Question: Can this software separate a relatively long track with two voices?

[phalexo]

I have tried "spleeter" and "speechbrain" with two different models. SpeechBrain works for a while but the throws an exception with size mismatch for tensors. spleeter simply fails to do anything. It does create rather large files but they don't contain anything, except a little noise in a couple places.

My current track is about 11.8 minutes, but longer would be better obviously.

[zhenhaoge]

If the spleeter fails to do anything, then you didn't set it right. Please make sure your path to call the model is correct.
The default setting of the spleeter will generate up to 10-minutes separated audio, but the 10-min threshold can be changed.

[phalexo]

Thanks. I have found a different approach.

…

On Fri, Jan 17, 2025 at 5:21 PM zhenhaoge ***@***.***> wrote: If the spleeter fails to do anything, then you didn't set it right. Please make sure your path to call the model is correct. The default setting of the spleeter will generate up to 10-minutes separated audio, but the 10-min threshold can be changed. — Reply to this email directly, view it on GitHub <#49 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABDD3ZLGGULO2QSNSSLKVM32LF7AHAVCNFSM6AAAAABPKNQC3GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOJZGMYDOOJVGE> . You are receiving this because you authored the thread.Message ID: ***@***.***>