fgnt / pb_chime5

Speech enhancement system for the CHiME-5 dinner party scenario
MIT License
108 stars 34 forks source link

new format for CHiME-6 #7

Closed sw005320 closed 4 years ago

sw005320 commented 4 years ago

hi,

Now, we're working on the CHiME-6 baseline, and try to include the GSS in the baseline recipe. In the CHiME-6, we plan to synchronize the audio files across arrays and use the same time across them. The audio files are modified accordingly and the JSON format will be changed as follows:

Could you make an option to follow such a format?

sw005320 commented 4 years ago

Added @jonbarker68 for this discussion.

boeddeker commented 4 years ago

Hi, sure, I can add such an option. This makes the code at many positions simpler.

sw005320 commented 4 years ago

Many thanks!

sw005320 commented 4 years ago

I just invited @boeddeker to our private repo. In the repo, we already have the version for the new format and test it with the beamformit etc. are working but not for GSS and want your help.

boeddeker commented 4 years ago

Thanks for the invitation, I got it.

I started to support your planned json style. I created a dummy json with your proposed format and adjusted the code to support it. The pb_chime5.database.chime5.create_json got a --chime6 flag and the pb_chime5.scripts.kaldi_run got a chime6=True option.

The test run worked in the chime6 mode.

I am not sure, if I find the time tomorrow to test it with your repository. On Friday I have some more time to test it.

sw005320 commented 4 years ago

This is very helpful. I think I can test your PR on my side at least.

boeddeker commented 4 years ago

Thanks. The PR is now also merged.

Hint: The default path for our json in the CHiME5 mode changed to cache/chime6.json (i.e. 5 -> 6)

sw005320 commented 4 years ago

This change is incorporated in our ongoing chime6 recipe and we confirmed that it is working.