do_tts.py - Why can't I use a custom voice?

[dekker33]

Can someone please tell me what I am doing wrong....

I have a folder called "train_trump" in the "voices" directory, with 3 ~10 second long recordings of trump talking, all floating point .wavs with a 22050 sample rate.

Anaconda Prompt:

C:\Users\me\anaconda3\tortoise-tts>python tortoise/do_tts.py --text "the virus is from china" --voice train_trump --preset fast

C:\Users\me\anaconda3\tortoise-tts\tortoise\utils\audio.py:17: WavFileWarning: Chunk (non-data) not understood, skipping it.
sampling_rate, data = read(full_path)
Traceback (most recent call last):
File "tortoise/do_tts.py", line 35, in
voice_samples, conditioning_latents = load_voices(voice_sel)
File "C:\Users\me\anaconda3\tortoise-tts\tortoise\utils\audio.py", line 124, in load_voices
clip, latent = load_voice(voice, extra_voice_dirs)
File "C:\Users\me\anaconda3\tortoise-tts\tortoise\utils\audio.py", line 111, in load_voice
c = load_audio(cond_path, 22050)
File "C:\Users\me\anaconda3\tortoise-tts\tortoise\utils\audio.py", line 31, in load_audio
audio, lsr = load_wav_to_torch(audiopath)
File "C:\Users\me\anaconda3\tortoise-tts\tortoise\utils\audio.py", line 25, in load_wav_to_torch
raise NotImplemented(f"Provided data dtype not supported: {data.dtype}")
TypeError: 'NotImplementedType' object is not callable

[neonbjb]

scipy.io.wavfile.read doesn't like your WAV format. I'd ask chatgpt about reasons why that might occur. Try using a quantized type like uint8.

[dekker33]

Thanks neon!
I normalized my audio and changed it from 64 floating point to 32 floating point!
I think it's working now