Error: bad sample width

[efoxxfiles]

Hi

Today bark infinity stop to work
Any idea?

tqdm._instances.clear()
full_generation_segments, audio_arr_segments, final_filename_will_be = api.generate_audio_long(text_prompt=Texto, history_prompt=history_prompt)

Audio(np.concatenate(audio_arr_segments), rate=generation.SAMPLE_RATE)

Error Traceback (most recent call last)

in <cell line: 8>()
6
7 tqdm._instances.clear()
----> 8 full_generation_segments, audio_arr_segments, final_filename_will_be = api.generate_audio_long(text_prompt=Texto, history_prompt=history_prompt)
9
10 Audio(np.concatenate(audio_arr_segments), rate=generation.SAMPLE_RATE)

5 frames

/usr/lib/python3.10/wave.py in setsampwidth(self, sampwidth)
351 raise Error('cannot change parameters after starting to write')
352 if sampwidth < 1 or sampwidth > 4:
--> 353 raise Error('bad sample width')
354 self._sampwidth = sampwidth
355

Error: bad sample width

[DiamondGlassDrill]

I get the same error.

Traceback (most recent call last):
File "K:\Voice\bark\bark_perform.py", line 136, in
main(namespace_args)
File "K:\Voice\bark\bark_perform.py", line 121, in main
api.generate_audio_long(**args_dict)
File "K:\Voice\bark\bark_infinity\api.py", line 1099, in generate_audio_long
write_one_segment(audio_arr = np.concatenate(audio_arr_segments), full_generation = full_generation_segments[0], **kwargs)
File "K:\Voice\bark\bark_infinity\api.py", line 360, in write_one_segment
write_seg_wav(filepath, audio_arr, **kwargs)
File "K:\Voice\bark\bark_infinity\api.py", line 409, in write_seg_wav
write_audiofile(filepath, audio_arr, **kwargs)
File "K:\Voice\bark\bark_infinity\api.py", line 438, in write_audiofile
audio.export(output_filepath, format="mp4", codec="aac")
File "C:\mambaforge\envs\clonevoice\lib\site-packages\pydub\audio_segment.py", line 890, in export
wave_data.setsampwidth(self.sample_width)
File "C:\mambaforge\envs\clonevoice\lib\wave.py", line 353, in setsampwidth
raise Error('bad sample width')
wave.Error: bad sample width
Exception ignored in: <function Wave_write.del at 0x000001F0B8798A60>

All generated .mp4 or .mp3 files are empty, where the .npz are being generated and contain data

[JonathanFly]

I think I have a fix. Make sure you aren't using MP4 format for now. I'll be online tomorrow Sunday evening will patch.

[JonathanFly]

This should be fixed. Let me know if works.

git pull to update

As far I can tell you only got the error if you tried to add silence between segments. Or if you used the CLI, where I accidentally left it defaulting to adding 0.25 seconds of silence.

[DiamondGlassDrill]

File is still empty. What I do is typing a text message and select a build in speaker, than I click generate it, it produces the .npz and an .mp3 but the .mp3 is empty.

[JonathanFly]

File is still empty. What I do is typing a text message and select a build in speaker, than I click generate it, it produces the .npz and an .mp3 but the .mp3 is empty.

Can you try setting the output format to 'wav' just to check that works at least?

Also just to double check you got the update, you typed git pull from the command line, same place you start python bark_webui.py?