Giters

breizhn / DTLN-aec

This Repostory contains the pretrained DTLN-aec model for real-time acoustic echo cancellation.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

up-sampling the phone data to 16k and then down-sampling to 8k Model effect will be better

JuneRen opened this issue · comments

commented

Thank you very much for your work. We first downsample the 16k AEC data set to 8k and train the model; then apply it to 8k phone data. The effect is not as good as up-sampling the phone data to 16k and then down-sampling to 8k. , What is the reason for this? Thank you

commented

Did you change the frame length and shift according to the sample rate? (Basically dividing by 2 since the sample rate is divided by 2)

commented

Did you change the frame length and shift according to the sample rate? (Basically dividing by 2 since the sample rate is divided by 2)

Thanks, the parameters are all divided by 2. Is the downsampling we used sox related to sox downsampling?