online TF outputs NAN - Githubissues

fgnt / nara_wpe

Different implementations of "Weighted Prediction Error" for speech dereverberation

MIT License

494 stars 164 forks source link

online TF outputs NAN #68

Open MMingabc opened 2 years ago

MMingabc commented 2 years ago

Hi, I am trying online WPE with my own 2-channel data. After 2 or 3 iterations outputs became NAN. The official example works properly. How do I fix this problem.

Many thanks!

boeddeker commented 2 years ago

Hi, difficult to say, what went wrong. I need more information to get an idea. Have you tried offline WPE. What ist the difference of your data, compared to the example? Does it contain speech? Large/small reverberation, noise, ...

MMingabc commented 2 years ago

Hi, boeddeker I am using FastRIR to generate Reverberated speech. The simulated room is 6x8x3 and the source speech was recorded in a professional studio. I would say the reverberation is pretty strong. The offline version is super stable and its performance is impressive.