search

posenhuang / deeplearningsourceseparation

Deep Recurrent Neural Networks for Source Separation
Other
365 stars 135 forks source link

could it be used in real time ? #29

Closed ucasiggcas closed 5 years ago

ucasiggcas commented 5 years ago

Dear, I want to use the method to solve the howling in RTC-Real Time Communication, so could I separate the noise and voice frame by frame ?

any reference will be appreciated.

thx

posenhuang commented 5 years ago

I believe it is possible! I think it might require writing inference code more efficiently and using hardware acceleration.

ucasiggcas commented 5 years ago

So I want to know that whether the input signal length can be modified and which parameter should I change if I need. thx

ucasiggcas commented 5 years ago

maybe you misunderstand the howling, it's very strong in RTC, like the pic down, image the up is howling and the down voice

posenhuang commented 5 years ago

I'm not familiar with this application, but it looks doable when looking at the spectrogram. You might need to check "Work on your data" section in the readme.