lhotse icon indicating copy to clipboard operation
lhotse copied to clipboard

Misleading feature type

Open csukuangfj opened this issue 2 years ago • 3 comments

In https://github.com/lhotse-speech/lhotse/blob/d9c4141319adb39f64684c762aa541467d25f7fc/lhotse/kaldi.py#L144-L145

It uses kaldiio as the feature type.

However, https://github.com/lhotse-speech/lhotse/blob/d9c4141319adb39f64684c762aa541467d25f7fc/lhotse/features/base.py#L407-L408 says possible types are fbank, mfcc, etc.

csukuangfj avatar Feb 15 '22 02:02 csukuangfj

Good catch! I think I hard-coded kaldiio in there because it might be tricky to determine whether the features imported from Kaldi are fbanks or mfccs (and we'd want to define computing energies and mixing differently for them). Any ideas?

EDIT: technically they can also be fbank+pitch, mfcc+pitch, etc... seems like a rabbit hole.

pzelasko avatar Feb 15 '22 13:02 pzelasko

How about making users provide it?

csukuangfj avatar Feb 15 '22 15:02 csukuangfj

+1

pzelasko avatar Feb 17 '22 01:02 pzelasko