Anonymous
8/18/2025, 7:23:50 AM
No.103621971
>>103621667
NTA but yes it does, though I haven't figured out how to set emotion markers yet. The model seems to just base it on the vibes of the text. There's not much documentation on how to control it. You might be able to get away with labeling the transcript of the reference audio as [angry] or [sad] or [intimate], but I haven't tried that yet. Generally it just tends to read out whatever tags there are, but maybe you could also do it audiobook style (ala. "She said with an excited tone, 'My dog died yesterday LETS GOOOO'" or something to that effect)
NTA but yes it does, though I haven't figured out how to set emotion markers yet. The model seems to just base it on the vibes of the text. There's not much documentation on how to control it. You might be able to get away with labeling the transcript of the reference audio as [angry] or [sad] or [intimate], but I haven't tried that yet. Generally it just tends to read out whatever tags there are, but maybe you could also do it audiobook style (ala. "She said with an excited tone, 'My dog died yesterday LETS GOOOO'" or something to that effect)