bark
bark copied to clipboard
How do you get the same voice?
How do you get the same voice?
I used "MALE:" but every time it is another voice. Do you have a "value" you can pass to the AI for "seed"? I understand this is just a start.
Look at the ReadMe at: https://github.com/suno-ai/bark#-voice-presets-and-voiceaudio-cloning
You can specify the voice in the generate_audio function.
yeah, try using history_prompt with a specific speaker. some might work better than others for voice consistency
Hey @TheCompAce,
You can specify a constant voice for generation using the history_prompt parameter in the generate_audio function...
The pattern is {lang_code}_speaker_{0-9}
Example:
audio_array = generate_audio(text_prompt, history_prompt="en_speaker_1")
It has been mentionned in the documentation !
OK thanks, maybe the project's head will take these questions and have ChatGPT make a readme file for them.
hahaha, i'll give it a try lol
Hey @TheCompAce,
You can specify a constant voice for generation using the
history_promptparameter in thegenerate_audiofunction... The pattern is{lang_code}_speaker_{0-9}Example:
audio_array = generate_audio(text_prompt, history_prompt="en_speaker_1")It has been mentionned in the documentation !
Where can I see all the speakers we can use and pass to history_prompt?
0-9 for all languages exist. you can load them in numpy and use the fine array with codec_decode to listen what they sound like
Hi all,
I noticed for East Indian languages, they have Hindi but not Tamil. Anyone know if there is a plan to include voices (male and female) for Bark?
Also, same question as above, it's tricky to get the same voice for every generation. Not sure if history_prompt really works.