtige aus der Linguistik findest du im Lexikonabschnitt, von A für Abaza, eine westkaukasische Sprache, bis Z für das Zyklusprinzip in einer bestimmten Syntax. [1.4 seconds pause] (contented sigh, 0.2 seconds) (exhausted groan, 0.9 seconds) [0.4 seconds pause]
- Tokens:
166
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[10.3 seconds duration] Alles Wichtige aus der Linguistik findest du im Lexikonabschnitt, von A für Abaza, eine westkaukasische Sprache, bis Z für das Zyklusprinzip in einer bestimmten Syntax. [1.4 seconds pause] (contented sigh, 0.2 seconds) (exhausted groan, 0.9 seconds) [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.8950.9360.0000.04013.28 s+0.08 s
SFT220.9591.0000.0000.04013.28 s+0.08 s
SFT230.7190.6170.0000.04013.28 s+0.08 s
SFT240.9261.0000.0000.04013.36 s+0.16 s
SFT2 + DPO210.9161.0000.0000.04013.36 s+0.16 s
SFT2 + DPO220.9431.0000.0000.12013.28 s+0.08 s
SFT2 + DPO230.9761.0000.0000.04013.36 s+0.16 s
SFT2 + DPO240.9581.0000.5680.04013.36 s+0.16 s
Contemptenscripted 18.9 s · 3 burst(s)
SFT2
emotion 0.903duration err 0.18 soverall 0.822 ± 0.115
SFT2 + DPO2
emotion 0.941duration err 0.18 soverall 0.896 ± 0.028
script and prompt
[0.5 seconds duration] So, uh, [0.2 seconds pause] (low mumble, 0.4 seconds) [0.3 seconds pause] [0.3 seconds duration] wow. [0.2 seconds pause] [4.1 seconds duration] And then that guy at the end who's basically burned alive, it's just unreal. [0.5 seconds pause] [1.8 seconds duration] I mean, the enemy must have known that, (low mumble, 0.3 seconds) [0.3 seconds pause] [0.7 seconds duration] uh, must have been, (low mumble, 0.2 seconds) [6.1 seconds duration] uh, given their head a shake when they realized that they were piling up their own dead to climb over to the fort. [0.4 seconds pause] [2.3 seconds duration] It's like, how many of your own have to be- [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is energised, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is mildly negative, slightly dominant, neutral openness; reads as intense contempt; style: casual, conversational; average recording, quiet background; 18.8s, EN.
SCRIPT:
[0.5 seconds duration] So, uh, [0.2 seconds pause] (low mumble, 0.4 seconds) [0.3 seconds pause] [0.3 seconds duration] wow. [0.2 seconds pause] [4.1 seconds duration] And then that guy at the end who's basically burned alive, it's just unreal. [0.5 seconds pause] [1.8 seconds duration] I mean, the enemy must have known that, (low mumble, 0.3 seconds) [0.3 seconds pause] [0.7 seconds duration] uh, must have been, (low mumble, 0.2 seconds) [6.1 seconds duration] uh, given their head a shake when they realized that they were piling up their own dead to climb over to the fort. [0.4 seconds pause] [2.3 seconds duration] It's like, how many of your own have to be- [0.3 seconds pause]
- Tokens:
234
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.5 seconds duration] So, uh, [0.2 seconds pause] (low mumble, 0.4 seconds) [0.3 seconds pause] [0.3 seconds duration] wow. [0.2 seconds pause] [4.1 seconds duration] And then that guy at the end who's basically burned alive, it's just unreal. [0.5 seconds pause] [1.8 seconds duration] I mean, the enemy must have known that, (low mumble, 0.3 seconds) [0.3 seconds pause] [0.7 seconds duration] uh, must have been, (low mumble, 0.2 seconds) [6.1 seconds duration] uh, given their head a shake when they realized that they were piling up their own dead to climb over to the fort. [0.4 seconds pause] [2.3 seconds duration] It's like, how many of your own have to be- [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.9621.0000.8310.04818.72 s-0.18 s
SFT220.9031.0000.7150.04818.72 s-0.18 s
SFT230.9811.0000.3890.04818.72 s-0.18 s
SFT240.7661.0000.4250.06518.72 s-0.18 s
SFT2 + DPO210.9061.0000.7740.04818.72 s-0.18 s
SFT2 + DPO220.9731.0000.8960.04818.72 s-0.18 s
SFT2 + DPO230.9091.0000.5700.04818.72 s-0.18 s
SFT2 + DPO240.9751.0000.5400.04818.72 s-0.18 s
Contentmentdescripted 12.9 s · 1 burst(s)
SFT2
emotion 0.245duration err 0.10 soverall 0.565 ± 0.031
SFT2 + DPO2
emotion 0.256duration err 0.10 soverall 0.400 ± 0.000
script and prompt
[6.3 seconds duration] Ich wünschte mir einfach, jemand wäre da, der mir sagen würde, wohin wir gehen, wo wir angefangen haben und warum das alles wichtig ist. [0.7 seconds pause] (contented sigh, 0.2 seconds) [0.5 seconds pause] [3.6 seconds duration] Es fühlt sich so einsam an, ohne diese beständige Stimme, die mich durch alles führt. [1.6 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Tallulah
- Instruction:
GENERAL: A voice notably high in asmr style. The delivery is distinctly ASMR-oriented and intimate, featuring clear whispered delivery, sensory focus, and a clearly soothing, relaxing register. The speaker sounds engaged with ASMR aesthetics, naturally soothing; reads as intense contentment.
SCRIPT:
[6.3 seconds duration] Ich wünschte mir einfach, jemand wäre da, der mir sagen würde, wohin wir gehen, wo wir angefangen haben und warum das alles wichtig ist. [0.7 seconds pause] (contented sigh, 0.2 seconds) [0.5 seconds pause] [3.6 seconds duration] Es fühlt sich so einsam an, ohne diese beständige Stimme, die mich durch alles führt. [1.6 seconds pause]
- Tokens:
160
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[6.3 seconds duration] Ich wünschte mir einfach, jemand wäre da, der mir sagen würde, wohin wir gehen, wo wir angefangen haben und warum das alles wichtig ist. [0.7 seconds pause] (contented sigh, 0.2 seconds) [0.5 seconds pause] [3.6 seconds duration] Es fühlt sich so einsam an, ohne diese beständige Stimme, die mich durch alles führt. [1.6 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1811.0000.5680.00012.80 s-0.10 s
SFT220.1811.0000.9610.00012.80 s-0.10 s
SFT230.4351.0000.9230.00012.80 s-0.10 s
SFT240.1811.0000.8520.00012.80 s-0.10 s
SFT2 + DPO210.1811.0000.0000.00012.80 s-0.10 s
SFT2 + DPO220.1811.0000.0000.00012.80 s-0.10 s
SFT2 + DPO230.1811.0000.0000.00012.80 s-0.10 s
SFT2 + DPO240.4781.0000.0000.00012.80 s-0.10 s
Contentmentenscripted 8.2 s · 1 burst(s)
SFT2
emotion 0.639duration err 0.20 soverall 0.482 ± 0.173
SFT2 + DPO2
emotion 0.776duration err 0.20 soverall 0.558 ± 0.138
script and prompt
(contented sigh, 0.2 seconds) [0.4 seconds pause] [4.1 seconds duration] But there are places, you know, where the planet feels like it's actually whispering to us, right here, right now. [0.2 seconds pause] [3.3 seconds duration] It's a connection that bypasses all the noise, a raw, undeniable conversation.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Aaron
- Instruction:
GENERAL: A voice very high in stance. The speaker completely dominates and monopolizes the acoustic space, delivering a relentless, aggressive lecture that talks entirely at the listener. There is absolutely zero room left for a response, effectively crushing a dialogue into a forcefully imposed wall of sound; reads as intense contentment.
SCRIPT:
(contented sigh, 0.2 seconds) [0.4 seconds pause] [4.1 seconds duration] But there are places, you know, where the planet feels like it's actually whispering to us, right here, right now. [0.2 seconds pause] [3.3 seconds duration] It's a connection that bypasses all the noise, a raw, undeniable conversation.
- Tokens:
105
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
(contented sigh, 0.2 seconds) [0.4 seconds pause] [4.1 seconds duration] But there are places, you know, where the planet feels like it's actually whispering to us, right here, right now. [0.2 seconds pause] [3.3 seconds duration] It's a connection that bypasses all the noise, a raw, undeniable conversation.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7521.0000.8520.0318.40 s+0.20 s
SFT220.1810.6650.1450.0008.40 s+0.20 s
SFT230.8920.5000.0000.0008.40 s+0.20 s
SFT240.7300.1400.9610.1258.40 s+0.20 s
SFT2 + DPO210.6511.0000.0000.0008.40 s+0.20 s
SFT2 + DPO220.7301.0000.0000.0008.40 s+0.20 s
SFT2 + DPO230.8230.6130.0000.0008.40 s+0.20 s
SFT2 + DPO240.9010.5880.7260.0008.40 s+0.20 s
Disappointmentdescripted 8.6 s · 1 burst(s)
SFT2
emotion 0.620duration err 0.08 soverall 0.511 ± 0.059
SFT2 + DPO2
emotion 0.567duration err 0.08 soverall 0.409 ± 0.100
script and prompt
(low mumble, 0.2 seconds) [6.7 seconds duration] Das klingt für mich auch so, dass der Film für dich ein gutes Gleichgewicht hat zwischen diesen, also was wir ja mal mit diesem Pendel aufgemacht haben. [1.4 seconds duration] Also es ist halt, also. [0.3 seconds duration] Ja.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense disappointment; style: casual, conversational; good recording, quiet background; 8.7s, DE.
SCRIPT:
(low mumble, 0.2 seconds) [6.7 seconds duration] Das klingt für mich auch so, dass der Film für dich ein gutes Gleichgewicht hat zwischen diesen, also was wir ja mal mit diesem Pendel aufgemacht haben. [1.4 seconds duration] Also es ist halt, also. [0.3 seconds duration] Ja.
- Tokens:
108
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(low mumble, 0.2 seconds) [6.7 seconds duration] Das klingt für mich auch so, dass der Film für dich ein gutes Gleichgewicht hat zwischen diesen, also was wir ja mal mit diesem Pendel aufgemacht haben. [1.4 seconds duration] Also es ist halt, also. [0.3 seconds duration] Ja.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.4821.0000.7870.0618.72 s+0.12 s
SFT220.7561.0000.3100.0918.64 s+0.04 s
SFT230.6241.0000.3360.1218.72 s+0.12 s
SFT240.6201.0000.8520.0918.64 s+0.04 s
SFT2 + DPO210.4701.0000.5680.0918.72 s+0.12 s
SFT2 + DPO220.7200.5190.8520.0308.64 s+0.04 s
SFT2 + DPO230.6110.8060.7260.0308.64 s+0.04 s
SFT2 + DPO240.4700.5000.3230.1218.72 s+0.12 s
Disappointmentenscripted 8.7 s · 1 burst(s)
SFT2
emotion 0.990duration err 0.10 soverall 0.811 ± 0.036
SFT2 + DPO2
emotion 0.996duration err 0.10 soverall 0.779 ± 0.072
script and prompt
[0.3 seconds pause] (ahem, 0.2 seconds) [0.0 seconds duration] I [4.3 seconds duration] must have been stained by those claws; I feel so deeply ashamed of myself. [0.4 seconds pause] [3.3 seconds duration] My body feels like it carries a heavy, undeniable disgrace. [0.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Jules
- Instruction:
GENERAL: A voice slightly expressing shame, a small, shame-filled voice, cringing and self-reproaching, unable to look up, building and building. The feeling is present but dialled down, not performed at full size. The feeling is fought down rather than shown: it is held in and only leaks through at the edges of phrases; reads as intense disappointment.
SCRIPT:
[0.3 seconds pause] (ahem, 0.2 seconds) [0.0 seconds duration] I [4.3 seconds duration] must have been stained by those claws; I feel so deeply ashamed of myself. [0.4 seconds pause] [3.3 seconds duration] My body feels like it carries a heavy, undeniable disgrace. [0.2 seconds pause]
- Tokens:
110
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.3 seconds pause] (ahem, 0.2 seconds) [0.0 seconds duration] I [4.3 seconds duration] must have been stained by those claws; I feel so deeply ashamed of myself. [0.4 seconds pause] [3.3 seconds duration] My body feels like it carries a heavy, undeniable disgrace. [0.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.9751.0000.2870.0008.80 s+0.10 s
SFT220.9941.0000.2870.0008.80 s+0.10 s
SFT230.9941.0000.0000.0008.80 s+0.10 s
SFT240.9951.0000.0000.0008.80 s+0.10 s
SFT2 + DPO210.9931.0000.2870.0008.80 s+0.10 s
SFT2 + DPO220.9981.0000.0000.0008.80 s+0.10 s
SFT2 + DPO230.9941.0000.3360.0008.80 s+0.10 s
SFT2 + DPO240.9991.0000.0000.1608.80 s+0.10 s
Disgustdescripted 16.9 s · 1 burst(s)
SFT2
emotion 0.712duration err 0.08 soverall 0.483 ± 0.186
SFT2 + DPO2
emotion 0.831duration err 0.08 soverall 0.564 ± 0.179
script and prompt
[6.7 seconds duration] Siehst du, wenn der Verkäufer den Angebotspreis zu niedrig ansetzt, schenkt er im Grunde nur Geld weg, was wirklich beeinflusst, ob der Verkauf reibungslos abläuft. [0.4 seconds pause] [3.1 seconds duration] Es ist klar, dass der Startpreis einen großen Einfluss auf den Gesamterfolg hat. [2.3 seconds pause] (contented sigh, 0.7 seconds) [3.7 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Cecilia
- Instruction:
GENERAL: A voice very high in warmth. The warmth is so overwhelmingly massive that the voice essentially melts into a gigantic, enveloping physical hug of pure low-frequency resonance. It is a completely bass-heavy, hyper-cozy, 'chocolate-like' acoustic environment that completely swallows the listener's ears; reads as intense disgust.
SCRIPT:
[6.7 seconds duration] Siehst du, wenn der Verkäufer den Angebotspreis zu niedrig ansetzt, schenkt er im Grunde nur Geld weg, was wirklich beeinflusst, ob der Verkauf reibungslos abläuft. [0.4 seconds pause] [3.1 seconds duration] Es ist klar, dass der Startpreis einen großen Einfluss auf den Gesamterfolg hat. [2.3 seconds pause] (contented sigh, 0.7 seconds) [3.7 seconds pause]
- Tokens:
212
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[6.7 seconds duration] Siehst du, wenn der Verkäufer den Angebotspreis zu niedrig ansetzt, schenkt er im Grunde nur Geld weg, was wirklich beeinflusst, ob der Verkauf reibungslos abläuft. [0.4 seconds pause] [3.1 seconds duration] Es ist klar, dass der Startpreis einen großen Einfluss auf den Gesamterfolg hat. [2.3 seconds pause] (contented sigh, 0.7 seconds) [3.7 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.9390.5000.0000.23716.96 s+0.06 s
SFT220.8311.0000.0001.23717.04 s+0.14 s
SFT230.8961.0000.4680.13216.96 s+0.06 s
SFT240.1811.0000.0000.13216.96 s+0.06 s
SFT2 + DPO210.9400.6230.1240.02617.04 s+0.14 s
SFT2 + DPO220.9420.6590.0000.00016.96 s+0.06 s
SFT2 + DPO230.5200.5120.2430.00016.96 s+0.06 s
SFT2 + DPO240.9230.8360.0000.07916.96 s+0.06 s
Disgustenscripted 12.7 s · 2 burst(s)
SFT2
emotion 0.339duration err 0.22 soverall 0.461 ± 0.038
SFT2 + DPO2
emotion 0.220duration err 0.22 soverall 0.481 ± 0.021
script and prompt
[0.3 seconds duration] Really? [0.4 seconds duration] Yes. [0.3 seconds pause] [2.0 seconds duration] Alec Wynn, that's a lot easier, uh, (ahem, 0.1 seconds) [2.2 seconds duration] says, when is the giveaway being announced? [0.5 seconds duration] And you know what? [2.1 seconds duration] I have a little bit of info on the giveaway. [0.3 seconds pause] [0.9 seconds duration] Who won the Elgato? [0.7 seconds duration] Who won the Elgato? [0.3 seconds pause] [0.3 seconds duration] Who won? [0.3 seconds duration] Uh, (low mumble, 0.5 seconds) [0.5 seconds duration] we have... [1.0 seconds duration] Neves won!
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is energised, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is positive, slightly dominant, slightly guarded; reads as intense disgust; style: casual, conversational; average recording, some background noise; mildly explicit content; 12.5s, EN.
SCRIPT:
[0.3 seconds duration] Really? [0.4 seconds duration] Yes. [0.3 seconds pause] [2.0 seconds duration] Alec Wynn, that's a lot easier, uh, (ahem, 0.1 seconds) [2.2 seconds duration] says, when is the giveaway being announced? [0.5 seconds duration] And you know what? [2.1 seconds duration] I have a little bit of info on the giveaway. [0.3 seconds pause] [0.9 seconds duration] Who won the Elgato? [0.7 seconds duration] Who won the Elgato? [0.3 seconds pause] [0.3 seconds duration] Who won? [0.3 seconds duration] Uh, (low mumble, 0.5 seconds) [0.5 seconds duration] we have... [1.0 seconds duration] Neves won!
- Tokens:
156
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.3 seconds duration] Really? [0.4 seconds duration] Yes. [0.3 seconds pause] [2.0 seconds duration] Alec Wynn, that's a lot easier, uh, (ahem, 0.1 seconds) [2.2 seconds duration] says, when is the giveaway being announced? [0.5 seconds duration] And you know what? [2.1 seconds duration] I have a little bit of info on the giveaway. [0.3 seconds pause] [0.9 seconds duration] Who won the Elgato? [0.7 seconds duration] Who won the Elgato? [0.3 seconds pause] [0.3 seconds duration] Who won? [0.3 seconds duration] Uh, (low mumble, 0.5 seconds) [0.5 seconds duration] we have... [1.0 seconds duration] Neves won!
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5631.0000.6840.22212.56 s-0.14 s
SFT220.1811.0000.3710.04412.48 s-0.22 s
SFT230.4291.0000.4630.15612.40 s-0.30 s
SFT240.1811.0000.7450.04412.48 s-0.22 s
SFT2 + DPO210.1811.0000.4670.06712.48 s-0.22 s
SFT2 + DPO220.4291.0000.5130.06712.48 s-0.22 s
SFT2 + DPO230.0891.0000.7610.06712.48 s-0.22 s
SFT2 + DPO240.1811.0000.5260.06712.48 s-0.22 s
Distressdescripted 15.9 s · 1 burst(s)
SFT2
emotion 0.628duration err 0.04 soverall 0.241 ± 0.118
SFT2 + DPO2
emotion 0.628duration err 0.08 soverall 0.211 ± 0.074
script and prompt
(surprised gasp, 0.1 seconds) [0.1 seconds duration] Um [5.4 seconds duration] ganz klar zu sein, ist für dich im Moment jedes Produkt, das Verstopfung verursachen könnte, absolut verboten, weil es allerlei Probleme [0.3 seconds pause] [9.8 seconds duration] mit deinem Verdauungssystem auslösen kann. [0.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Rose
- Instruction:
GENERAL: A voice notably high in monologue style. The delivery is distinctly monologic and introspective, featuring clear solo structure, introspective phrasing, and a clearly one-person-show register. The speaker sounds engaged with monologue aesthetics, naturally introspective; reads as intense distress.
SCRIPT:
(surprised gasp, 0.1 seconds) [0.1 seconds duration] Um [5.4 seconds duration] ganz klar zu sein, ist für dich im Moment jedes Produkt, das Verstopfung verursachen könnte, absolut verboten, weil es allerlei Probleme [0.3 seconds pause] [9.8 seconds duration] mit deinem Verdauungssystem auslösen kann. [0.2 seconds pause]
- Tokens:
199
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(surprised gasp, 0.1 seconds) [0.1 seconds duration] Um [5.4 seconds duration] ganz klar zu sein, ist für dich im Moment jedes Produkt, das Verstopfung verursachen könnte, absolut verboten, weil es allerlei Probleme [0.3 seconds pause] [9.8 seconds duration] mit deinem Verdauungssystem auslösen kann. [0.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6281.0000.3100.51915.92 s+0.02 s
SFT220.6281.0000.0920.14815.92 s+0.02 s
SFT230.6280.5000.2750.03715.92 s+0.02 s
SFT240.6280.5000.1843.81516.00 s+0.10 s
SFT2 + DPO210.6281.0000.2170.40715.92 s+0.02 s
SFT2 + DPO220.6281.0000.1562.92616.08 s+0.18 s
SFT2 + DPO230.6280.5000.0890.14815.92 s+0.02 s
SFT2 + DPO240.6280.5770.0000.00016.00 s+0.10 s
Distressenscripted 7.8 s · 1 burst(s)
SFT2
emotion 0.944duration err 0.04 soverall 0.828 ± 0.149
SFT2 + DPO2
emotion 0.976duration err 0.04 soverall 0.865 ± 0.075
script and prompt
(contented sigh, 0.2 seconds) [0.3 seconds pause] [4.3 seconds duration] Please, for the love of everything, do not fill this space with rubbish or hurt the forum in any other way. [0.4 seconds pause] [2.6 seconds duration] I feel such embarrassment if someone ruins what we've built here.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Bernard
- Instruction:
GENERAL: A voice slightly expressing shame, a small, shame-filled voice, cringing and self-reproaching, unable to look up, building and building. The feeling is present but dialled down, not performed at full size; reads as intense distress.
SCRIPT:
(contented sigh, 0.2 seconds) [0.3 seconds pause] [4.3 seconds duration] Please, for the love of everything, do not fill this space with rubbish or hurt the forum in any other way. [0.4 seconds pause] [2.6 seconds duration] I feel such embarrassment if someone ruins what we've built here.
- Tokens:
98
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
(contented sigh, 0.2 seconds) [0.3 seconds pause] [4.3 seconds duration] Please, for the love of everything, do not fill this space with rubbish or hurt the forum in any other way. [0.4 seconds pause] [2.6 seconds duration] I feel such embarrassment if someone ruins what we've built here.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.9550.9630.7870.0007.84 s+0.04 s
SFT220.9300.5410.8870.0007.84 s+0.04 s
SFT230.9160.5000.0000.0007.84 s+0.04 s
SFT240.9761.0000.8870.0007.84 s+0.04 s
SFT2 + DPO210.9581.0000.1770.0007.84 s+0.04 s
SFT2 + DPO220.9771.0000.9610.0007.84 s+0.04 s
SFT2 + DPO230.9731.0000.0000.0007.84 s+0.04 s
SFT2 + DPO240.9961.0000.2870.0007.84 s+0.04 s
Doubtdescripted 22.7 s · 3 burst(s)
SFT2
emotion 0.784duration err 0.10 soverall 0.643 ± 0.188
SFT2 + DPO2
emotion 0.782duration err 0.10 soverall 0.571 ± 0.158
script and prompt
[1.8 seconds duration] Die machen jetzt schon solche Geschichten da. [0.3 seconds pause] [1.8 seconds duration] Das mit Abstürzen und, (low mumble, 0.1 seconds) [1.7 seconds duration] äh, keiner kümmert sich und so. [0.4 seconds pause] [1.6 seconds duration] Und nicht mal irgendwie, dass, (ahem, 0.1 seconds) [3.3 seconds duration] ich hab nichts davon gehört, dass sie sich irgendwo bei irgendwem entschuldigen. [0.6 seconds pause] [2.8 seconds duration] Und die Leute haben da jetzt fünf Jahre eingezahlt, ja. [0.4 seconds pause] [3.0 seconds duration] Die haben fünf Jahre den Web-Space nach und nach erkauft. [0.6 seconds pause] [0.3 seconds duration] Und, (low mumble, 0.3 seconds) [2.4 seconds duration] ähm, jetzt stell dir vor, die sind irgendwann weg. [0.7 seconds pause] [0.1 seconds duration] Ja. [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, slightly rough, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense doubt; style: conversational, casual; good recording, quiet background; 22.7s, DE.
SCRIPT:
[1.8 seconds duration] Die machen jetzt schon solche Geschichten da. [0.3 seconds pause] [1.8 seconds duration] Das mit Abstürzen und, (low mumble, 0.1 seconds) [1.7 seconds duration] äh, keiner kümmert sich und so. [0.4 seconds pause] [1.6 seconds duration] Und nicht mal irgendwie, dass, (ahem, 0.1 seconds) [3.3 seconds duration] ich hab nichts davon gehört, dass sie sich irgendwo bei irgendwem entschuldigen. [0.6 seconds pause] [2.8 seconds duration] Und die Leute haben da jetzt fünf Jahre eingezahlt, ja. [0.4 seconds pause] [3.0 seconds duration] Die haben fünf Jahre den Web-Space nach und nach erkauft. [0.6 seconds pause] [0.3 seconds duration] Und, (low mumble, 0.3 seconds) [2.4 seconds duration] ähm, jetzt stell dir vor, die sind irgendwann weg. [0.7 seconds pause] [0.1 seconds duration] Ja. [0.4 seconds pause]
- Tokens:
283
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[1.8 seconds duration] Die machen jetzt schon solche Geschichten da. [0.3 seconds pause] [1.8 seconds duration] Das mit Abstürzen und, (low mumble, 0.1 seconds) [1.7 seconds duration] äh, keiner kümmert sich und so. [0.4 seconds pause] [1.6 seconds duration] Und nicht mal irgendwie, dass, (ahem, 0.1 seconds) [3.3 seconds duration] ich hab nichts davon gehört, dass sie sich irgendwo bei irgendwem entschuldigen. [0.6 seconds pause] [2.8 seconds duration] Und die Leute haben da jetzt fünf Jahre eingezahlt, ja. [0.4 seconds pause] [3.0 seconds duration] Die haben fünf Jahre den Web-Space nach und nach erkauft. [0.6 seconds pause] [0.3 seconds duration] Und, (low mumble, 0.3 seconds) [2.4 seconds duration] ähm, jetzt stell dir vor, die sind irgendwann weg. [0.7 seconds pause] [0.1 seconds duration] Ja. [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.9391.0000.5620.12122.56 s-0.14 s
SFT220.6741.0000.3110.12122.64 s-0.06 s
SFT230.9181.0000.6810.10622.64 s-0.06 s
SFT240.6031.0000.6340.13622.56 s-0.14 s
SFT2 + DPO210.8181.0000.3600.98522.56 s-0.14 s
SFT2 + DPO220.9221.0000.5540.12122.64 s-0.06 s
SFT2 + DPO230.7011.0000.8080.10622.64 s-0.06 s
SFT2 + DPO240.6891.0000.5940.12122.56 s-0.14 s
Doubtenscripted 19.5 s · 1 burst(s)
SFT2
emotion 0.397duration err 0.02 soverall 0.409 ± 0.024
SFT2 + DPO2
emotion 0.276duration err 0.02 soverall 0.399 ± 0.017
script and prompt
[0.2 seconds pause] [2.5 seconds duration] No, no, I'll be disappointing people, so. [0.9 seconds pause] [2.4 seconds duration] So you were like, I'll just go until I can't go anymore. [0.4 seconds pause] [2.6 seconds duration] And then sure enough, like, the signing was like, hours. [0.5 seconds pause] [4.5 seconds duration] And at the end of the hours, plural, I looked over and there you were. [0.6 seconds pause] [4.8 seconds duration] Head down, in your arms, greasy hair splayed all about. (ahem, 0.1 seconds)
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is mildly positive, slightly submissive, neutral openness; reads as intense doubt; style: casual, playful; average recording, quiet background; mildly explicit content; 19.7s, EN.
SCRIPT:
[0.2 seconds pause] [2.5 seconds duration] No, no, I'll be disappointing people, so. [0.9 seconds pause] [2.4 seconds duration] So you were like, I'll just go until I can't go anymore. [0.4 seconds pause] [2.6 seconds duration] And then sure enough, like, the signing was like, hours. [0.5 seconds pause] [4.5 seconds duration] And at the end of the hours, plural, I looked over and there you were. [0.6 seconds pause] [4.8 seconds duration] Head down, in your arms, greasy hair splayed all about. (ahem, 0.1 seconds)
- Tokens:
245
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.2 seconds pause] [2.5 seconds duration] No, no, I'll be disappointing people, so. [0.9 seconds pause] [2.4 seconds duration] So you were like, I'll just go until I can't go anymore. [0.4 seconds pause] [2.6 seconds duration] And then sure enough, like, the signing was like, hours. [0.5 seconds pause] [4.5 seconds duration] And at the end of the hours, plural, I looked over and there you were. [0.6 seconds pause] [4.8 seconds duration] Head down, in your arms, greasy hair splayed all about. (ahem, 0.1 seconds)
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5801.0000.2870.01919.52 s+0.02 s
SFT220.4701.0000.0000.01919.52 s+0.02 s
SFT230.0661.0000.0000.01919.52 s+0.02 s
SFT240.4701.0000.0000.00019.52 s+0.02 s
SFT2 + DPO210.0241.0000.1570.03719.52 s+0.02 s
SFT2 + DPO220.0661.0000.2250.07419.52 s+0.02 s
SFT2 + DPO230.5441.0000.0000.05619.52 s+0.02 s
SFT2 + DPO240.4701.0000.0000.03719.52 s+0.02 s
Elationdescripted 19.8 s · 1 burst(s)
SFT2
emotion 0.473duration err 0.06 soverall 0.470 ± 0.081
SFT2 + DPO2
emotion 0.742duration err 0.06 soverall 0.591 ± 0.112
script and prompt
(low mumble, 0.1 seconds) [0.7 seconds duration] eeehm, [0.5 seconds pause] [1.8 seconds duration] lizenziert von niemandem. [3.3 seconds duration] Das ganze gibt's also entsprechend nur als Fans abzuschauen. [0.4 seconds pause] [4.3 seconds duration] Zufälligerweise auf Deutsch bei einer Fans-Up-Gruppe, die genauso heißt wie dieser Podcast hier. [0.7 seconds duration] Die haben (ahem) [0.3 seconds pause] [1.4 seconds duration] uns den Namen geklaut, die blöden Penner. [2.2 seconds duration] Diese Schweine, die haben einfach den Namen geklaut, ey. [1.4 seconds duration] Echt, ich werd gleich mal verklagen, ey. [0.5 seconds pause] [1.1 seconds duration] Das weißte du nur ins Bezirk. [1.1 seconds duration] Echt mal, wer draus ist denn hier?
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, slightly submissive, slightly guarded; reads as intense elation; style: casual, conversational; good recording, quiet background; 19.7s, DE.
SCRIPT:
(low mumble, 0.1 seconds) [0.7 seconds duration] eeehm, [0.5 seconds pause] [1.8 seconds duration] lizenziert von niemandem. [3.3 seconds duration] Das ganze gibt's also entsprechend nur als Fans abzuschauen. [0.4 seconds pause] [4.3 seconds duration] Zufälligerweise auf Deutsch bei einer Fans-Up-Gruppe, die genauso heißt wie dieser Podcast hier. [0.7 seconds duration] Die haben (ahem) [0.3 seconds pause] [1.4 seconds duration] uns den Namen geklaut, die blöden Penner. [2.2 seconds duration] Diese Schweine, die haben einfach den Namen geklaut, ey. [1.4 seconds duration] Echt, ich werd gleich mal verklagen, ey. [0.5 seconds pause] [1.1 seconds duration] Das weißte du nur ins Bezirk. [1.1 seconds duration] Echt mal, wer draus ist denn hier?
- Tokens:
246
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(low mumble, 0.1 seconds) [0.7 seconds duration] eeehm, [0.5 seconds pause] [1.8 seconds duration] lizenziert von niemandem. [3.3 seconds duration] Das ganze gibt's also entsprechend nur als Fans abzuschauen. [0.4 seconds pause] [4.3 seconds duration] Zufälligerweise auf Deutsch bei einer Fans-Up-Gruppe, die genauso heißt wie dieser Podcast hier. [0.7 seconds duration] Die haben (ahem) [0.3 seconds pause] [1.4 seconds duration] uns den Namen geklaut, die blöden Penner. [2.2 seconds duration] Diese Schweine, die haben einfach den Namen geklaut, ey. [1.4 seconds duration] Echt, ich werd gleich mal verklagen, ey. [0.5 seconds pause] [1.1 seconds duration] Das weißte du nur ins Bezirk. [1.1 seconds duration] Echt mal, wer draus ist denn hier?
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7401.0000.5070.09119.76 s-0.04 s
SFT220.0081.0000.5070.22719.68 s-0.12 s
SFT230.5061.0000.1640.12119.76 s-0.04 s
SFT240.6371.0000.7260.10619.76 s-0.04 s
SFT2 + DPO210.8431.0000.5490.07619.76 s-0.04 s
SFT2 + DPO220.5061.0000.4870.07619.84 s+0.04 s
SFT2 + DPO230.8790.8210.2650.21219.68 s-0.12 s
SFT2 + DPO240.7401.0000.2540.13619.76 s-0.04 s
Elationenscripted 7.5 s · 1 burst(s)
SFT2
emotion 0.142duration err 0.46 soverall 0.315 ± 0.071
SFT2 + DPO2
emotion 0.140duration err 0.46 soverall 0.256 ± 0.076
script and prompt
[3.8 seconds duration] Nobody in my doctor's office speaking is, I gotta get Rosetta Stone to fill out my prescription. [2.4 seconds duration] This is, this is from Dr. (contented sigh, 0.5 seconds) [0.3 seconds duration] Krrrrr. [0.5 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is energised, brisk, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, slightly rough, thin; average clarity, some disfluency, wide pitch range, normal breath; affect is positive, slightly dominant, fairly guarded; reads as intense elation; style: casual, playful; below-average recording, some background noise; mildly explicit content; 7.0s, EN.
SCRIPT:
[3.8 seconds duration] Nobody in my doctor's office speaking is, I gotta get Rosetta Stone to fill out my prescription. [2.4 seconds duration] This is, this is from Dr. (contented sigh, 0.5 seconds) [0.3 seconds duration] Krrrrr. [0.5 seconds pause]
- Tokens:
88
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[3.8 seconds duration] Nobody in my doctor's office speaking is, I gotta get Rosetta Stone to fill out my prescription. [2.4 seconds duration] This is, this is from Dr. (contented sigh, 0.5 seconds) [0.3 seconds duration] Krrrrr. [0.5 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0080.5250.1640.2087.04 s-0.46 s
SFT220.0081.0000.3100.1677.04 s-0.46 s
SFT230.0081.0000.0000.2087.04 s-0.46 s
SFT240.5431.0000.0000.2087.04 s-0.46 s
SFT2 + DPO210.0081.0000.0000.2087.04 s-0.46 s
SFT2 + DPO220.5350.5080.0830.1257.04 s-0.46 s
SFT2 + DPO230.0080.5000.0000.2087.04 s-0.46 s
SFT2 + DPO240.0081.0000.0000.2087.04 s-0.46 s
Embarrassmentdescripted 16.4 s · 0 burst(s)
SFT2
emotion 0.484duration err 0.00 soverall 0.377 ± 0.007
SFT2 + DPO2
emotion 0.313duration err 0.00 soverall 0.367 ± 0.005
script and prompt
[0.8 seconds duration] Ja, hi zusammen. [0.3 seconds pause] [0.8 seconds duration] Ich bin Nils Hoffmann. [1.9 seconds duration] Ich bin der Geschäftsführer von Possible. [0.7 seconds pause] [4.7 seconds duration] Wir arbeiten auch an der Schnittstelle zwischen Verwaltung und Tech-Unternehmen, äh, (ahem) [7.2 seconds duration] machen da Research-Projekte, Beratungsprojekte, führen Startup-Programme durch, genau, und versuchen, die Verwaltung ein Stück innovativer zu machen.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense embarrassment; style: monologue, conversational; good recording, quiet background; 16.4s, DE.
SCRIPT:
[0.8 seconds duration] Ja, hi zusammen. [0.3 seconds pause] [0.8 seconds duration] Ich bin Nils Hoffmann. [1.9 seconds duration] Ich bin der Geschäftsführer von Possible. [0.7 seconds pause] [4.7 seconds duration] Wir arbeiten auch an der Schnittstelle zwischen Verwaltung und Tech-Unternehmen, äh, (ahem) [7.2 seconds duration] machen da Research-Projekte, Beratungsprojekte, führen Startup-Programme durch, genau, und versuchen, die Verwaltung ein Stück innovativer zu machen.
- Tokens:
205
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.8 seconds duration] Ja, hi zusammen. [0.3 seconds pause] [0.8 seconds duration] Ich bin Nils Hoffmann. [1.9 seconds duration] Ich bin der Geschäftsführer von Possible. [0.7 seconds pause] [4.7 seconds duration] Wir arbeiten auch an der Schnittstelle zwischen Verwaltung und Tech-Unternehmen, äh, (ahem) [7.2 seconds duration] machen da Research-Projekte, Beratungsprojekte, führen Startup-Programme durch, genau, und versuchen, die Verwaltung ein Stück innovativer zu machen.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5221.0000.0000.04516.40 s+0.00 s
SFT220.6211.0000.0000.04516.40 s+0.00 s
SFT230.1721.0000.0000.09116.40 s+0.00 s
SFT240.6231.0000.0000.06816.40 s+0.00 s
SFT2 + DPO210.1851.0000.0000.11416.40 s+0.00 s
SFT2 + DPO220.3811.0000.0000.09116.40 s+0.00 s
SFT2 + DPO230.3631.0000.0000.06816.40 s+0.00 s
SFT2 + DPO240.3251.0000.0000.09116.40 s+0.00 s
Embarrassmentenscripted 9.4 s · 1 burst(s)
SFT2
emotion 0.161duration err 0.04 soverall 0.382 ± 0.013
SFT2 + DPO2
emotion 0.156duration err 0.04 soverall 0.376 ± 0.008
script and prompt
[2.2 seconds duration] Completely gone off the radar. [0.5 seconds pause] [1.3 seconds duration] They get a private investigator. [2.0 seconds duration] The guy's only got PO boxes and stuff. [0.3 seconds duration] Oh really? [0.2 seconds pause] [1.6 seconds duration] And they find him in the mountains (ahem, 0.3 seconds) [0.7 seconds duration] and he's a hermit. [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is mildly negative, slightly dominant, neutral openness; reads as intense embarrassment; style: casual, conversational; average recording, quiet background; mildly explicit content; 9.5s, EN.
SCRIPT:
[2.2 seconds duration] Completely gone off the radar. [0.5 seconds pause] [1.3 seconds duration] They get a private investigator. [2.0 seconds duration] The guy's only got PO boxes and stuff. [0.3 seconds duration] Oh really? [0.2 seconds pause] [1.6 seconds duration] And they find him in the mountains (ahem, 0.3 seconds) [0.7 seconds duration] and he's a hermit. [0.3 seconds pause]
- Tokens:
118
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[2.2 seconds duration] Completely gone off the radar. [0.5 seconds pause] [1.3 seconds duration] They get a private investigator. [2.0 seconds duration] The guy's only got PO boxes and stuff. [0.3 seconds duration] Oh really? [0.2 seconds pause] [1.6 seconds duration] And they find him in the mountains (ahem, 0.3 seconds) [0.7 seconds duration] and he's a hermit. [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0631.0000.0000.0659.44 s+0.04 s
SFT220.2111.0000.0000.0979.44 s+0.04 s
SFT230.2111.0000.0000.0329.44 s+0.04 s
SFT240.1591.0000.0000.0009.44 s+0.04 s
SFT2 + DPO210.1721.0000.0000.0659.44 s+0.04 s
SFT2 + DPO220.1351.0000.0000.0979.44 s+0.04 s
SFT2 + DPO230.2381.0000.0000.0659.36 s-0.04 s
SFT2 + DPO240.0801.0000.0000.0329.44 s+0.04 s
Emotional Numbnessdescripted 9.0 s · 1 burst(s)
SFT2
emotion 0.556duration err 0.04 soverall 0.490 ± 0.099
SFT2 + DPO2
emotion 0.616duration err 0.04 soverall 0.382 ± 0.095
script and prompt
[0.2 seconds duration] Ah, [0.2 seconds pause] (ahem, 0.4 seconds) [0.5 seconds pause] [6.1 seconds duration] ist wahrscheinlich auch, weil ich das so früh gesehen hab und mich das so lange begleitet, aber der geht mir tatsächlich bis heute unter die Haut. [0.5 seconds pause] [1.1 seconds duration] Wie, wie stehst du dazu?
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense emotional numbness; style: conversational, casual; average recording, quiet background; 9.0s, DE.
SCRIPT:
[0.2 seconds duration] Ah, [0.2 seconds pause] (ahem, 0.4 seconds) [0.5 seconds pause] [6.1 seconds duration] ist wahrscheinlich auch, weil ich das so früh gesehen hab und mich das so lange begleitet, aber der geht mir tatsächlich bis heute unter die Haut. [0.5 seconds pause] [1.1 seconds duration] Wie, wie stehst du dazu?
- Tokens:
112
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.2 seconds duration] Ah, [0.2 seconds pause] (ahem, 0.4 seconds) [0.5 seconds pause] [6.1 seconds duration] ist wahrscheinlich auch, weil ich das so früh gesehen hab und mich das so lange begleitet, aber der geht mir tatsächlich bis heute unter die Haut. [0.5 seconds pause] [1.1 seconds duration] Wie, wie stehst du dazu?
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1801.0000.0000.0628.96 s-0.04 s
SFT220.8061.0000.2170.0948.96 s-0.04 s
SFT230.7711.0000.0000.1258.96 s-0.04 s
SFT240.4671.0000.3100.0948.96 s-0.04 s
SFT2 + DPO210.5771.0000.2750.0948.96 s-0.04 s
SFT2 + DPO220.6651.0000.1690.1258.96 s-0.04 s
SFT2 + DPO230.6861.0000.3360.0948.96 s-0.04 s
SFT2 + DPO240.5341.0000.2541.0008.96 s-0.04 s
Emotional Numbnessenscripted 17.1 s · 1 burst(s)
SFT2
emotion 0.518duration err 0.06 soverall 0.493 ± 0.079
SFT2 + DPO2
emotion 0.551duration err 0.06 soverall 0.578 ± 0.093
script and prompt
[1.3 seconds duration] Michael Hanfleet who works at (ahem, 0.2 seconds) [5.1 seconds duration] Watt Culture has this great line that I've parroted every single time we've talked about punk on collision or anything like that. [0.5 seconds pause] [5.3 seconds duration] It's a really great summation of the whole thing, which is you either get YCM punk is worth it or you don't and there's no in between. [0.6 seconds pause] [1.5 seconds duration] And I think that's so accurate. [0.6 seconds pause] [1.7 seconds duration] And in this scenario, [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is positive, slightly submissive, neutral openness; reads as intense emotional numbness; style: conversational, casual; average recording, quiet background; 17.1s, EN.
SCRIPT:
[1.3 seconds duration] Michael Hanfleet who works at (ahem, 0.2 seconds) [5.1 seconds duration] Watt Culture has this great line that I've parroted every single time we've talked about punk on collision or anything like that. [0.5 seconds pause] [5.3 seconds duration] It's a really great summation of the whole thing, which is you either get YCM punk is worth it or you don't and there's no in between. [0.6 seconds pause] [1.5 seconds duration] And I think that's so accurate. [0.6 seconds pause] [1.7 seconds duration] And in this scenario, [0.3 seconds pause]
- Tokens:
213
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[1.3 seconds duration] Michael Hanfleet who works at (ahem, 0.2 seconds) [5.1 seconds duration] Watt Culture has this great line that I've parroted every single time we've talked about punk on collision or anything like that. [0.5 seconds pause] [5.3 seconds duration] It's a really great summation of the whole thing, which is you either get YCM punk is worth it or you don't and there's no in between. [0.6 seconds pause] [1.5 seconds duration] And I think that's so accurate. [0.6 seconds pause] [1.7 seconds duration] And in this scenario, [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.4671.0000.3100.03117.04 s-0.06 s
SFT220.5891.0000.9230.00017.04 s-0.06 s
SFT230.6111.0000.8520.03117.04 s-0.06 s
SFT240.4061.0000.0000.03117.04 s-0.06 s
SFT2 + DPO210.5091.0000.9230.00017.04 s-0.06 s
SFT2 + DPO220.3041.0000.9230.03117.04 s-0.06 s
SFT2 + DPO230.6411.0000.2440.00017.04 s-0.06 s
SFT2 + DPO240.7511.0000.8520.03117.04 s-0.06 s
Fatigue Exhaustiondescripted 9.4 s · 1 burst(s)
SFT2
emotion 0.478duration err 0.20 soverall 0.368 ± 0.085
SFT2 + DPO2
emotion 0.606duration err 0.20 soverall 0.415 ± 0.017
script and prompt
[1.5 seconds duration] ja, die hat nicht diese Stimme. [0.6 seconds duration] What the fuck? [6.9 seconds duration] Ich meine, die hat auch in Resiour eine relativ hohe Stimme, aber eher so eine lieblich hohe Stimme und nicht so eine, so eine kreischende, äh, (exhausted groan, 0.4 seconds)
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is positive, slightly dominant, neutral openness; reads as intense fatigue exhaustion; style: conversational, casual; average recording, some background noise; 9.6s, DE.
SCRIPT:
[1.5 seconds duration] ja, die hat nicht diese Stimme. [0.6 seconds duration] What the fuck? [6.9 seconds duration] Ich meine, die hat auch in Resiour eine relativ hohe Stimme, aber eher so eine lieblich hohe Stimme und nicht so eine, so eine kreischende, äh, (exhausted groan, 0.4 seconds)
- Tokens:
120
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[1.5 seconds duration] ja, die hat nicht diese Stimme. [0.6 seconds duration] What the fuck? [6.9 seconds duration] Ich meine, die hat auch in Resiour eine relativ hohe Stimme, aber eher so eine lieblich hohe Stimme und nicht so eine, so eine kreischende, äh, (exhausted groan, 0.4 seconds)
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5651.0000.3360.0869.60 s+0.20 s
SFT220.4550.5000.2540.1149.60 s+0.20 s
SFT230.2291.0000.2170.1439.60 s+0.20 s
SFT240.6641.0000.2250.0869.60 s+0.20 s
SFT2 + DPO210.6641.0000.2070.1149.60 s+0.20 s
SFT2 + DPO220.5391.0000.2980.1149.60 s+0.20 s
SFT2 + DPO230.6111.0000.4490.1149.60 s+0.20 s
SFT2 + DPO240.6111.0000.1760.1149.60 s+0.20 s
Fatigue Exhaustionenscripted 12.6 s · 1 burst(s)
SFT2
emotion 0.423duration err 0.20 soverall 0.488 ± 0.026
SFT2 + DPO2
emotion 0.466duration err 0.20 soverall 0.482 ± 0.021
script and prompt
[2.2 seconds duration] And I also, I would like to take back, uh, (low mumble, 0.2 seconds) [0.4 seconds pause] [1.3 seconds duration] me shitting on Papa Murphy's. [2.5 seconds duration] Kyle was bringing up Papa Murphy's, and I was like, (ahem) [6.0 seconds duration] oh what the fuck, that's not even one of the- and then I looked it up, and it is one of the biggest pizza places in the country.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense fatigue exhaustion; style: conversational, casual; average recording, some background noise; 12.5s, EN.
SCRIPT:
[2.2 seconds duration] And I also, I would like to take back, uh, (low mumble, 0.2 seconds) [0.4 seconds pause] [1.3 seconds duration] me shitting on Papa Murphy's. [2.5 seconds duration] Kyle was bringing up Papa Murphy's, and I was like, (ahem) [6.0 seconds duration] oh what the fuck, that's not even one of the- and then I looked it up, and it is one of the biggest pizza places in the country.
- Tokens:
155
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[2.2 seconds duration] And I also, I would like to take back, uh, (low mumble, 0.2 seconds) [0.4 seconds pause] [1.3 seconds duration] me shitting on Papa Murphy's. [2.5 seconds duration] Kyle was bringing up Papa Murphy's, and I was like, (ahem) [6.0 seconds duration] oh what the fuck, that's not even one of the- and then I looked it up, and it is one of the biggest pizza places in the country.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5391.0000.4650.01912.40 s-0.20 s
SFT220.3081.0000.5070.01912.40 s-0.20 s
SFT230.5391.0000.6700.01912.40 s-0.20 s
SFT240.3081.0000.3380.03812.40 s-0.20 s
SFT2 + DPO210.5651.0000.3960.05712.40 s-0.20 s
SFT2 + DPO220.3081.0000.7560.07512.40 s-0.20 s
SFT2 + DPO230.4251.0000.5070.03812.40 s-0.20 s
SFT2 + DPO240.5651.0000.5270.05712.40 s-0.20 s
Feardescripted 14.2 s · 1 burst(s)
SFT2
emotion 0.515duration err 0.22 soverall 0.394 ± 0.145
SFT2 + DPO2
emotion 0.630duration err 0.14 soverall 0.529 ± 0.083
script and prompt
[0.6 seconds pause] (exhausted groan, 0.1 seconds) [1.5 seconds duration] Ach, [5.5 seconds duration] wieder das wahre Gefühl des Gewinnanteils zu spüren, statt dieser starren, unbarmherzigen Sichtweise. [0.6 seconds pause] [5.9 seconds duration] Ich sehne mich nach einer Chance, die feste Zahlung über die nur des einzigen Jahres hart erarbeiteten Gewinne hinaus zu sehen.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Andreas
- Instruction:
GENERAL: A voice unmistakably expressing longing, an aching, yearning voice, full of wistful longing, reaching for what is out of reach, with every word; reads as intense fear.
SCRIPT:
[0.6 seconds pause] (exhausted groan, 0.1 seconds) [1.5 seconds duration] Ach, [5.5 seconds duration] wieder das wahre Gefühl des Gewinnanteils zu spüren, statt dieser starren, unbarmherzigen Sichtweise. [0.6 seconds pause] [5.9 seconds duration] Ich sehne mich nach einer Chance, die feste Zahlung über die nur des einzigen Jahres hart erarbeiteten Gewinne hinaus zu sehen.
- Tokens:
179
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.6 seconds pause] (exhausted groan, 0.1 seconds) [1.5 seconds duration] Ach, [5.5 seconds duration] wieder das wahre Gefühl des Gewinnanteils zu spüren, statt dieser starren, unbarmherzigen Sichtweise. [0.6 seconds pause] [5.9 seconds duration] Ich sehne mich nach einer Chance, die feste Zahlung über die nur des einzigen Jahres hart erarbeiteten Gewinne hinaus zu sehen.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6821.0000.7260.11414.40 s+0.20 s
SFT220.5780.7480.0000.00014.32 s+0.12 s
SFT230.7560.6370.5490.00014.32 s+0.12 s
SFT240.0440.5610.0000.08614.64 s+0.44 s
SFT2 + DPO210.8231.0000.3360.11414.40 s+0.20 s
SFT2 + DPO220.4501.0000.2750.00014.32 s+0.12 s
SFT2 + DPO230.5851.0000.7260.05714.32 s+0.12 s
SFT2 + DPO240.6620.9970.2980.00014.32 s+0.12 s
Fearenscripted 17.1 s · 0 burst(s)
SFT2
emotion 0.278duration err 0.02 soverall 0.311 ± 0.064
SFT2 + DPO2
emotion 0.267duration err 0.02 soverall 0.347 ± 0.002
script and prompt
[1.7 seconds duration] Phil, uhm, (low mumble) [0.5 seconds pause] [8.7 seconds duration] Kip Lewis, guys like that, Shane Whittier, you know, you start kinda seeing, okay, who's staying, who's shuffling, all the rest of that stuff, because you're gonna have some dogs that come in, uhm, (low mumble) [1.5 seconds duration] and could potentially come in. [4.7 seconds duration] So, yeaah, like I said, I think, I think you see that, I think it bodes well for us, but,
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is subdued, normal-paced, relaxed, fairly steady; timbre is neutral-toned, neutral-bright, slightly rough, balanced body; somewhat unclear, frequent disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense fear; style: casual, conversational; average recording, quiet background; 17.2s, EN.
SCRIPT:
[1.7 seconds duration] Phil, uhm, (low mumble) [0.5 seconds pause] [8.7 seconds duration] Kip Lewis, guys like that, Shane Whittier, you know, you start kinda seeing, okay, who's staying, who's shuffling, all the rest of that stuff, because you're gonna have some dogs that come in, uhm, (low mumble) [1.5 seconds duration] and could potentially come in. [4.7 seconds duration] So, yeaah, like I said, I think, I think you see that, I think it bodes well for us, but,
- Tokens:
214
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[1.7 seconds duration] Phil, uhm, (low mumble) [0.5 seconds pause] [8.7 seconds duration] Kip Lewis, guys like that, Shane Whittier, you know, you start kinda seeing, okay, who's staying, who's shuffling, all the rest of that stuff, because you're gonna have some dogs that come in, uhm, (low mumble) [1.5 seconds duration] and could potentially come in. [4.7 seconds duration] So, yeaah, like I said, I think, I think you see that, I think it bodes well for us, but,
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1571.0000.0000.14817.12 s+0.02 s
SFT220.2381.0000.0000.14817.12 s+0.02 s
SFT230.2651.0000.0001.00017.12 s+0.02 s
SFT240.4501.0000.0000.14817.12 s+0.02 s
SFT2 + DPO210.2651.0000.0000.14817.12 s+0.02 s
SFT2 + DPO220.4141.0000.0000.16417.12 s+0.02 s
SFT2 + DPO230.0441.0000.0000.14817.12 s+0.02 s
SFT2 + DPO240.3441.0000.0000.14817.12 s+0.02 s
Helplessnessdescripted 10.0 s · 1 burst(s)
SFT2
emotion 0.957duration err 0.00 soverall 0.825 ± 0.131
SFT2 + DPO2
emotion 0.966duration err 0.00 soverall 0.819 ± 0.080
script and prompt
[5.9 seconds duration] Ich konnte nicht einmal den Grand Marnier hinzufügen; der Gedanke an diese Migräne bringt so ein schweres Gefühl zurück. [0.6 seconds pause] (contented sigh, 0.2 seconds) [0.5 seconds pause] [2.6 seconds duration] Also habe ich dieses Mal ganz auf Alkohol verzichtet. [0.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Alex
- Instruction:
GENERAL: A voice mildly expressing sadness, a soft, grief-worn voice, heavy and slow, on the edge of tears, impossible to hide. The feeling is present but dialled down, not performed at full size. The feeling is fought down rather than shown: it is held in and only leaks through at the edges of phrases; reads as intense helplessness.
SCRIPT:
[5.9 seconds duration] Ich konnte nicht einmal den Grand Marnier hinzufügen; der Gedanke an diese Migräne bringt so ein schweres Gefühl zurück. [0.6 seconds pause] (contented sigh, 0.2 seconds) [0.5 seconds pause] [2.6 seconds duration] Also habe ich dieses Mal ganz auf Alkohol verzichtet. [0.2 seconds pause]
- Tokens:
125
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[5.9 seconds duration] Ich konnte nicht einmal den Grand Marnier hinzufügen; der Gedanke an diese Migräne bringt so ein schweres Gefühl zurück. [0.6 seconds pause] (contented sigh, 0.2 seconds) [0.5 seconds pause] [2.6 seconds duration] Also habe ich dieses Mal ganz auf Alkohol verzichtet. [0.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.9600.6440.0000.03610.00 s+0.00 s
SFT220.9531.0000.8870.03610.00 s+0.00 s
SFT230.9511.0000.0000.03610.00 s+0.00 s
SFT240.9651.0000.9230.03610.00 s+0.00 s
SFT2 + DPO210.9781.0000.0000.03610.00 s+0.00 s
SFT2 + DPO220.9761.0000.0000.03610.00 s+0.00 s
SFT2 + DPO230.9501.0000.0000.03610.00 s+0.00 s
SFT2 + DPO240.9591.0000.9610.03610.00 s+0.00 s
Helplessnessenscripted 19.2 s · 1 burst(s)
SFT2
emotion 0.807duration err 0.08 soverall 0.675 ± 0.112
SFT2 + DPO2
emotion 0.779duration err 0.08 soverall 0.618 ± 0.066
script and prompt
[0.5 seconds pause] [4.1 seconds duration] Yeah, I actually had some surgery and then they told me that they'd put something up my bum. [0.8 seconds pause] [0.4 seconds duration] Really? [2.4 seconds duration] Well, they put like painkillers medication. [2.0 seconds duration] So I had, I think I had my elbow and my, (low mumble, 0.1 seconds) [7.2 seconds duration] um, my wrist cracked back in cause it was like bent, fell off my skateboard and I didn't, it didn't break my arm. [1.7 seconds duration] It bowed my arm because
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly negative, neutral stance, neutral openness; reads as intense helplessness; style: casual, conversational; average recording, quiet background; 19.3s, EN.
SCRIPT:
[0.5 seconds pause] [4.1 seconds duration] Yeah, I actually had some surgery and then they told me that they'd put something up my bum. [0.8 seconds pause] [0.4 seconds duration] Really? [2.4 seconds duration] Well, they put like painkillers medication. [2.0 seconds duration] So I had, I think I had my elbow and my, (low mumble, 0.1 seconds) [7.2 seconds duration] um, my wrist cracked back in cause it was like bent, fell off my skateboard and I didn't, it didn't break my arm. [1.7 seconds duration] It bowed my arm because
- Tokens:
240
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.5 seconds pause] [4.1 seconds duration] Yeah, I actually had some surgery and then they told me that they'd put something up my bum. [0.8 seconds pause] [0.4 seconds duration] Really? [2.4 seconds duration] Well, they put like painkillers medication. [2.0 seconds duration] So I had, I think I had my elbow and my, (low mumble, 0.1 seconds) [7.2 seconds duration] um, my wrist cracked back in cause it was like bent, fell off my skateboard and I didn't, it didn't break my arm. [1.7 seconds duration] It bowed my arm because
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7231.0000.2660.12519.28 s+0.08 s
SFT220.8051.0000.3380.10919.28 s+0.08 s
SFT230.8551.0000.2170.03119.28 s+0.08 s
SFT240.8451.0000.6190.04719.28 s+0.08 s
SFT2 + DPO210.7851.0000.5950.12519.28 s+0.08 s
SFT2 + DPO220.7621.0000.1920.14119.28 s+0.08 s
SFT2 + DPO230.7751.0000.1630.10919.28 s+0.08 s
SFT2 + DPO240.7941.0000.4870.03119.28 s+0.08 s
Hope Enthusiasm Optimismdescripted 16.5 s · 1 burst(s)
SFT2
emotion 0.159duration err 0.34 soverall 0.374 ± 0.117
SFT2 + DPO2
emotion 0.122duration err 0.10 soverall 0.477 ± 0.005
script and prompt
[0.7 seconds duration] Und, ähm, (low mumble, 0.1 seconds) [0.3 seconds pause] [5.2 seconds duration] er wurde ja noch ganz kurz für eine vierte Amtszeit wiedergewählt, aber ist dann sehr kurz in die vierte Amtszeit gestorben [0.6 seconds pause] [9.3 seconds duration] und noch während er gelebt hat, wurde das quasi schon beschränkt, dass danach Präsidenten nur noch auf zwei Amtszeiten gewählt werden dürfen, weil das dann denen dann doch zu viel war mit Roosevelt. [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense hope enthusiasm optimism; style: monologue, casual; good recording, quiet background; 16.4s, DE.
SCRIPT:
[0.7 seconds duration] Und, ähm, (low mumble, 0.1 seconds) [0.3 seconds pause] [5.2 seconds duration] er wurde ja noch ganz kurz für eine vierte Amtszeit wiedergewählt, aber ist dann sehr kurz in die vierte Amtszeit gestorben [0.6 seconds pause] [9.3 seconds duration] und noch während er gelebt hat, wurde das quasi schon beschränkt, dass danach Präsidenten nur noch auf zwei Amtszeiten gewählt werden dürfen, weil das dann denen dann doch zu viel war mit Roosevelt. [0.3 seconds pause]
- Tokens:
205
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.7 seconds duration] Und, ähm, (low mumble, 0.1 seconds) [0.3 seconds pause] [5.2 seconds duration] er wurde ja noch ganz kurz für eine vierte Amtszeit wiedergewählt, aber ist dann sehr kurz in die vierte Amtszeit gestorben [0.6 seconds pause] [9.3 seconds duration] und noch während er gelebt hat, wurde das quasi schon beschränkt, dass danach Präsidenten nur noch auf zwei Amtszeiten gewählt werden dürfen, weil das dann denen dann doch zu viel war mit Roosevelt. [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.2721.0000.6440.01816.40 s-0.10 s
SFT220.1220.5000.2360.05415.44 s-1.06 s
SFT230.1220.5050.5270.07116.40 s-0.10 s
SFT240.1221.0000.4870.08916.40 s-0.10 s
SFT2 + DPO210.1221.0000.4870.03616.40 s-0.10 s
SFT2 + DPO220.1221.0000.4680.01816.40 s-0.10 s
SFT2 + DPO230.1221.0000.4870.05416.40 s-0.10 s
SFT2 + DPO240.1221.0000.4870.05416.40 s-0.10 s
Hope Enthusiasm Optimismenscripted 13.4 s · 3 burst(s)
SFT2
emotion 0.765duration err 0.04 soverall 0.648 ± 0.133
SFT2 + DPO2
emotion 0.816duration err 0.04 soverall 0.739 ± 0.080
script and prompt
[0.2 seconds pause] [1.7 seconds duration] Yeah, Valentine's Day gift guide all under [1.2 seconds pause] (low mumble, 0.4 seconds) [3.8 seconds duration] $5. Someone tweeted at us the other day and was like, you know, that, uh, (low mumble, 0.2 seconds) [0.2 seconds pause] [2.0 seconds duration] gamer tech guide was really good. [1.3 seconds duration] You guys should make that into a series. [0.3 seconds duration] And I went, (effort grunt, 0.2 seconds) [0.1 seconds duration] ha! [1.2 seconds pause] [0.6 seconds duration] We already have.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, relaxed, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, slightly submissive, neutral openness; reads as intense hope enthusiasm optimism; style: casual, conversational; good recording, quiet background; mildly explicit content; 13.4s, EN.
SCRIPT:
[0.2 seconds pause] [1.7 seconds duration] Yeah, Valentine's Day gift guide all under [1.2 seconds pause] (low mumble, 0.4 seconds) [3.8 seconds duration] $5. Someone tweeted at us the other day and was like, you know, that, uh, (low mumble, 0.2 seconds) [0.2 seconds pause] [2.0 seconds duration] gamer tech guide was really good. [1.3 seconds duration] You guys should make that into a series. [0.3 seconds duration] And I went, (effort grunt, 0.2 seconds) [0.1 seconds duration] ha! [1.2 seconds pause] [0.6 seconds duration] We already have.
- Tokens:
167
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.2 seconds pause] [1.7 seconds duration] Yeah, Valentine's Day gift guide all under [1.2 seconds pause] (low mumble, 0.4 seconds) [3.8 seconds duration] $5. Someone tweeted at us the other day and was like, you know, that, uh, (low mumble, 0.2 seconds) [0.2 seconds pause] [2.0 seconds duration] gamer tech guide was really good. [1.3 seconds duration] You guys should make that into a series. [0.3 seconds duration] And I went, (effort grunt, 0.2 seconds) [0.1 seconds duration] ha! [1.2 seconds pause] [0.6 seconds duration] We already have.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6411.0000.5980.11613.36 s-0.04 s
SFT220.7501.0000.6610.07013.44 s+0.04 s
SFT230.7611.0000.4520.04713.44 s+0.04 s
SFT240.9091.0000.5930.09313.36 s-0.04 s
SFT2 + DPO210.7721.0000.4660.04713.36 s-0.04 s
SFT2 + DPO220.7931.0000.6760.04713.44 s+0.04 s
SFT2 + DPO230.8041.0000.5040.07013.36 s-0.04 s
SFT2 + DPO240.8971.0000.5850.04713.44 s+0.04 s
Impatience and Irritabilitydescripted 17.0 s · 0 burst(s)
SFT2
emotion 0.315duration err 0.04 soverall 0.352 ± 0.005
SFT2 + DPO2
emotion 0.301duration err 0.04 soverall 0.346 ± 0.004
script and prompt
[2.2 seconds duration] Genau, und dann wird's halt auch ein bisschen streitbarer, ne? [4.3 seconds duration] Also ich, so, so sehr ich dir zustimme im Bezug auf, ähm, (low mumble) [0.4 seconds pause] [6.8 seconds duration] Mensch, der, der Herr Ehlers hätte aber da doch hier nochmal ein bisschen mehr Fleisch an den Knochen bringen können im Bezug auf was meint er denn jetzt mit OEP. [0.9 seconds pause] [2.1 seconds duration] Trotzdem kann ich mich irgendwie sozusagen, [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense impatience and irritability; style: casual, conversational; good recording, quiet background; 17.0s, DE.
SCRIPT:
[2.2 seconds duration] Genau, und dann wird's halt auch ein bisschen streitbarer, ne? [4.3 seconds duration] Also ich, so, so sehr ich dir zustimme im Bezug auf, ähm, (low mumble) [0.4 seconds pause] [6.8 seconds duration] Mensch, der, der Herr Ehlers hätte aber da doch hier nochmal ein bisschen mehr Fleisch an den Knochen bringen können im Bezug auf was meint er denn jetzt mit OEP. [0.9 seconds pause] [2.1 seconds duration] Trotzdem kann ich mich irgendwie sozusagen, [0.3 seconds pause]
- Tokens:
213
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[2.2 seconds duration] Genau, und dann wird's halt auch ein bisschen streitbarer, ne? [4.3 seconds duration] Also ich, so, so sehr ich dir zustimme im Bezug auf, ähm, (low mumble) [0.4 seconds pause] [6.8 seconds duration] Mensch, der, der Herr Ehlers hätte aber da doch hier nochmal ein bisschen mehr Fleisch an den Knochen bringen können im Bezug auf was meint er denn jetzt mit OEP. [0.9 seconds pause] [2.1 seconds duration] Trotzdem kann ich mich irgendwie sozusagen, [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.3921.0000.0000.15517.04 s+0.04 s
SFT220.3461.0000.0000.12117.04 s+0.04 s
SFT230.2441.0000.0000.15517.04 s+0.04 s
SFT240.2761.0000.0000.12117.04 s+0.04 s
SFT2 + DPO210.1671.0000.0000.15517.04 s+0.04 s
SFT2 + DPO220.4341.0000.0000.17217.04 s+0.04 s
SFT2 + DPO230.2881.0000.0000.15517.04 s+0.04 s
SFT2 + DPO240.3161.0000.0000.13817.04 s+0.04 s
Impatience and Irritabilityenscripted 15.0 s · 1 burst(s)
SFT2
emotion 0.139duration err 0.08 soverall 0.460 ± 0.011
SFT2 + DPO2
emotion 0.134duration err 0.08 soverall 0.457 ± 0.011
script and prompt
[2.5 seconds duration] Yeah, and so then, that's why you see people go, uhm, (low mumble, 0.1 seconds) [0.3 seconds pause] [1.1 seconds duration] yeaah, I'm gonna go start my own studio. [0.3 seconds pause] [5.8 seconds duration] Or I'm gonna go do this because I wanna go make a thing that, this is why I fell in love with game development, and I don't wanna do it at a big place anymore. [0.7 seconds pause] [2.8 seconds duration] And, I mean, I would say, I feel like that's what happened with Valve. [0.6 seconds pause] [0.5 seconds duration] In some way. [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, slightly submissive, neutral openness; reads as intense impatience and irritability; style: conversational, casual; average recording, quiet background; 14.9s, EN.
SCRIPT:
[2.5 seconds duration] Yeah, and so then, that's why you see people go, uhm, (low mumble, 0.1 seconds) [0.3 seconds pause] [1.1 seconds duration] yeaah, I'm gonna go start my own studio. [0.3 seconds pause] [5.8 seconds duration] Or I'm gonna go do this because I wanna go make a thing that, this is why I fell in love with game development, and I don't wanna do it at a big place anymore. [0.7 seconds pause] [2.8 seconds duration] And, I mean, I would say, I feel like that's what happened with Valve. [0.6 seconds pause] [0.5 seconds duration] In some way. [0.3 seconds pause]
- Tokens:
186
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[2.5 seconds duration] Yeah, and so then, that's why you see people go, uhm, (low mumble, 0.1 seconds) [0.3 seconds pause] [1.1 seconds duration] yeaah, I'm gonna go start my own studio. [0.3 seconds pause] [5.8 seconds duration] Or I'm gonna go do this because I wanna go make a thing that, this is why I fell in love with game development, and I don't wanna do it at a big place anymore. [0.7 seconds pause] [2.8 seconds duration] And, I mean, I would say, I feel like that's what happened with Valve. [0.6 seconds pause] [0.5 seconds duration] In some way. [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0091.0000.5270.14114.88 s-0.12 s
SFT220.1401.0000.6440.14114.88 s-0.12 s
SFT230.0641.0000.6190.14114.96 s-0.04 s
SFT240.3421.0000.6980.14114.96 s-0.04 s
SFT2 + DPO210.1051.0000.5490.14114.88 s-0.12 s
SFT2 + DPO220.0701.0000.7560.16914.96 s-0.04 s
SFT2 + DPO230.1401.0000.6440.14114.96 s-0.04 s
SFT2 + DPO240.2201.0000.5490.14114.88 s-0.12 s
Infatuationdescripted 12.4 s · 1 burst(s)
SFT2
emotion 0.372duration err 0.08 soverall 0.322 ± 0.128
SFT2 + DPO2
emotion 0.182duration err 0.08 soverall 0.420 ± 0.021
script and prompt
[0.2 seconds pause] [5.5 seconds duration] Also, das zielt wie andere Konsensmethoden darauf ab, mehr Blöcke für die Kette zu generieren. [0.7 seconds pause] (contented sigh, 0.3 seconds) [1.7 seconds pause] [1.8 seconds duration] Im Grunde ist das der ganze Sinn. [2.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Moritz
- Instruction:
GENERAL: A voice very high in throat resonance. Throat resonance is maximally developed and overwhelmingly dominant, completely saturating the voice with deep, cavernous pharyngeal resonance. The acoustic output is almost entirely derived from throat vibration, creating a heavily warmed, deeply throaty vocal quality; reads as intense infatuation.
SCRIPT:
[0.2 seconds pause] [5.5 seconds duration] Also, das zielt wie andere Konsensmethoden darauf ab, mehr Blöcke für die Kette zu generieren. [0.7 seconds pause] (contented sigh, 0.3 seconds) [1.7 seconds pause] [1.8 seconds duration] Im Grunde ist das der ganze Sinn. [2.2 seconds pause]
- Tokens:
154
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.2 seconds pause] [5.5 seconds duration] Also, das zielt wie andere Konsensmethoden darauf ab, mehr Blöcke für die Kette zu generieren. [0.7 seconds pause] (contented sigh, 0.3 seconds) [1.7 seconds pause] [1.8 seconds duration] Im Grunde ist das der ganze Sinn. [2.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0641.0000.0000.00012.32 s-0.08 s
SFT220.5471.0000.2570.00012.32 s-0.08 s
SFT230.1521.0000.0002.50012.32 s-0.08 s
SFT240.7230.5140.0000.00012.32 s-0.08 s
SFT2 + DPO210.0641.0000.0000.00012.32 s-0.08 s
SFT2 + DPO220.4761.0000.2440.04512.32 s-0.08 s
SFT2 + DPO230.0371.0000.0000.00012.32 s-0.08 s
SFT2 + DPO240.1521.0000.2440.00012.32 s-0.08 s
Infatuationenscripted 17.1 s · 1 burst(s)
SFT2
emotion 0.226duration err 0.10 soverall 0.428 ± 0.034
SFT2 + DPO2
emotion 0.100duration err 0.06 soverall 0.409 ± 0.023
script and prompt
[1.7 seconds duration] Because they're not managed. [1.0 seconds duration] They're not regulated. [0.3 seconds pause] [1.0 seconds duration] And wolves are the same. [1.3 seconds duration] They don't care about tomorrow. [1.2 seconds duration] They don't think about, (contented sigh, 0.3 seconds) [2.2 seconds duration] okay, there's 10 elk out here in the field. [0.8 seconds pause] [1.8 seconds duration] I've got to eat for the next 10 weeks. [3.4 seconds duration] I need to conserve these elk and make sure I've got, you know, one a week. [1.9 seconds duration] They're going to go out there and kill all 10 of them if they can. [0.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly negative, slightly dominant, slightly guarded; reads as intense infatuation; style: casual, conversational; good recording, quiet background; 17.1s, EN.
SCRIPT:
[1.7 seconds duration] Because they're not managed. [1.0 seconds duration] They're not regulated. [0.3 seconds pause] [1.0 seconds duration] And wolves are the same. [1.3 seconds duration] They don't care about tomorrow. [1.2 seconds duration] They don't think about, (contented sigh, 0.3 seconds) [2.2 seconds duration] okay, there's 10 elk out here in the field. [0.8 seconds pause] [1.8 seconds duration] I've got to eat for the next 10 weeks. [3.4 seconds duration] I need to conserve these elk and make sure I've got, you know, one a week. [1.9 seconds duration] They're going to go out there and kill all 10 of them if they can. [0.2 seconds pause]
- Tokens:
213
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[1.7 seconds duration] Because they're not managed. [1.0 seconds duration] They're not regulated. [0.3 seconds pause] [1.0 seconds duration] And wolves are the same. [1.3 seconds duration] They don't care about tomorrow. [1.2 seconds duration] They don't think about, (contented sigh, 0.3 seconds) [2.2 seconds duration] okay, there's 10 elk out here in the field. [0.8 seconds pause] [1.8 seconds duration] I've got to eat for the next 10 weeks. [3.4 seconds duration] I need to conserve these elk and make sure I've got, you know, one a week. [1.9 seconds duration] They're going to go out there and kill all 10 of them if they can. [0.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.3491.0000.3360.02917.04 s-0.06 s
SFT220.2061.0000.3360.00016.96 s-0.14 s
SFT230.2711.0000.0000.00016.96 s-0.14 s
SFT240.0771.0000.0000.02917.04 s-0.06 s
SFT2 + DPO210.1521.0000.0000.00017.04 s-0.06 s
SFT2 + DPO220.0121.0000.0000.00017.04 s-0.06 s
SFT2 + DPO230.1091.0000.0000.02917.04 s-0.06 s
SFT2 + DPO240.1291.0000.2440.00017.04 s-0.06 s
Interestdescripted 15.8 s · 0 burst(s)
SFT2
emotion 0.641duration err 0.06 soverall 0.402 ± 0.042
SFT2 + DPO2
emotion 0.620duration err 0.04 soverall 0.498 ± 0.181
script and prompt
[7.4 seconds duration] Ihr habt es gefühlt, Pfannenstil, daß die zweite Hälfte der Odyssee von besonderer Schönheit und Größe ist. [0.8 seconds pause] [1.3 seconds duration] Wie, der (low mumble) [5.0 seconds duration] Heimgekehrte wird als ein fahrender Bettler an seinem eigenen Herde misshandelt. [0.6 seconds pause] [0.3 seconds duration] Wie, [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An elderly masculine voice; delivery is very low-energy, slow, relaxed, fairly steady; timbre is neutral-toned, slightly dark, rough, thin; slurred, frequent disfluency, fairly narrow pitch, audible breath; affect is neutral, neutral stance, slightly guarded; reads as intense interest; style: monologue, whispered; below-average recording, quiet background; 15.8s, DE.
SCRIPT:
[7.4 seconds duration] Ihr habt es gefühlt, Pfannenstil, daß die zweite Hälfte der Odyssee von besonderer Schönheit und Größe ist. [0.8 seconds pause] [1.3 seconds duration] Wie, der (low mumble) [5.0 seconds duration] Heimgekehrte wird als ein fahrender Bettler an seinem eigenen Herde misshandelt. [0.6 seconds pause] [0.3 seconds duration] Wie, [0.4 seconds pause]
- Tokens:
198
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[7.4 seconds duration] Ihr habt es gefühlt, Pfannenstil, daß die zweite Hälfte der Odyssee von besonderer Schönheit und Größe ist. [0.8 seconds pause] [1.3 seconds duration] Wie, der (low mumble) [5.0 seconds duration] Heimgekehrte wird als ein fahrender Bettler an seinem eigenen Herde misshandelt. [0.6 seconds pause] [0.3 seconds duration] Wie, [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5350.6731.0000.09715.84 s+0.04 s
SFT220.8270.0081.0000.06515.84 s+0.04 s
SFT230.5690.5091.0000.12915.84 s+0.04 s
SFT240.6330.5001.0000.09715.92 s+0.12 s
SFT2 + DPO210.6330.5001.0000.06515.84 s+0.04 s
SFT2 + DPO220.8111.0001.0000.06515.84 s+0.04 s
SFT2 + DPO230.3720.5051.0000.12915.84 s+0.04 s
SFT2 + DPO240.6640.6581.0000.06515.84 s+0.04 s
Interestenscripted 21.4 s · 2 burst(s)
SFT2
emotion 0.472duration err 0.22 soverall 0.295 ± 0.167
SFT2 + DPO2
emotion 0.545duration err 0.16 soverall 0.187 ± 0.030
script and prompt
[5.4 seconds pause] [0.6 seconds duration] In [3.0 seconds pause] (ahem, 0.4 seconds) [1.1 seconds pause] [7.1 seconds duration] adult populations, a frequently encountered complaint of an unpleasant odor can often be traced back to the manifestation of specific underlying [1.6 seconds pause] (low mumble, 0.4 seconds) [0.5 seconds pause] [0.6 seconds duration] pathologies. [0.7 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Juliet
- Instruction:
GENERAL: A voice very high in teacher/didactic style. The delivery is ultra-pedagogical and intensely teaching-like, sounding exactly like a master teacher with maximum instructional affectation. The speaker maintains extreme pedagogical pacing and clarity, uses highly structured, educational phrasing throughout; reads as intense interest.
SCRIPT:
[5.4 seconds pause] [0.6 seconds duration] In [3.0 seconds pause] (ahem, 0.4 seconds) [1.1 seconds pause] [7.1 seconds duration] adult populations, a frequently encountered complaint of an unpleasant odor can often be traced back to the manifestation of specific underlying [1.6 seconds pause] (low mumble, 0.4 seconds) [0.5 seconds pause] [0.6 seconds duration] pathologies. [0.7 seconds pause]
- Tokens:
265
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[5.4 seconds pause] [0.6 seconds duration] In [3.0 seconds pause] (ahem, 0.4 seconds) [1.1 seconds pause] [7.1 seconds duration] adult populations, a frequently encountered complaint of an unpleasant odor can often be traced back to the manifestation of specific underlying [1.6 seconds pause] (low mumble, 0.4 seconds) [0.5 seconds pause] [0.6 seconds duration] pathologies. [0.7 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.4031.0000.0000.87021.20 s-0.20 s
SFT220.7491.0000.2520.04321.12 s-0.28 s
SFT230.5601.0000.1351.04321.20 s-0.20 s
SFT240.1750.5000.1990.39121.20 s-0.20 s
SFT2 + DPO210.7800.5130.2481.52221.20 s-0.20 s
SFT2 + DPO220.5701.0000.0001.26121.28 s-0.12 s
SFT2 + DPO230.2291.0000.0001.60921.36 s-0.04 s
SFT2 + DPO240.5991.0000.1400.82621.12 s-0.28 s
Intoxication Altered States of Consciousnessdescripted 17.1 s · 1 burst(s)
SFT2
emotion 0.641duration err 0.14 soverall 0.599 ± 0.047
SFT2 + DPO2
emotion 0.619duration err 0.14 soverall 0.598 ± 0.203
script and prompt
[2.3 seconds duration] Also, das ist jetzt halt auch wieder so eine Sache. [1.0 seconds duration] Jetzt bin ich noch in Europa. [3.2 seconds duration] Jetzt weiß ich, okay, jetzt kommt aus jeder Steckdose auch was raus, so, ne. [1.0 seconds pause] [1.3 seconds duration] Das ist mir auch bewusst, so. [1.1 seconds duration] Ja, hier weiß ich das. (low mumble, 0.2 seconds) [0.2 seconds pause] [3.8 seconds duration] Woanders weiß ich nicht, aber deswegen mach ich, deswegen ist es ja auch ein Abenteuer, so, ne. [2.7 seconds duration] Deswegen, na, deswegen muss man ja auch gucken. [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense intoxication altered states of consciousness; style: casual, monologue; average recording, quiet background; 17.1s, DE.
SCRIPT:
[2.3 seconds duration] Also, das ist jetzt halt auch wieder so eine Sache. [1.0 seconds duration] Jetzt bin ich noch in Europa. [3.2 seconds duration] Jetzt weiß ich, okay, jetzt kommt aus jeder Steckdose auch was raus, so, ne. [1.0 seconds pause] [1.3 seconds duration] Das ist mir auch bewusst, so. [1.1 seconds duration] Ja, hier weiß ich das. (low mumble, 0.2 seconds) [0.2 seconds pause] [3.8 seconds duration] Woanders weiß ich nicht, aber deswegen mach ich, deswegen ist es ja auch ein Abenteuer, so, ne. [2.7 seconds duration] Deswegen, na, deswegen muss man ja auch gucken. [0.3 seconds pause]
- Tokens:
213
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[2.3 seconds duration] Also, das ist jetzt halt auch wieder so eine Sache. [1.0 seconds duration] Jetzt bin ich noch in Europa. [3.2 seconds duration] Jetzt weiß ich, okay, jetzt kommt aus jeder Steckdose auch was raus, so, ne. [1.0 seconds pause] [1.3 seconds duration] Das ist mir auch bewusst, so. [1.1 seconds duration] Ja, hier weiß ich das. (low mumble, 0.2 seconds) [0.2 seconds pause] [3.8 seconds duration] Woanders weiß ich nicht, aber deswegen mach ich, deswegen ist es ja auch ein Abenteuer, so, ne. [2.7 seconds duration] Deswegen, na, deswegen muss man ja auch gucken. [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.2331.0000.9610.13616.96 s-0.14 s
SFT220.7811.0000.6410.13616.96 s-0.14 s
SFT230.8361.0000.0000.13616.96 s-0.14 s
SFT240.7131.0000.9230.12116.96 s-0.14 s
SFT2 + DPO210.7691.0000.9610.15216.96 s-0.14 s
SFT2 + DPO220.9171.0000.9610.12116.96 s-0.14 s
SFT2 + DPO230.5751.0000.6700.12116.96 s-0.14 s
SFT2 + DPO240.2141.0000.0000.13616.96 s-0.14 s
Intoxication Altered States of Consciousnessenscripted 5.6 s · 1 burst(s)
SFT2
emotion 0.484duration err 0.08 soverall 0.331 ± 0.073
SFT2 + DPO2
emotion 0.530duration err 0.08 soverall 0.251 ± 0.040
script and prompt
[0.4 seconds pause] [2.7 seconds duration] There it was, that small spot in the dirt she'd been digging. [0.5 seconds pause] (sharp inhale, 0.4 seconds) [1.4 seconds duration] Exactly what she'd been searching for. [0.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Michael
- Instruction:
GENERAL: A voice extremely low in mixed resonance. The resonance is entirely one-dimensional and heavily isolated to a single cavity, creating a highly imbalanced and acoustically disconnected sound. One resonance source completely dominates while all others are suppressed, resulting in a sonically lopsided profile that lacks integration; reads as intense intoxication altered states of consciousness.
SCRIPT:
[0.4 seconds pause] [2.7 seconds duration] There it was, that small spot in the dirt she'd been digging. [0.5 seconds pause] (sharp inhale, 0.4 seconds) [1.4 seconds duration] Exactly what she'd been searching for. [0.2 seconds pause]
- Tokens:
69
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.4 seconds pause] [2.7 seconds duration] There it was, that small spot in the dirt she'd been digging. [0.5 seconds pause] (sharp inhale, 0.4 seconds) [1.4 seconds duration] Exactly what she'd been searching for. [0.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5991.0000.0000.0005.52 s-0.08 s
SFT220.2700.5770.0000.0005.52 s-0.08 s
SFT230.6370.7310.0000.0005.52 s-0.08 s
SFT240.4291.0000.0000.0005.52 s-0.08 s
SFT2 + DPO210.2871.0000.0000.2785.52 s-0.08 s
SFT2 + DPO220.5360.5900.0000.0005.52 s-0.08 s
SFT2 + DPO230.6200.5030.0000.0005.52 s-0.08 s
SFT2 + DPO240.6800.5480.0000.0565.52 s-0.08 s
Jealousy and Envydescripted 22.4 s · 1 burst(s)
SFT2
emotion 0.004duration err 0.12 soverall 0.400 ± 0.015
SFT2 + DPO2
emotion 0.004duration err 0.08 soverall 0.406 ± 0.006
script and prompt
(low mumble, 0.1 seconds) [0.7 seconds duration] euhm, [1.1 seconds duration] ja, nee, das ist total nett. [0.2 seconds duration] Matthias (ahem) [0.8 seconds duration] Meyer von Mensch. [1.7 seconds pause] [11.4 seconds duration] Oink, oink, klatschen, äh, genau, nee, das ist total netter, und, äh, da fällt mir ein, auch, äh, kann man auch durchaus erwähnen, da waren wir nämlich heute nicht, die haben einen sehr schönen, sehr großen, schön begrünten Garten auch, [0.6 seconds pause] [5.8 seconds duration] wo man im Sommer wunderbar hin kann, auch mit, ich glaube, ein Wasserspiel an irgendeiner Stelle auch, total nett mit so einem
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, relaxed, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, frequent disfluency, wide pitch range, light breath; affect is mildly positive, slightly submissive, neutral openness; reads as intense jealousy and envy; style: conversational, casual; good recording, quiet background; 22.5s, DE.
SCRIPT:
(low mumble, 0.1 seconds) [0.7 seconds duration] euhm, [1.1 seconds duration] ja, nee, das ist total nett. [0.2 seconds duration] Matthias (ahem) [0.8 seconds duration] Meyer von Mensch. [1.7 seconds pause] [11.4 seconds duration] Oink, oink, klatschen, äh, genau, nee, das ist total netter, und, äh, da fällt mir ein, auch, äh, kann man auch durchaus erwähnen, da waren wir nämlich heute nicht, die haben einen sehr schönen, sehr großen, schön begrünten Garten auch, [0.6 seconds pause] [5.8 seconds duration] wo man im Sommer wunderbar hin kann, auch mit, ich glaube, ein Wasserspiel an irgendeiner Stelle auch, total nett mit so einem
- Tokens:
281
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(low mumble, 0.1 seconds) [0.7 seconds duration] euhm, [1.1 seconds duration] ja, nee, das ist total nett. [0.2 seconds duration] Matthias (ahem) [0.8 seconds duration] Meyer von Mensch. [1.7 seconds pause] [11.4 seconds duration] Oink, oink, klatschen, äh, genau, nee, das ist total netter, und, äh, da fällt mir ein, auch, äh, kann man auch durchaus erwähnen, da waren wir nämlich heute nicht, die haben einen sehr schönen, sehr großen, schön begrünten Garten auch, [0.6 seconds pause] [5.8 seconds duration] wo man im Sommer wunderbar hin kann, auch mit, ich glaube, ein Wasserspiel an irgendeiner Stelle auch, total nett mit so einem
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0041.0000.3020.11022.56 s+0.16 s
SFT220.0041.0000.2580.11022.56 s+0.16 s
SFT230.0041.0000.1900.08222.48 s+0.08 s
SFT240.0041.0000.1060.12322.48 s+0.08 s
SFT2 + DPO210.0041.0000.2580.11022.48 s+0.08 s
SFT2 + DPO220.0041.0000.3020.11022.48 s+0.08 s
SFT2 + DPO230.0041.0000.2380.12322.48 s+0.08 s
SFT2 + DPO240.0041.0000.2480.11022.48 s+0.08 s
Jealousy and Envyenscripted 11.2 s · 1 burst(s)
SFT2
emotion 0.004duration err 0.24 soverall 0.483 ± 0.025
SFT2 + DPO2
emotion 0.004duration err 0.12 soverall 0.449 ± 0.026
script and prompt
[1.1 seconds duration] Yeah, let's talk about it. (low mumble, 0.2 seconds) [0.2 seconds duration] Uh, [0.7 seconds pause] [5.4 seconds duration] also, can you guys please do me a favor and talk about the most interesting thing right off the top so we can name the podcast something appropriate? [0.7 seconds duration] Sure. [0.3 seconds duration] Geez. [0.9 seconds duration] What are we gonna talk about? [1.0 seconds duration] What's the most interesting thing? [0.7 seconds duration] I got a haircut.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is energised, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is positive, slightly dominant, slightly guarded; reads as intense jealousy and envy; style: casual, conversational; average recording, some background noise; 11.0s, EN.
SCRIPT:
[1.1 seconds duration] Yeah, let's talk about it. (low mumble, 0.2 seconds) [0.2 seconds duration] Uh, [0.7 seconds pause] [5.4 seconds duration] also, can you guys please do me a favor and talk about the most interesting thing right off the top so we can name the podcast something appropriate? [0.7 seconds duration] Sure. [0.3 seconds duration] Geez. [0.9 seconds duration] What are we gonna talk about? [1.0 seconds duration] What's the most interesting thing? [0.7 seconds duration] I got a haircut.
- Tokens:
137
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[1.1 seconds duration] Yeah, let's talk about it. (low mumble, 0.2 seconds) [0.2 seconds duration] Uh, [0.7 seconds pause] [5.4 seconds duration] also, can you guys please do me a favor and talk about the most interesting thing right off the top so we can name the podcast something appropriate? [0.7 seconds duration] Sure. [0.3 seconds duration] Geez. [0.9 seconds duration] What are we gonna talk about? [1.0 seconds duration] What's the most interesting thing? [0.7 seconds duration] I got a haircut.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0041.0000.5950.05910.96 s-0.24 s
SFT220.0041.0000.7560.05910.96 s-0.24 s
SFT230.0041.0000.3990.05910.96 s-0.24 s
SFT240.0041.0000.5270.07810.96 s-0.24 s
SFT2 + DPO210.0041.0000.5490.09811.04 s-0.16 s
SFT2 + DPO220.0041.0000.7260.13711.28 s+0.08 s
SFT2 + DPO230.0041.0000.5070.13711.20 s+0.00 s
SFT2 + DPO240.0041.0000.1700.05910.96 s-0.24 s
Longingdescripted 14.1 s · 0 burst(s)
SFT2
emotion 0.329duration err 0.14 soverall 0.343 ± 0.070
SFT2 + DPO2
emotion 0.296duration err 0.14 soverall 0.383 ± 0.008
script and prompt
[1.1 seconds duration] Also, was ist meine Meinung dazu? [0.3 seconds pause] [4.5 seconds duration] Ich finde, dass dieses piece mehr growth eigentlich klar sein sollte. [0.8 seconds pause] [1.7 seconds duration] Wir gehen halt in Iterationen vor. [0.3 seconds pause] [3.4 seconds duration] Also, ich hoffe, dass sich das in den letzten, jetzt fast 25 Jahren, [2.0 seconds duration] ähm, durchgesetzt hat. (low mumble)
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, frequent disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense longing; style: monologue, casual; average recording, quiet background; 14.3s, DE.
SCRIPT:
[1.1 seconds duration] Also, was ist meine Meinung dazu? [0.3 seconds pause] [4.5 seconds duration] Ich finde, dass dieses piece mehr growth eigentlich klar sein sollte. [0.8 seconds pause] [1.7 seconds duration] Wir gehen halt in Iterationen vor. [0.3 seconds pause] [3.4 seconds duration] Also, ich hoffe, dass sich das in den letzten, jetzt fast 25 Jahren, [2.0 seconds duration] ähm, durchgesetzt hat. (low mumble)
- Tokens:
178
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[1.1 seconds duration] Also, was ist meine Meinung dazu? [0.3 seconds pause] [4.5 seconds duration] Ich finde, dass dieses piece mehr growth eigentlich klar sein sollte. [0.8 seconds pause] [1.7 seconds duration] Wir gehen halt in Iterationen vor. [0.3 seconds pause] [3.4 seconds duration] Also, ich hoffe, dass sich das in den letzten, jetzt fast 25 Jahren, [2.0 seconds duration] ähm, durchgesetzt hat. (low mumble)
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0661.0000.0000.05114.24 s+0.14 s
SFT220.2861.0000.0000.05114.24 s+0.14 s
SFT230.5441.0000.0000.02614.24 s+0.14 s
SFT240.4210.5700.0000.02614.24 s+0.14 s
SFT2 + DPO210.0971.0000.0000.02614.24 s+0.14 s
SFT2 + DPO220.5291.0000.0000.07714.24 s+0.14 s
SFT2 + DPO230.2981.0000.0000.02614.24 s+0.14 s
SFT2 + DPO240.2581.0000.0000.05114.24 s+0.14 s
Longingenscripted 13.8 s · 1 burst(s)
SFT2
emotion 0.164duration err 0.08 soverall 0.357 ± 0.052
SFT2 + DPO2
emotion 0.120duration err 0.06 soverall 0.196 ± 0.033
script and prompt
[5.9 seconds duration] Think of remarketing: we gather user data, building specific lists, which lets us then reach out with targeted sales emails. [0.3 seconds pause] [2.6 seconds duration] It's how we make sure the right message gets to the right people. [2.0 seconds pause] (surprised gasp, 0.7 seconds) [2.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Josephine
- Instruction:
GENERAL: A voice notably high in pitch range. The pitch moves expressively across a wide range of notes, actively utilizing significant fundamental frequency highs and lows to project personality. The voice sounds lively, colorful, and deeply engaged, creating a dynamic and engaging melodic map; reads as intense longing.
SCRIPT:
[5.9 seconds duration] Think of remarketing: we gather user data, building specific lists, which lets us then reach out with targeted sales emails. [0.3 seconds pause] [2.6 seconds duration] It's how we make sure the right message gets to the right people. [2.0 seconds pause] (surprised gasp, 0.7 seconds) [2.3 seconds pause]
- Tokens:
173
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[5.9 seconds duration] Think of remarketing: we gather user data, building specific lists, which lets us then reach out with targeted sales emails. [0.3 seconds pause] [2.6 seconds duration] It's how we make sure the right message gets to the right people. [2.0 seconds pause] (surprised gasp, 0.7 seconds) [2.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0761.0000.1240.06113.84 s+0.04 s
SFT220.2581.0000.2650.09113.92 s+0.12 s
SFT230.3190.6890.2250.00013.92 s+0.12 s
SFT240.0020.5880.4150.09113.76 s-0.04 s
SFT2 + DPO210.0660.5190.0000.03013.84 s+0.04 s
SFT2 + DPO220.0250.5010.0000.03013.92 s+0.12 s
SFT2 + DPO230.3300.5440.1510.03013.84 s+0.04 s
SFT2 + DPO240.0570.3680.0000.00013.84 s+0.04 s
Malevolence Malicedescripted 18.2 s · 2 burst(s)
SFT2
emotion 0.458duration err 0.04 soverall 0.311 ± 0.100
SFT2 + DPO2
emotion 0.339duration err 0.04 soverall 0.257 ± 0.048
script and prompt
[1.7 seconds duration] Und es hier eben auch, äh, (ahem) [8.8 seconds duration] die dementsprechenden Analysen braucht und Gegenstrategien und grad auch Feminismus so krass angegriffen wird und, (low mumble, 0.1 seconds) [2.0 seconds duration] ähm, hier einfach ganz, (low mumble, 0.1 seconds) [5.1 seconds duration] äh, wichtige BündnispartnerInnen und Allianzen wichtig und notwendig sind. [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult somewhat feminine voice; delivery is subdued, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, neutral openness; reads as intense malevolence malice; style: monologue, casual; average recording, quiet background; 18.3s, DE.
SCRIPT:
[1.7 seconds duration] Und es hier eben auch, äh, (ahem) [8.8 seconds duration] die dementsprechenden Analysen braucht und Gegenstrategien und grad auch Feminismus so krass angegriffen wird und, (low mumble, 0.1 seconds) [2.0 seconds duration] ähm, hier einfach ganz, (low mumble, 0.1 seconds) [5.1 seconds duration] äh, wichtige BündnispartnerInnen und Allianzen wichtig und notwendig sind. [0.4 seconds pause]
- Tokens:
228
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[1.7 seconds duration] Und es hier eben auch, äh, (ahem) [8.8 seconds duration] die dementsprechenden Analysen braucht und Gegenstrategien und grad auch Feminismus so krass angegriffen wird und, (low mumble, 0.1 seconds) [2.0 seconds duration] ähm, hier einfach ganz, (low mumble, 0.1 seconds) [5.1 seconds duration] äh, wichtige BündnispartnerInnen und Allianzen wichtig und notwendig sind. [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6131.0000.5050.14718.24 s+0.04 s
SFT220.2470.5000.2580.11818.24 s+0.04 s
SFT230.5310.5210.1100.14718.24 s+0.04 s
SFT240.4401.0000.1770.14718.16 s-0.04 s
SFT2 + DPO210.0400.5020.3580.11818.16 s-0.04 s
SFT2 + DPO220.5460.5070.3130.11818.24 s+0.04 s
SFT2 + DPO230.0630.5060.1900.14718.24 s+0.04 s
SFT2 + DPO240.7050.5000.4490.11818.24 s+0.04 s
Malevolence Maliceenscripted 11.3 s · 3 burst(s)
SFT2
emotion 0.152duration err 0.06 soverall 0.457 ± 0.034
SFT2 + DPO2
emotion 0.553duration err 0.05 soverall 0.408 ± 0.043
script and prompt
(low mumble, 0.3 seconds) [0.6 seconds pause] [1.9 seconds duration] Please just look at me for a moment. [0.6 seconds pause] (contented sigh, 0.3 seconds) [0.9 seconds pause] [4.3 seconds duration] I need someone to see me, to really see me, you know? [0.5 seconds pause] (contented sigh, 0.1 seconds) [1.8 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Kilian
- Instruction:
GENERAL: A single isolated slurping noises, produced on its own rather than inside speech, involuntary and natural; reads as intense malevolence malice.
SCRIPT:
(low mumble, 0.3 seconds) [0.6 seconds pause] [1.9 seconds duration] Please just look at me for a moment. [0.6 seconds pause] (contented sigh, 0.3 seconds) [0.9 seconds pause] [4.3 seconds duration] I need someone to see me, to really see me, you know? [0.5 seconds pause] (contented sigh, 0.1 seconds) [1.8 seconds pause]
- Tokens:
142
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
(low mumble, 0.3 seconds) [0.6 seconds pause] [1.9 seconds duration] Please just look at me for a moment. [0.6 seconds pause] (contented sigh, 0.3 seconds) [0.9 seconds pause] [4.3 seconds duration] I need someone to see me, to really see me, you know? [0.5 seconds pause] (contented sigh, 0.1 seconds) [1.8 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1081.0000.4800.00011.36 s+0.06 s
SFT220.0321.0000.1270.05011.36 s+0.06 s
SFT230.0961.0000.3840.05011.36 s+0.06 s
SFT240.3741.0000.4860.05011.36 s+0.06 s
SFT2 + DPO210.6661.0000.2420.00011.36 s+0.06 s
SFT2 + DPO220.4591.0000.0000.00011.36 s+0.06 s
SFT2 + DPO230.8411.0000.0001.00011.28 s-0.02 s
SFT2 + DPO240.2471.0000.2290.10011.36 s+0.06 s
Paindescripted 13.6 s · 2 burst(s)
SFT2
emotion 0.937duration err 0.14 soverall 0.691 ± 0.111
SFT2 + DPO2
emotion 0.950duration err 0.14 soverall 0.805 ± 0.016
script and prompt
[7.3 seconds duration] Selbst wenn es schrecklich war, selbst wenn es so weh tat, gab es in dieser toxischen Liebe immer noch etwas Echtes. [0.6 seconds pause] (contented sigh, 0.2 seconds) [0.7 seconds pause] (contented sigh, 0.2 seconds) [1.3 seconds pause] [3.3 seconds duration] Du musst nur diesen Teil sehen, dieses winzige bisschen davon.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Torsten
- Instruction:
GENERAL: A voice extremely low in whisper-talk style. The delivery is intensely loud and completely non-whispered, entirely rejecting any whispered or breath-based element. The acoustic profile shows zero whisper characteristics or breathy quality, sounding harsh and utterly non-whispered; reads as intense pain.
SCRIPT:
[7.3 seconds duration] Selbst wenn es schrecklich war, selbst wenn es so weh tat, gab es in dieser toxischen Liebe immer noch etwas Echtes. [0.6 seconds pause] (contented sigh, 0.2 seconds) [0.7 seconds pause] (contented sigh, 0.2 seconds) [1.3 seconds pause] [3.3 seconds duration] Du musst nur diesen Teil sehen, dieses winzige bisschen davon.
- Tokens:
168
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[7.3 seconds duration] Selbst wenn es schrecklich war, selbst wenn es so weh tat, gab es in dieser toxischen Liebe immer noch etwas Echtes. [0.6 seconds pause] (contented sigh, 0.2 seconds) [0.7 seconds pause] (contented sigh, 0.2 seconds) [1.3 seconds pause] [3.3 seconds duration] Du musst nur diesen Teil sehen, dieses winzige bisschen davon.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.8820.6030.0000.00013.44 s-0.16 s
SFT220.9850.8440.0000.32313.52 s-0.08 s
SFT230.9671.0000.0000.00013.44 s-0.16 s
SFT240.9131.0000.0000.00013.44 s-0.16 s
SFT2 + DPO210.9780.9900.0000.00013.44 s-0.16 s
SFT2 + DPO220.9731.0000.0000.00013.44 s-0.16 s
SFT2 + DPO230.8641.0000.4470.00013.44 s-0.16 s
SFT2 + DPO240.9851.0000.0000.00013.52 s-0.08 s
Painenscripted 8.6 s · 1 burst(s)
SFT2
emotion 0.903duration err 0.24 soverall 0.806 ± 0.120
SFT2 + DPO2
emotion 0.940duration err 0.20 soverall 0.942 ± 0.023
script and prompt
[0.4 seconds pause] (contented sigh, 0.2 seconds) [0.7 seconds pause] [4.2 seconds duration] I genuinely believed it was just a rough patch, a period of deep sadness, for years. [0.7 seconds pause] [2.0 seconds duration] Some sort of low mood, I kept telling myself. [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Renate
- Instruction:
GENERAL: A voice somewhat low in arousal. The speaker is entirely awake but is consciously holding back their physiological energy, maintaining a highly reserved and strictly controlled vocal output. The intensity of the acoustic signal is deliberately suppressed, typical of a cautious, highly formal, or hyper-professional setting; reads as intense pain.
SCRIPT:
[0.4 seconds pause] (contented sigh, 0.2 seconds) [0.7 seconds pause] [4.2 seconds duration] I genuinely believed it was just a rough patch, a period of deep sadness, for years. [0.7 seconds pause] [2.0 seconds duration] Some sort of low mood, I kept telling myself. [0.4 seconds pause]
- Tokens:
110
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.4 seconds pause] (contented sigh, 0.2 seconds) [0.7 seconds pause] [4.2 seconds duration] I genuinely believed it was just a rough patch, a period of deep sadness, for years. [0.7 seconds pause] [2.0 seconds duration] Some sort of low mood, I kept telling myself. [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.8741.0000.2870.3608.96 s+0.36 s
SFT220.8771.0000.5490.0008.80 s+0.20 s
SFT230.9461.0000.4320.0008.80 s+0.20 s
SFT240.9171.0000.3400.0008.80 s+0.20 s
SFT2 + DPO210.9031.0000.6980.0008.80 s+0.20 s
SFT2 + DPO220.9680.9290.6700.0008.80 s+0.20 s
SFT2 + DPO230.9481.0000.8190.0008.80 s+0.20 s
SFT2 + DPO240.9431.0000.7870.0008.80 s+0.20 s
Pleasure Ecstasydescripted 12.3 s · 1 burst(s)
SFT2
emotion 0.317duration err 0.08 soverall 0.237 ± 0.108
SFT2 + DPO2
emotion 0.288duration err 0.10 soverall 0.323 ± 0.131
script and prompt
[6.7 seconds duration] Lay-Wetten oder so ein paar andere Begriffe sind diese clevere Art, wie Leute kostenlose Wetten und diese süßen Buchmacheranreize schnappen. [0.5 seconds pause] [3.6 seconds duration] Das ist eine echte Technik, die Leute nutzen, um von diesen Aktionen zu profitieren. [0.7 seconds pause] (low mumble, 0.8 seconds)
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Brandon
- Instruction:
GENERAL: A voice notably high in storytelling style. The delivery is distinctly narrative and storytelling-oriented, featuring clear plot structure, storytelling cadence, and a clearly narrative-focused, entertaining register. The speaker sounds engaged with storytelling, naturally engaging; reads as intense pleasure ecstasy.
SCRIPT:
[6.7 seconds duration] Lay-Wetten oder so ein paar andere Begriffe sind diese clevere Art, wie Leute kostenlose Wetten und diese süßen Buchmacheranreize schnappen. [0.5 seconds pause] [3.6 seconds duration] Das ist eine echte Technik, die Leute nutzen, um von diesen Aktionen zu profitieren. [0.7 seconds pause] (low mumble, 0.8 seconds)
- Tokens:
154
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[6.7 seconds duration] Lay-Wetten oder so ein paar andere Begriffe sind diese clevere Art, wie Leute kostenlose Wetten und diese süßen Buchmacheranreize schnappen. [0.5 seconds pause] [3.6 seconds duration] Das ist eine echte Technik, die Leute nutzen, um von diesen Aktionen zu profitieren. [0.7 seconds pause] (low mumble, 0.8 seconds)
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.2880.5570.0000.11412.40 s+0.10 s
SFT220.2880.5260.0001.00012.48 s+0.18 s
SFT230.2880.6310.0000.05712.32 s+0.02 s
SFT240.4071.0000.1390.05712.32 s+0.02 s
SFT2 + DPO210.2880.5190.0001.00012.40 s+0.10 s
SFT2 + DPO220.2881.0000.0000.11412.40 s+0.10 s
SFT2 + DPO230.2881.0000.0000.05712.40 s+0.10 s
SFT2 + DPO240.2881.0000.4490.08612.40 s+0.10 s
Pleasure Ecstasyenscripted 8.7 s · 1 burst(s)
SFT2
emotion 0.778duration err 0.10 soverall 0.570 ± 0.253
SFT2 + DPO2
emotion 0.800duration err 0.10 soverall 0.499 ± 0.168
script and prompt
(breathy giggle, 0.1 seconds) [0.2 seconds duration] These [5.7 seconds duration] guidelines are a framework, and I'm proud to hand you the power to sharpen them or ease them where you see fit. [0.7 seconds pause] [2.0 seconds duration] That's the beauty of having the final say.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Sam
- Instruction:
GENERAL: A voice overwhelmingly expressing pride, a proud, self-assured voice, chest swelling with pride, triumphantly boastful, impossible to hide. The feeling is fought down rather than shown: it is held in and only leaks through at the edges of phrases; reads as intense pleasure ecstasy.
SCRIPT:
(breathy giggle, 0.1 seconds) [0.2 seconds duration] These [5.7 seconds duration] guidelines are a framework, and I'm proud to hand you the power to sharpen them or ease them where you see fit. [0.7 seconds pause] [2.0 seconds duration] That's the beauty of having the final say.
- Tokens:
110
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
(breathy giggle, 0.1 seconds) [0.2 seconds duration] These [5.7 seconds duration] guidelines are a framework, and I'm proud to hand you the power to sharpen them or ease them where you see fit. [0.7 seconds pause] [2.0 seconds duration] That's the beauty of having the final say.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7471.0000.2540.2588.80 s+0.10 s
SFT220.9551.0000.5710.0008.80 s+0.10 s
SFT230.5410.5530.0000.0008.80 s+0.10 s
SFT240.8700.7820.0000.0008.80 s+0.10 s
SFT2 + DPO210.9170.7070.0000.0008.80 s+0.10 s
SFT2 + DPO220.8010.5530.0000.0008.80 s+0.10 s
SFT2 + DPO230.6610.5630.0000.0008.80 s+0.10 s
SFT2 + DPO240.8200.6800.3230.0008.80 s+0.10 s
Pridedescripted 10.3 s · 1 burst(s)
SFT2
emotion 0.559duration err 0.14 soverall 0.430 ± 0.030
SFT2 + DPO2
emotion 0.427duration err 0.14 soverall 0.358 ± 0.107
script and prompt
[3.3 seconds duration] ihre Identität, nämlich des Türkentums der türkischen Sprache. [0.7 seconds pause] [0.3 seconds duration] Und (surprised gasp, 0.2 seconds) [5.4 seconds duration] ich glaube, es geht gar nicht wirklich um Sprache, Kultur oder wie Murat jetzt mit diesem Beispiel kam. [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense pride; style: monologue, casual; average recording, quiet background; 10.2s, DE.
SCRIPT:
[3.3 seconds duration] ihre Identität, nämlich des Türkentums der türkischen Sprache. [0.7 seconds pause] [0.3 seconds duration] Und (surprised gasp, 0.2 seconds) [5.4 seconds duration] ich glaube, es geht gar nicht wirklich um Sprache, Kultur oder wie Murat jetzt mit diesem Beispiel kam. [0.4 seconds pause]
- Tokens:
127
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[3.3 seconds duration] ihre Identität, nämlich des Türkentums der türkischen Sprache. [0.7 seconds pause] [0.3 seconds duration] Und (surprised gasp, 0.2 seconds) [5.4 seconds duration] ich glaube, es geht gar nicht wirklich um Sprache, Kultur oder wie Murat jetzt mit diesem Beispiel kam. [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5591.0000.2980.00010.16 s-0.14 s
SFT220.5591.0000.0000.00010.16 s-0.14 s
SFT230.5591.0000.0000.00010.16 s-0.14 s
SFT240.5591.0000.2980.00010.16 s-0.14 s
SFT2 + DPO210.5590.5840.0000.00010.16 s-0.14 s
SFT2 + DPO220.5591.0000.3230.00010.16 s-0.14 s
SFT2 + DPO230.5591.0000.3230.00010.16 s-0.14 s
SFT2 + DPO240.0320.5190.3100.00010.16 s-0.14 s
Prideenscripted 14.7 s · 1 burst(s)
SFT2
emotion 0.674duration err 0.10 soverall 0.477 ± 0.112
SFT2 + DPO2
emotion 0.671duration err 0.12 soverall 0.455 ± 0.081
script and prompt
[4.7 seconds duration] Why don't you tell us off the air, dude, I hate that we're gonna put this company out of business if we reveal this to the public. [0.5 seconds duration] Exactly. [0.2 seconds pause] [0.4 seconds duration] But where was it? [0.9 seconds duration] I would never, because... [0.9 seconds pause] [1.3 seconds duration] Was it the shop right on Castor Avenue? [0.2 seconds pause] [0.2 seconds duration] No. (ahem, 0.1 seconds) [1.8 seconds pause] [0.7 seconds duration] It's in New Jersey. [0.6 seconds pause] [1.8 seconds duration] I'm not going to fucking Jersey, dude. [0.4 seconds duration] The hell?
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, slightly rough, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is neutral, slightly dominant, neutral openness; reads as intense pride; style: casual, playful; average recording, quiet background; mildly explicit content; 14.9s, EN.
SCRIPT:
[4.7 seconds duration] Why don't you tell us off the air, dude, I hate that we're gonna put this company out of business if we reveal this to the public. [0.5 seconds duration] Exactly. [0.2 seconds pause] [0.4 seconds duration] But where was it? [0.9 seconds duration] I would never, because... [0.9 seconds pause] [1.3 seconds duration] Was it the shop right on Castor Avenue? [0.2 seconds pause] [0.2 seconds duration] No. (ahem, 0.1 seconds) [1.8 seconds pause] [0.7 seconds duration] It's in New Jersey. [0.6 seconds pause] [1.8 seconds duration] I'm not going to fucking Jersey, dude. [0.4 seconds duration] The hell?
- Tokens:
185
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[4.7 seconds duration] Why don't you tell us off the air, dude, I hate that we're gonna put this company out of business if we reveal this to the public. [0.5 seconds duration] Exactly. [0.2 seconds pause] [0.4 seconds duration] But where was it? [0.9 seconds duration] I would never, because... [0.9 seconds pause] [1.3 seconds duration] Was it the shop right on Castor Avenue? [0.2 seconds pause] [0.2 seconds duration] No. (ahem, 0.1 seconds) [1.8 seconds pause] [0.7 seconds duration] It's in New Jersey. [0.6 seconds pause] [1.8 seconds duration] I'm not going to fucking Jersey, dude. [0.4 seconds duration] The hell?
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7341.0000.0000.03414.80 s+0.10 s
SFT220.5591.0000.0000.06914.80 s+0.10 s
SFT230.8431.0000.0000.10314.80 s+0.10 s
SFT240.5591.0000.0000.06914.80 s+0.10 s
SFT2 + DPO210.6251.0000.0000.03414.80 s+0.10 s
SFT2 + DPO220.7841.0000.0000.06914.80 s+0.10 s
SFT2 + DPO230.5591.0000.0000.06914.88 s+0.18 s
SFT2 + DPO240.7171.0000.0000.05214.80 s+0.10 s
Reliefdescripted 17.4 s · 2 burst(s)
SFT2
emotion 0.260duration err 0.04 soverall 0.407 ± 0.045
SFT2 + DPO2
emotion 0.237duration err 0.04 soverall 0.356 ± 0.079
script and prompt
(ahem, 0.2 seconds) [3.8 seconds duration] Die dann hier plötzlich ständig zu Tesla eingeladen und zu Apple eingeladen. [0.4 seconds duration] Und (low mumble, 0.2 seconds) [4.0 seconds duration] was glaubst du, warum die ganzen Podcasts jetzt meinen, die müssten jetzt hier sich eh Autos kaufen? [0.4 seconds pause] [3.3 seconds duration] Weil sie halt keine, weil keine Sau sich mehr für iPhone-Themen interessiert. [4.1 seconds duration] Wenn bockten das die 50. Einstellung und das 50. Subbehördeil? [1.0 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, slightly dominant, slightly guarded; reads as intense relief; style: casual, conversational; good recording, quiet background; 17.4s, DE.
SCRIPT:
(ahem, 0.2 seconds) [3.8 seconds duration] Die dann hier plötzlich ständig zu Tesla eingeladen und zu Apple eingeladen. [0.4 seconds duration] Und (low mumble, 0.2 seconds) [4.0 seconds duration] was glaubst du, warum die ganzen Podcasts jetzt meinen, die müssten jetzt hier sich eh Autos kaufen? [0.4 seconds pause] [3.3 seconds duration] Weil sie halt keine, weil keine Sau sich mehr für iPhone-Themen interessiert. [4.1 seconds duration] Wenn bockten das die 50. Einstellung und das 50. Subbehördeil? [1.0 seconds pause]
- Tokens:
217
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(ahem, 0.2 seconds) [3.8 seconds duration] Die dann hier plötzlich ständig zu Tesla eingeladen und zu Apple eingeladen. [0.4 seconds duration] Und (low mumble, 0.2 seconds) [4.0 seconds duration] was glaubst du, warum die ganzen Podcasts jetzt meinen, die müssten jetzt hier sich eh Autos kaufen? [0.4 seconds pause] [3.3 seconds duration] Weil sie halt keine, weil keine Sau sich mehr für iPhone-Themen interessiert. [4.1 seconds duration] Wenn bockten das die 50. Einstellung und das 50. Subbehördeil? [1.0 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.0661.0000.0000.09417.36 s-0.04 s
SFT220.3870.7250.6600.13217.44 s+0.04 s
SFT230.3871.0000.7100.13217.36 s-0.04 s
SFT240.1981.0000.2150.07517.36 s-0.04 s
SFT2 + DPO210.3091.0000.7691.00017.36 s-0.04 s
SFT2 + DPO220.0531.0000.4980.15117.36 s-0.04 s
SFT2 + DPO230.4061.0000.5400.17017.44 s+0.04 s
SFT2 + DPO240.1811.0000.7691.00017.36 s-0.04 s
Reliefenscripted 18.7 s · 1 burst(s)
SFT2
emotion 0.406duration err 0.02 soverall 0.553 ± 0.021
SFT2 + DPO2
emotion 0.560duration err 0.02 soverall 0.525 ± 0.118
script and prompt
[3.7 seconds duration] And they're serving down time like every day and they're like, yeah, we're going to kill a whole bunch more. [0.9 seconds duration] This is ridiculous. [2.4 seconds duration] Like this is, this is insane. [6.5 seconds duration] Like every single time there's a giant big record and Eve it's insane, but this is even more insane than it was before. [0.5 seconds pause] [3.3 seconds duration] Part of my problem with knowledge on this is I haven't played Eve in quite a while now. [0.7 seconds pause] (low mumble, 0.4 seconds) [0.3 seconds duration] Um.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is energised, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is positive, slightly dominant, neutral openness; reads as intense relief; style: casual, playful; average recording, quiet background; 18.7s, EN.
SCRIPT:
[3.7 seconds duration] And they're serving down time like every day and they're like, yeah, we're going to kill a whole bunch more. [0.9 seconds duration] This is ridiculous. [2.4 seconds duration] Like this is, this is insane. [6.5 seconds duration] Like every single time there's a giant big record and Eve it's insane, but this is even more insane than it was before. [0.5 seconds pause] [3.3 seconds duration] Part of my problem with knowledge on this is I haven't played Eve in quite a while now. [0.7 seconds pause] (low mumble, 0.4 seconds) [0.3 seconds duration] Um.
- Tokens:
233
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[3.7 seconds duration] And they're serving down time like every day and they're like, yeah, we're going to kill a whole bunch more. [0.9 seconds duration] This is ridiculous. [2.4 seconds duration] Like this is, this is insane. [6.5 seconds duration] Like every single time there's a giant big record and Eve it's insane, but this is even more insane than it was before. [0.5 seconds pause] [3.3 seconds duration] Part of my problem with knowledge on this is I haven't played Eve in quite a while now. [0.7 seconds pause] (low mumble, 0.4 seconds) [0.3 seconds duration] Um.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.4891.0000.8190.02818.72 s+0.02 s
SFT220.4731.0000.6410.01418.72 s+0.02 s
SFT230.0231.0001.0000.04218.72 s+0.02 s
SFT240.6381.0001.0000.05618.72 s+0.02 s
SFT2 + DPO210.2531.0001.0000.00018.72 s+0.02 s
SFT2 + DPO220.6871.0000.9230.98618.72 s+0.02 s
SFT2 + DPO230.7101.0000.6670.01418.72 s+0.02 s
SFT2 + DPO240.5911.0000.8190.01418.72 s+0.02 s
Sadnessdescripted 17.5 s · 1 burst(s)
SFT2
emotion 0.657duration err 0.06 soverall 0.248 ± 0.101
SFT2 + DPO2
emotion 0.745duration err 0.02 soverall 0.355 ± 0.175
script and prompt
[5.2 seconds duration] Die Siedlungen haben viel mehr Trubel als der tiefe Wald, und diese Vögel genießen ein gutes Gespräch ganz besonders. [0.6 seconds pause] [1.8 seconds duration] Das merkt man schon daran, wie sehr sie reden wollen. [0.8 seconds pause] (ahem, 0.2 seconds) [8.9 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Sharon
- Instruction:
GENERAL: A voice notably high in harmonicity. The voice is highly resonant and exceptionally focused, featuring a massive fundamental pitch and a remarkably low acoustic noise floor. The tonal energy is highly efficient and distinctly ringing, characteristic of a highly trained public speaker maximizing their vocal fold closure; reads as intense sadness.
SCRIPT:
[5.2 seconds duration] Die Siedlungen haben viel mehr Trubel als der tiefe Wald, und diese Vögel genießen ein gutes Gespräch ganz besonders. [0.6 seconds pause] [1.8 seconds duration] Das merkt man schon daran, wie sehr sie reden wollen. [0.8 seconds pause] (ahem, 0.2 seconds) [8.9 seconds pause]
- Tokens:
219
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[5.2 seconds duration] Die Siedlungen haben viel mehr Trubel als der tiefe Wald, und diese Vögel genießen ein gutes Gespräch ganz besonders. [0.6 seconds pause] [1.8 seconds duration] Das merkt man schon daran, wie sehr sie reden wollen. [0.8 seconds pause] (ahem, 0.2 seconds) [8.9 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5930.5000.6150.03417.52 s+0.02 s
SFT220.7320.5000.0000.65517.60 s+0.10 s
SFT230.7121.0000.0000.34517.52 s+0.02 s
SFT240.5930.5190.0000.89717.60 s+0.10 s
SFT2 + DPO210.7370.6900.0000.79317.52 s+0.02 s
SFT2 + DPO220.5930.5020.0000.34517.52 s+0.02 s
SFT2 + DPO230.7711.0000.0000.00017.52 s+0.02 s
SFT2 + DPO240.8811.0000.0000.72417.52 s+0.02 s
Sadnessenscripted 15.6 s · 1 burst(s)
SFT2
emotion 0.593duration err 0.06 soverall 0.428 ± 0.013
SFT2 + DPO2
emotion 0.593duration err 0.08 soverall 0.409 ± 0.036
script and prompt
[0.9 seconds duration] Yeah, I know. [0.3 seconds duration] It is. [0.5 seconds duration] It's annoying. [0.4 seconds duration] It is. [1.1 seconds duration] Alright, and here we go. [1.6 seconds duration] Alright, so let's have some fun with just some, like, (ahem, 0.1 seconds) [2.1 seconds duration] uh, things to go around, like, awards for the year. [0.5 seconds pause] [0.3 seconds duration] Alright? [0.5 seconds duration] Oh. [0.6 seconds duration] First one. [0.5 seconds pause] [1.5 seconds duration] Just MVP right now. [0.6 seconds pause] [0.3 seconds duration] Who do you think? [0.4 seconds duration] Tyree Kill. [0.7 seconds pause] [2.3 seconds duration] You're gonna take Tyree Kill over Josh Allen right now? [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is energised, brisk, neutral tension, moderately variable; timbre is neutral-toned, slightly bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is positive, slightly dominant, slightly guarded; reads as intense sadness; style: casual, conversational; average recording, some background noise; mildly explicit content; 15.5s, EN.
SCRIPT:
[0.9 seconds duration] Yeah, I know. [0.3 seconds duration] It is. [0.5 seconds duration] It's annoying. [0.4 seconds duration] It is. [1.1 seconds duration] Alright, and here we go. [1.6 seconds duration] Alright, so let's have some fun with just some, like, (ahem, 0.1 seconds) [2.1 seconds duration] uh, things to go around, like, awards for the year. [0.5 seconds pause] [0.3 seconds duration] Alright? [0.5 seconds duration] Oh. [0.6 seconds duration] First one. [0.5 seconds pause] [1.5 seconds duration] Just MVP right now. [0.6 seconds pause] [0.3 seconds duration] Who do you think? [0.4 seconds duration] Tyree Kill. [0.7 seconds pause] [2.3 seconds duration] You're gonna take Tyree Kill over Josh Allen right now? [0.4 seconds pause]
- Tokens:
193
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.9 seconds duration] Yeah, I know. [0.3 seconds duration] It is. [0.5 seconds duration] It's annoying. [0.4 seconds duration] It is. [1.1 seconds duration] Alright, and here we go. [1.6 seconds duration] Alright, so let's have some fun with just some, like, (ahem, 0.1 seconds) [2.1 seconds duration] uh, things to go around, like, awards for the year. [0.5 seconds pause] [0.3 seconds duration] Alright? [0.5 seconds duration] Oh. [0.6 seconds duration] First one. [0.5 seconds pause] [1.5 seconds duration] Just MVP right now. [0.6 seconds pause] [0.3 seconds duration] Who do you think? [0.4 seconds duration] Tyree Kill. [0.7 seconds pause] [2.3 seconds duration] You're gonna take Tyree Kill over Josh Allen right now? [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5931.0000.7260.24115.52 s-0.08 s
SFT220.5931.0000.6190.29315.60 s+0.00 s
SFT230.5931.0000.4840.15515.52 s-0.08 s
SFT240.5931.0000.6700.22415.52 s-0.08 s
SFT2 + DPO210.5931.0000.1440.19015.52 s-0.08 s
SFT2 + DPO220.5931.0000.6980.20715.52 s-0.08 s
SFT2 + DPO230.5931.0000.6980.22415.52 s-0.08 s
SFT2 + DPO240.5931.0000.2350.15515.52 s-0.08 s
Sexual Lustdescripted 21.3 s · 1 burst(s)
SFT2
emotion 0.184duration err 0.18 soverall 0.516 ± 0.101
SFT2 + DPO2
emotion 0.364duration err 0.20 soverall 0.477 ± 0.073
script and prompt
[5.2 seconds duration] Weil du weißt doch, wir bewegen uns heutzutage immerhin in einer Welt, in der wir immer mehr Daten über uns preisgeben. [0.4 seconds pause] [2.2 seconds duration] Überall kommen uns Angebote unter die Nase, [0.6 seconds pause] (low mumble, 0.2 seconds) [10.6 seconds duration] die angeblich kostenlos sind von irgendwelchen Diensten, die man uns dann erweisen möchte, sei es E-Mail, Chat, Nachrichten, Werbung, Gutscheine, Freifahrten, whatsoever. [0.4 seconds pause] [1.7 seconds duration] Aber in Wirklichkeit sind wir ja das Produkt.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, slightly dominant, slightly guarded; reads as intense sexual lust; style: conversational, casual; good recording, quiet background; 21.2s, DE.
SCRIPT:
[5.2 seconds duration] Weil du weißt doch, wir bewegen uns heutzutage immerhin in einer Welt, in der wir immer mehr Daten über uns preisgeben. [0.4 seconds pause] [2.2 seconds duration] Überall kommen uns Angebote unter die Nase, [0.6 seconds pause] (low mumble, 0.2 seconds) [10.6 seconds duration] die angeblich kostenlos sind von irgendwelchen Diensten, die man uns dann erweisen möchte, sei es E-Mail, Chat, Nachrichten, Werbung, Gutscheine, Freifahrten, whatsoever. [0.4 seconds pause] [1.7 seconds duration] Aber in Wirklichkeit sind wir ja das Produkt.
- Tokens:
264
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[5.2 seconds duration] Weil du weißt doch, wir bewegen uns heutzutage immerhin in einer Welt, in der wir immer mehr Daten über uns preisgeben. [0.4 seconds pause] [2.2 seconds duration] Überall kommen uns Angebote unter die Nase, [0.6 seconds pause] (low mumble, 0.2 seconds) [10.6 seconds duration] die angeblich kostenlos sind von irgendwelchen Diensten, die man uns dann erweisen möchte, sei es E-Mail, Chat, Nachrichten, Werbung, Gutscheine, Freifahrten, whatsoever. [0.4 seconds pause] [1.7 seconds duration] Aber in Wirklichkeit sind wir ja das Produkt.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1721.0000.8190.00021.12 s-0.18 s
SFT220.0871.0000.8190.00021.12 s-0.18 s
SFT230.4640.5070.6980.00021.12 s-0.18 s
SFT240.0121.0000.9610.00021.12 s-0.18 s
SFT2 + DPO210.6031.0000.3360.00021.12 s-0.18 s
SFT2 + DPO220.3590.5000.8190.00021.04 s-0.26 s
SFT2 + DPO230.3481.0000.8190.01721.12 s-0.18 s
SFT2 + DPO240.1441.0000.6150.00021.12 s-0.18 s
Sexual Lustenscripted 8.8 s · 0 burst(s)
SFT2
emotion 0.690duration err 0.08 soverall 0.610 ± 0.139
SFT2 + DPO2
emotion 0.608duration err 0.08 soverall 0.506 ± 0.078
script and prompt
[0.6 seconds duration] But now I'm (ahem) [1.4 seconds duration] gonna really take my time. [3.7 seconds duration] And he brings the dog down, which, everyone's like, that dog's scary. [0.2 seconds pause] [0.6 seconds duration] I want that dog. [1.3 seconds pause] [0.8 seconds duration] That dog is amazing. [0.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, wide pitch range, light breath; affect is positive, slightly dominant, neutral openness; reads as intense sexual lust; style: casual, conversational; average recording, some background noise; mildly explicit content; 8.8s, EN.
SCRIPT:
[0.6 seconds duration] But now I'm (ahem) [1.4 seconds duration] gonna really take my time. [3.7 seconds duration] And he brings the dog down, which, everyone's like, that dog's scary. [0.2 seconds pause] [0.6 seconds duration] I want that dog. [1.3 seconds pause] [0.8 seconds duration] That dog is amazing. [0.2 seconds pause]
- Tokens:
109
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[0.6 seconds duration] But now I'm (ahem) [1.4 seconds duration] gonna really take my time. [3.7 seconds duration] And he brings the dog down, which, everyone's like, that dog's scary. [0.2 seconds pause] [0.6 seconds duration] I want that dog. [1.3 seconds pause] [0.8 seconds duration] That dog is amazing. [0.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7261.0001.0000.0718.72 s-0.08 s
SFT220.5971.0000.0000.0718.72 s-0.08 s
SFT230.7461.0001.0000.0368.72 s-0.08 s
SFT240.6921.0001.0000.0008.72 s-0.08 s
SFT2 + DPO210.6251.0001.0000.0718.72 s-0.08 s
SFT2 + DPO220.5491.0000.0000.0718.72 s-0.08 s
SFT2 + DPO230.6031.0001.0000.0718.72 s-0.08 s
SFT2 + DPO240.6551.0001.0000.1438.72 s-0.08 s
Shamedescripted 10.5 s · 1 burst(s)
SFT2
emotion 0.122duration err 0.18 soverall 0.369 ± 0.097
SFT2 + DPO2
emotion 0.331duration err 0.18 soverall 0.522 ± 0.180
script and prompt
[1.0 seconds duration] Ja, was soll ich sagen? [1.0 seconds duration] Der Alex, hm, (low mumble, 0.1 seconds) [0.6 seconds pause] [5.7 seconds duration] Alexander Vogt hat sich auch auf meinen Kommentar zurückgemeldet, hat gesagt Danke und er hätte diese Folge, äh, (ahem) [2.1 seconds duration] die Folge dieser Woche einfach mal da durchgejagt.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly positive, neutral stance, slightly guarded; reads as intense shame; style: conversational, playful; good recording, quiet background; 10.3s, DE.
SCRIPT:
[1.0 seconds duration] Ja, was soll ich sagen? [1.0 seconds duration] Der Alex, hm, (low mumble, 0.1 seconds) [0.6 seconds pause] [5.7 seconds duration] Alexander Vogt hat sich auch auf meinen Kommentar zurückgemeldet, hat gesagt Danke und er hätte diese Folge, äh, (ahem) [2.1 seconds duration] die Folge dieser Woche einfach mal da durchgejagt.
- Tokens:
129
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[1.0 seconds duration] Ja, was soll ich sagen? [1.0 seconds duration] Der Alex, hm, (low mumble, 0.1 seconds) [0.6 seconds pause] [5.7 seconds duration] Alexander Vogt hat sich auch auf meinen Kommentar zurückgemeldet, hat gesagt Danke und er hätte diese Folge, äh, (ahem) [2.1 seconds duration] die Folge dieser Woche einfach mal da durchgejagt.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.2591.0000.3250.08810.32 s-0.18 s
SFT220.1441.0000.2010.17610.32 s-0.18 s
SFT230.0661.0000.4650.05910.32 s-0.18 s
SFT240.0190.5000.2070.14710.32 s-0.18 s
SFT2 + DPO210.1161.0000.4290.11810.32 s-0.18 s
SFT2 + DPO220.9151.0000.3960.05910.32 s-0.18 s
SFT2 + DPO230.2591.0000.0000.05910.32 s-0.18 s
SFT2 + DPO240.0351.0000.3520.05910.32 s-0.18 s
Shameenscripted 13.0 s · 1 burst(s)
SFT2
emotion 0.518duration err 0.12 soverall 0.518 ± 0.025
SFT2 + DPO2
emotion 0.548duration err 0.12 soverall 0.548 ± 0.035
script and prompt
[1.0 seconds pause] [5.0 seconds duration] I come closer to being able to give specifics on this, like when people say, what would it take to convince you that God exists? [2.7 seconds duration] And my answer is always, I don't know, but if there's a God it knows. [0.9 seconds duration] You can't do that here. [0.9 seconds pause] (low mumble, 0.1 seconds) [0.2 seconds duration] Um, [0.4 seconds pause] [1.8 seconds duration] well I probably can't give specifics because
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense shame; style: conversational, casual; average recording, quiet background; 13.2s, EN.
SCRIPT:
[1.0 seconds pause] [5.0 seconds duration] I come closer to being able to give specifics on this, like when people say, what would it take to convince you that God exists? [2.7 seconds duration] And my answer is always, I don't know, but if there's a God it knows. [0.9 seconds duration] You can't do that here. [0.9 seconds pause] (low mumble, 0.1 seconds) [0.2 seconds duration] Um, [0.4 seconds pause] [1.8 seconds duration] well I probably can't give specifics because
- Tokens:
164
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[1.0 seconds pause] [5.0 seconds duration] I come closer to being able to give specifics on this, like when people say, what would it take to convince you that God exists? [2.7 seconds duration] And my answer is always, I don't know, but if there's a God it knows. [0.9 seconds duration] You can't do that here. [0.9 seconds pause] (low mumble, 0.1 seconds) [0.2 seconds duration] Um, [0.4 seconds pause] [1.8 seconds duration] well I probably can't give specifics because
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6571.0000.6190.01913.12 s+0.12 s
SFT220.6911.0000.4470.00013.12 s+0.12 s
SFT230.3801.0000.5490.03813.12 s+0.12 s
SFT240.3421.0000.5490.01913.12 s+0.12 s
SFT2 + DPO210.6961.0000.6440.00013.12 s+0.12 s
SFT2 + DPO220.7321.0000.1390.01913.12 s+0.12 s
SFT2 + DPO230.3001.0000.6980.01913.12 s+0.12 s
SFT2 + DPO240.4631.0000.5950.01913.12 s+0.12 s
Sournessdescripted 12.4 s · 0 burst(s)
SFT2
emotion 0.663duration err 0.00 soverall 0.346 ± 0.144
SFT2 + DPO2
emotion 0.489duration err 0.02 soverall 0.371 ± 0.166
script and prompt
[0.7 seconds duration] Also, (contented sigh) [0.5 seconds pause] [7.4 seconds duration] ich denke wirklich, wir sollten diese blockierten Wörter in der Nachricht zumindest farblich kennzeichnen, weißt du, damit die Leute sie sehen und sie meiden können. [0.4 seconds pause] [2.2 seconds duration] Das würde die Benutzerfreundlichkeit sehr verbessern. [1.2 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Emerson
- Instruction:
GENERAL: A voice notably high in focus. The speaker employs clear, highly engaged 'Direct Address,' actively ensuring their acoustic energy firmly lands on the listener. The voice possesses a strong forward projection that metaphorically 'looks' right at the audience, demanding their attention and cognitive participation; reads as intense sourness.
SCRIPT:
[0.7 seconds duration] Also, (contented sigh) [0.5 seconds pause] [7.4 seconds duration] ich denke wirklich, wir sollten diese blockierten Wörter in der Nachricht zumindest farblich kennzeichnen, weißt du, damit die Leute sie sehen und sie meiden können. [0.4 seconds pause] [2.2 seconds duration] Das würde die Benutzerfreundlichkeit sehr verbessern. [1.2 seconds pause]
- Tokens:
155
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[0.7 seconds duration] Also, (contented sigh) [0.5 seconds pause] [7.4 seconds duration] ich denke wirklich, wir sollten diese blockierten Wörter in der Nachricht zumindest farblich kennzeichnen, weißt du, damit die Leute sie sehen und sie meiden können. [0.4 seconds pause] [2.2 seconds duration] Das würde die Benutzerfreundlichkeit sehr verbessern. [1.2 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6860.5241.0000.96912.40 s+0.00 s
SFT220.7660.5051.0000.00012.40 s+0.00 s
SFT230.6940.6310.0000.00012.40 s+0.00 s
SFT240.5040.5910.0000.00012.40 s+0.00 s
SFT2 + DPO210.7940.5101.0000.09412.48 s+0.08 s
SFT2 + DPO220.5831.0000.0000.03112.40 s+0.00 s
SFT2 + DPO230.3911.0000.0000.00012.40 s+0.00 s
SFT2 + DPO240.1870.0741.0001.00012.40 s+0.00 s
Sournessenscripted 16.5 s · 1 burst(s)
SFT2
emotion 0.176duration err 0.10 soverall 0.519 ± 0.066
SFT2 + DPO2
emotion 0.292duration err 0.12 soverall 0.474 ± 0.050
script and prompt
[5.9 seconds duration] You know, you're at the baseline, you're at now another baseline, and you have to then get all these other skillsets. [7.5 seconds duration] So within the teams, it's where there's any gaps, so whether it's language demolitions, you know, I was the forward air controller, so mine was anything to do with air, uh, [0.3 seconds pause] (low mumble, 0.3 seconds) [2.5 seconds duration] was me, so I was the FAC within ours.
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, neutral tension, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is mildly negative, neutral stance, slightly guarded; reads as intense sourness; style: conversational, casual; good recording, quiet background; 16.5s, EN.
SCRIPT:
[5.9 seconds duration] You know, you're at the baseline, you're at now another baseline, and you have to then get all these other skillsets. [7.5 seconds duration] So within the teams, it's where there's any gaps, so whether it's language demolitions, you know, I was the forward air controller, so mine was anything to do with air, uh, [0.3 seconds pause] (low mumble, 0.3 seconds) [2.5 seconds duration] was me, so I was the FAC within ours.
- Tokens:
205
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[5.9 seconds duration] You know, you're at the baseline, you're at now another baseline, and you have to then get all these other skillsets. [7.5 seconds duration] So within the teams, it's where there's any gaps, so whether it's language demolitions, you know, I was the forward air controller, so mine was anything to do with air, uh, [0.3 seconds pause] (low mumble, 0.3 seconds) [2.5 seconds duration] was me, so I was the FAC within ours.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.1341.0000.9230.09816.40 s-0.10 s
SFT220.1341.0001.0000.04916.40 s-0.10 s
SFT230.4191.0000.3380.14816.40 s-0.10 s
SFT240.0151.0001.0000.06616.40 s-0.10 s
SFT2 + DPO210.0081.0000.6980.04916.40 s-0.10 s
SFT2 + DPO220.4191.0000.3780.21316.40 s-0.10 s
SFT2 + DPO230.3911.0000.5680.08216.32 s-0.18 s
SFT2 + DPO240.3511.0000.7870.08216.40 s-0.10 s
Teasingdescripted 17.5 s · 2 burst(s)
SFT2
emotion 0.562duration err 0.06 soverall 0.556 ± 0.023
SFT2 + DPO2
emotion 0.533duration err 0.06 soverall 0.524 ± 0.028
script and prompt
(low mumble, 0.4 seconds) [2.9 seconds duration] Ich hab jetzt selten Konzepte erlebt, die ganz anders sind, glaube ich. [0.4 seconds pause] [5.8 seconds duration] Aber wenn ich jetzt in einen Raum reingehen würde und da ist so ein Smartboard mit so komischen digitalen Stiften, da würde ich da erstmal auch [0.2 seconds pause] (low mumble, 0.3 seconds) [1.6 seconds duration] gucken, wie ich das Zeug benutze. [0.5 seconds pause] [5.1 seconds duration] Und das ist, glaube ich, bei diesen digitalen Tools, die sind, also es gibt noch nicht so viele [0.3 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: An adult masculine voice; delivery is normally alert, normal-paced, slightly relaxed, fairly steady; timbre is neutral-toned, neutral-bright, fairly smooth, balanced body; average clarity, some disfluency, moderate pitch range, light breath; affect is neutral, neutral stance, slightly guarded; reads as intense teasing; style: casual, monologue; good recording, quiet background; 17.5s, DE.
SCRIPT:
(low mumble, 0.4 seconds) [2.9 seconds duration] Ich hab jetzt selten Konzepte erlebt, die ganz anders sind, glaube ich. [0.4 seconds pause] [5.8 seconds duration] Aber wenn ich jetzt in einen Raum reingehen würde und da ist so ein Smartboard mit so komischen digitalen Stiften, da würde ich da erstmal auch [0.2 seconds pause] (low mumble, 0.3 seconds) [1.6 seconds duration] gucken, wie ich das Zeug benutze. [0.5 seconds pause] [5.1 seconds duration] Und das ist, glaube ich, bei diesen digitalen Tools, die sind, also es gibt noch nicht so viele [0.3 seconds pause]
- Tokens:
219
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
(low mumble, 0.4 seconds) [2.9 seconds duration] Ich hab jetzt selten Konzepte erlebt, die ganz anders sind, glaube ich. [0.4 seconds pause] [5.8 seconds duration] Aber wenn ich jetzt in einen Raum reingehen würde und da ist so ein Smartboard mit so komischen digitalen Stiften, da würde ich da erstmal auch [0.2 seconds pause] (low mumble, 0.3 seconds) [1.6 seconds duration] gucken, wie ich das Zeug benutze. [0.5 seconds pause] [5.1 seconds duration] Und das ist, glaube ich, bei diesen digitalen Tools, die sind, also es gibt noch nicht so viele [0.3 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5041.0000.8210.01617.44 s-0.06 s
SFT220.7341.0000.3350.01617.44 s-0.06 s
SFT230.5041.0000.7610.04817.44 s-0.06 s
SFT240.5041.0000.9240.06517.44 s-0.06 s
SFT2 + DPO210.5041.0000.6980.03217.44 s-0.06 s
SFT2 + DPO220.5041.0000.9090.03217.44 s-0.06 s
SFT2 + DPO230.6211.0000.4620.01617.44 s-0.06 s
SFT2 + DPO240.5041.0000.6690.01617.44 s-0.06 s
Teasingenscripted 13.7 s · 1 burst(s)
SFT2
emotion 0.642duration err 0.04 soverall 0.494 ± 0.067
SFT2 + DPO2
emotion 0.641duration err 0.04 soverall 0.488 ± 0.041
script and prompt
[1.5 seconds duration] Nah, they do, but it's real. (chuckle, 0.2 seconds) [1.2 seconds duration] That's what they be doing. [2.2 seconds duration] Like first two years guaranteed or something. [1.2 seconds duration] Team option, team option. [0.7 seconds duration] And he's like, damn. [0.5 seconds pause] [1.7 seconds duration] So they don't even get to the money fast no more. [3.4 seconds duration] But like back in the day, like when West Matthews and them got paid fast. [1.1 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult masculine voice; delivery is normally alert, normal-paced, neutral tension, moderately variable; timbre is neutral-toned, very dark, slightly rough, balanced body; somewhat unclear, frequent disfluency, wide pitch range, normal breath; affect is positive, slightly dominant, neutral openness; reads as intense teasing; style: casual, playful; below-average recording, some background noise; mildly explicit content; 13.7s, EN.
SCRIPT:
[1.5 seconds duration] Nah, they do, but it's real. (chuckle, 0.2 seconds) [1.2 seconds duration] That's what they be doing. [2.2 seconds duration] Like first two years guaranteed or something. [1.2 seconds duration] Team option, team option. [0.7 seconds duration] And he's like, damn. [0.5 seconds pause] [1.7 seconds duration] So they don't even get to the money fast no more. [3.4 seconds duration] But like back in the day, like when West Matthews and them got paid fast. [1.1 seconds pause]
- Tokens:
171
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[1.5 seconds duration] Nah, they do, but it's real. (chuckle, 0.2 seconds) [1.2 seconds duration] That's what they be doing. [2.2 seconds duration] Like first two years guaranteed or something. [1.2 seconds duration] Team option, team option. [0.7 seconds duration] And he's like, damn. [0.5 seconds pause] [1.7 seconds duration] So they don't even get to the money fast no more. [3.4 seconds duration] But like back in the day, like when West Matthews and them got paid fast. [1.1 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.5861.0000.7870.05813.68 s-0.02 s
SFT220.6611.0000.0000.03813.60 s-0.10 s
SFT230.7841.0000.1700.11513.68 s-0.02 s
SFT240.5361.0000.5070.07713.68 s-0.02 s
SFT2 + DPO210.6731.0000.5950.07713.68 s-0.02 s
SFT2 + DPO220.5761.0000.4290.03813.60 s-0.10 s
SFT2 + DPO230.6551.0000.7560.03813.68 s-0.02 s
SFT2 + DPO240.6591.0000.1760.03813.68 s-0.02 s
Thankfulness Gratitudedescripted 16.4 s · 1 burst(s)
SFT2
emotion 0.795duration err 0.04 soverall 0.560 ± 0.116
SFT2 + DPO2
emotion 0.780duration err 0.06 soverall 0.548 ± 0.101
script and prompt
[6.0 seconds duration] Ihre Mutter hat ihr diesen leuchtend roten Strampler mit einer kleinen Mütze genäht, und die Krankenschwester hat ihn ihr sofort angezogen. [5.9 seconds pause] [2.7 seconds duration] Deshalb haben wir den Namen gewählt; es fühlte sich einfach richtig an. [1.2 seconds pause] (childlike giggle, 0.2 seconds) [0.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Sheila
- Instruction:
GENERAL: A voice very high in volatility. The emotional state is completely chaotic and deeply shattered, featuring violent, rapid-fire affective switching that fundamentally destabilizes the entire acoustic signal. The speaker cycles through manic, extreme emotional extremes every few seconds, entirely destroying any sense of a consistent psychological baseline; reads as intense thankfulness gratitude.
SCRIPT:
[6.0 seconds duration] Ihre Mutter hat ihr diesen leuchtend roten Strampler mit einer kleinen Mütze genäht, und die Krankenschwester hat ihn ihr sofort angezogen. [5.9 seconds pause] [2.7 seconds duration] Deshalb haben wir den Namen gewählt; es fühlte sich einfach richtig an. [1.2 seconds pause] (childlike giggle, 0.2 seconds) [0.4 seconds pause]
- Tokens:
205
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[6.0 seconds duration] Ihre Mutter hat ihr diesen leuchtend roten Strampler mit einer kleinen Mütze genäht, und die Krankenschwester hat ihn ihr sofort angezogen. [5.9 seconds pause] [2.7 seconds duration] Deshalb haben wir den Namen gewählt; es fühlte sich einfach richtig an. [1.2 seconds pause] (childlike giggle, 0.2 seconds) [0.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.8181.0000.0000.06116.40 s+0.00 s
SFT220.8071.0000.0000.39416.40 s+0.00 s
SFT230.8861.0000.0000.09116.48 s+0.08 s
SFT240.6691.0000.0000.00016.32 s-0.08 s
SFT2 + DPO210.6091.0000.0000.03016.48 s+0.08 s
SFT2 + DPO220.7911.0000.2240.06116.48 s+0.08 s
SFT2 + DPO230.8751.0000.0630.21216.40 s+0.00 s
SFT2 + DPO240.8440.7620.0000.15216.48 s+0.08 s
Thankfulness Gratitudeenscripted 4.1 s · 2 burst(s)
SFT2
emotion 0.907duration err 0.10 soverall 0.699 ± 0.019
SFT2 + DPO2
emotion 0.438duration err 0.10 soverall 0.453 ± 0.083
script and prompt
[2.6 seconds duration] I spend a lot of time with Adam and Owen, (sharp inhale, 0.3 seconds) [0.6 seconds pause] [0.2 seconds duration] and (low mumble, 0.1 seconds) [0.3 seconds duration] uhm,
full prompt as the model sees it
<user_inst>
- Reference(s):
None
- Instruction:
GENERAL: A young adult feminine voice; delivery is normally alert, measured, relaxed, moderately variable; timbre is neutral-toned, dark, fairly smooth, thin; somewhat unclear, frequent disfluency, moderate pitch range, normal breath; affect is mildly negative, submissive, neutral openness; reads as intense thankfulness gratitude; style: casual, conversational; average recording, quiet background; 4.1s, EN.
SCRIPT:
[2.6 seconds duration] I spend a lot of time with Adam and Owen, (sharp inhale, 0.3 seconds) [0.6 seconds pause] [0.2 seconds duration] and (low mumble, 0.1 seconds) [0.3 seconds duration] uhm,
- Tokens:
50
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[2.6 seconds duration] I spend a lot of time with Adam and Owen, (sharp inhale, 0.3 seconds) [0.6 seconds pause] [0.2 seconds duration] and (low mumble, 0.1 seconds) [0.3 seconds duration] uhm,
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.8501.0000.3710.1674.00 s-0.10 s
SFT220.9581.0000.2680.1674.00 s-0.10 s
SFT230.9361.0000.3210.2504.00 s-0.10 s
SFT240.8831.0000.1840.1674.00 s-0.10 s
SFT2 + DPO210.3331.0000.3250.0834.00 s-0.10 s
SFT2 + DPO220.3101.0000.2070.1674.00 s-0.10 s
SFT2 + DPO230.3611.0000.3990.1674.00 s-0.10 s
SFT2 + DPO240.7491.0000.4240.0834.00 s-0.10 s
Triumphdescripted 10.2 s · 1 burst(s)
SFT2
emotion 0.353duration err 0.20 soverall 0.460 ± 0.075
SFT2 + DPO2
emotion 0.302duration err 0.20 soverall 0.381 ± 0.024
script and prompt
[3.3 seconds duration] In zwei Wochen sollten wir bereit sein, mit der wichtigen Arbeit zu beginnen. [0.4 seconds pause] [3.9 seconds duration] Das gibt uns genug Zeit, uns gründlich auf die nächste Phase vorzubereiten. [1.0 seconds pause] (contented sigh, 0.2 seconds) [1.4 seconds pause]
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Larry
- Instruction:
GENERAL: A voice somewhat low in head resonance. Head resonance is minimally present, with only a subtle layer of upper brightness supporting the overall acoustic profile. The voice retains some natural sparkle but lacks the full, bright head resonance characteristic of maximally engaged delivery; reads as intense triumph.
SCRIPT:
[3.3 seconds duration] In zwei Wochen sollten wir bereit sein, mit der wichtigen Arbeit zu beginnen. [0.4 seconds pause] [3.9 seconds duration] Das gibt uns genug Zeit, uns gründlich auf die nächste Phase vorzubereiten. [1.0 seconds pause] (contented sigh, 0.2 seconds) [1.4 seconds pause]
- Tokens:
130
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
German
- Text:
[3.3 seconds duration] In zwei Wochen sollten wir bereit sein, mit der wichtigen Arbeit zu beginnen. [0.4 seconds pause] [3.9 seconds duration] Das gibt uns genug Zeit, uns gründlich auf die nächste Phase vorzubereiten. [1.0 seconds pause] (contented sigh, 0.2 seconds) [1.4 seconds pause]
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.6851.0000.0000.00010.40 s+0.20 s
SFT220.0251.0000.9230.00010.40 s+0.20 s
SFT230.0531.0000.0000.00010.40 s+0.20 s
SFT240.6501.0000.0000.00010.40 s+0.20 s
SFT2 + DPO210.7250.6550.0000.00010.40 s+0.20 s
SFT2 + DPO220.0981.0000.0000.00010.40 s+0.20 s
SFT2 + DPO230.2850.8550.0000.00010.40 s+0.20 s
SFT2 + DPO240.0981.0000.0000.00010.40 s+0.20 s
Triumphenscripted 8.5 s · 1 burst(s)
SFT2
emotion 0.740duration err 0.02 soverall 0.524 ± 0.217
SFT2 + DPO2
emotion 0.907duration err 0.02 soverall 0.742 ± 0.124
script and prompt
[3.9 seconds duration] November was the only month the headline appeared: The last sporting taboo. [0.8 seconds pause] (contented sigh, 0.2 seconds) [0.4 seconds pause] [3.2 seconds duration] That finally being over feels like such a weight lifted.
full prompt as the model sees it
<user_inst>
- Reference(s):
Speaker: Rodrigo
- Instruction:
GENERAL: A voice deeply expressing relief, a voice flooding with relief, exhaling the tension, grateful it is finally over, in every breath. The feeling is fought down rather than shown: it is held in and only leaks through at the edges of phrases; reads as intense triumph.
SCRIPT:
[3.9 seconds duration] November was the only month the headline appeared: The last sporting taboo. [0.8 seconds pause] (contented sigh, 0.2 seconds) [0.4 seconds pause] [3.2 seconds duration] That finally being over feels like such a weight lifted.
- Tokens:
106
- Quality:
None
- Sound Event:
None
- Ambient Sound:
None
- Language:
English
- Text:
[3.9 seconds duration] November was the only month the headline appeared: The last sporting taboo. [0.8 seconds pause] (contented sigh, 0.2 seconds) [0.4 seconds pause] [3.2 seconds duration] That finally being over feels like such a weight lifted.
</user_inst>
modelcompletionemotion pctqualityburstWERlengtherr
SFT210.7940.6090.0000.0008.48 s-0.02 s
SFT220.9081.0000.0000.0008.48 s-0.02 s
SFT230.2850.5620.0000.0918.48 s-0.02 s
SFT240.9730.5340.0000.0008.48 s-0.02 s
SFT2 + DPO210.8900.5730.0000.0008.48 s-0.02 s
SFT2 + DPO220.9021.0000.8870.0458.48 s-0.02 s
SFT2 + DPO230.8981.0000.0000.0458.48 s-0.02 s
SFT2 + DPO240.9380.6410.0000.0008.48 s-0.02 s

How to read the numbers

Provenance

Prompt text is drawn only from sources that are already public: emolia, kartoffelphon and mls and the LAION voice profiles. Material that was deliberately left out of those releases is excluded here too, by an allowlist rather than a blocklist. All audio on this page is generated by the model; none of it is source recording.