Open mthrok opened 4 years ago
This is very cool. espnet also basically supports most of these examples, and this effort can be aligned with our interest as well.
One only comment is that Callhome is used for diarization, not for ASR. Also, another suggestion is to consider the pitch feature, which is effective not only for tonal languages but for the other languages.
Hi @sw005320
Thanks for the comment.
One only comment is that Callhome is used for diarization, not for ASR.
Noted. will reflect it.
Also, another suggestion is to consider the pitch feature, which is effective not only for tonal languages but for the other languages.
Do you mean adding an equivalent implementation of compute-kaldi-pitch-feats
to torchaudio
?
Do you mean adding an equivalent implementation of
compute-kaldi-pitch-feats
totorchaudio
?
Yes, exactly. That would be very cool, but it should not be a very high priority though.
We found that the pitch feature always improved the performance for several tonal languages (e.g., Chinese), and did not degrade the performance for the other languages. So, espnet1 decided to use log Mel filterbank + pitch features as default. However, the pitch feature extraction is rather complicated, and we had some difficulties in making this pitch feature extraction fully written by torch functions. So, espnet2 decided to only use log Mel filterbank features, instead. We still observe a slight degradation of the ASR performance, but that can be mitigated by some tuning. We're now moving to espnet2 so we don't need it in the long term, but probably it is quite beneficial for the short term or people keep to use espnet1.
Do you mean adding an equivalent implementation of
compute-kaldi-pitch-feats
totorchaudio
?Yes, exactly. That would be very cool, but it should not be a very high priority though.
We found that the pitch feature always improved the performance for several tonal languages (e.g., Chinese), and did not degrade the performance for the other languages. So, espnet1 decided to use log Mel filterbank + pitch features as default. However, the pitch feature extraction is rather complicated, and we had some difficulties in making this pitch feature extraction fully written by torch functions. So, espnet2 decided to only use log Mel filterbank features, instead. We still observe a slight degradation of the ASR performance, but that can be mitigated by some tuning. We're now moving to espnet2 so we don't need it in the long term, but probably it is quite beneficial for the short term or people keep to use espnet1.
I see. I created the issue https://github.com/pytorch/audio/issues/686 to track this.
@sw005320
espnet also basically supports most of these examples, and this effort can be aligned with our interest as well.
If spent has similar testing, could you give me the pointer to it??
Sorry, I did not notice it. We did not specifically check the feature extraction compatibility test. But we’ll also use torchaudio features and will have make such a test in the near future.
@mthrok since this is so related to https://github.com/pytorch/audio/issues/689, I would like to work on this as well.
@mthrok since this is so related to #689, I would like to work on this as well.
Sure! Thanks for signing up!
Test parameters for Kaldi fbank were generated using this script, but this does not necessarily generate values in valid range, so we need to revise them.
First, we need to test the parity of default values.
Second, we should cover values found in Kaldi examples. such as
Then we can also add perturbation
--frame-length
and--frame-shift
: integer value around 10 - 50 [ms] if perturbing--preemphasis-coefficient
: less than 1.0, around 0.90 - 0.99--low-freq
: around 20 - 50 (and less than--high-freq
)See also https://github.com/pytorch/audio/pull/672