search

Enny1991 / beamformers

Easy to use Beamformers for multi-channel speech separation/enhancement
MIT License
184 stars 48 forks source link

using length of optional target for reconstruction #2

Closed 5ke closed 1 year ago

5ke commented 3 years ago

Thank you for making your repository available.

For the MVDR and MSNR beamformers (in beamformers.py), the target is optional, but at the same time it is required to reconstruct the wave form. Wouldn't it be an option to obtain the length from the mixture instead?

Enny1991 commented 3 years ago

Yes absolutely, this is actually a bug that I solved already locally, let me push it.

5ke commented 3 years ago

Thanks!