-
Notifications
You must be signed in to change notification settings - Fork 2.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add ConferencingSpeech 2021 recipe to egs2 #4192
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
some minor comments
egs2/README.md
Outdated
@@ -19,6 +19,7 @@ See: https://espnet.github.io/espnet/espnet2_tutorial.html#recipes-using-espnet2 | |||
| chime4 | The 4th CHiME Speech Separation and Recognition Challenge | ASR/Multichannel ASR | ENG | http://spandh.dcs.shef.ac.uk/chime_challenge/chime2016/ | | | |||
| cmu_indic | CMU INDIC | TTS | 7 languages | http://festvox.org/cmu_indic/ | | | |||
| commonvoice | The Mozilla Common Voice | ASR | 13 languages | https://voice.mozilla.org/datasets | | | |||
| conferencingspeech21 | Far-field Multi-channel Speech Enhancement Challenge for Video Conferencing (ConferencingSpeech 2021) | SE | 2 languages | https://tea-lab.qq.com/conferencingspeech-2021 | | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
2 languages -> ENG, CMN
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
OK. This is more clear.
@LiChenda, can I ask you to review this PR? |
@sw005320 Sure, I can review this PR. |
path_noise = noise_data[path_noise] | ||
out.write( | ||
f"{path_clean} {start_time} {path_noise} " | ||
f"/path/{args.tag}/{path_rir}.wav {snr} {scale}\n" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Does "/path/" exist on filesystem?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is a placeholder. We don't actually use this path for simulation. Because this python script is intended for generating the simulation configuration from an already generated data.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh, I see.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM!
Codecov Report
@@ Coverage Diff @@
## master #4192 +/- ##
=======================================
Coverage 80.71% 80.71%
=======================================
Files 453 453
Lines 39575 39575
=======================================
Hits 31944 31944
Misses 7631 7631
Flags with carried forward coverage won't be shown. Click here to find out more. 📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more |
|
|
Thanks, @Emrys365! |
I was not sure if it is appropriate to upload my local model. Because it was trained based on a local branch of espnet, where a customized Preprocessor was defined and used in espnet2/train/preprocessor.py.
|
I see. |
ConferencingSpeech 2021 is a far-field multi-channel speech enhancement challenge.
The training data includes several English and Mandarin corpora, while test data are semi-real and real recordings in Mandarin.
This PR adds the enh recipe for task2 (Multi-channel speech enhancement with multiple distributed microphone arrays).