default audio sample loading breaks on windows
sildater opened this issue · comments
on windows I get the following error when exporting a model via save_neutone_model:
RuntimeError: Error opening <_io.BytesIO object at 0x0000019212E07C20>: Format not recognised.
It stems from loading the default audio mp3s in lines 129, 132, 135 in audio.py, the bytesIO object has no format and can't be read by soundfile (line 87). It works on ubuntu.
@bogdanteleaga does this look familiar when you were making the audio sample changes? Perhaps there's a quick fix here
@sildater thanks for raising this issue! While we look into a fix, you should still be able to export your model by setting dump_sample=False and submission=False when calling the save_neutone_model method.
@sildater Hello! Sorry for the delay, I just tried this out on a Windows machine right now and it seems like it's working. Can you check you have the latest soundfile (I pulled 0.12.1 on a fresh install) and you can use it to read mp3 files with soundfile.read?
Closing due to no reply.
Hi! Thanks for the reply and sorry for the late answer! I could fix it with a fresh install of soundfile as you suggested.