Georgi Dzhambazov

reduce dependecy on htkmfc

1

make sure extracting MFCC with essentia same as damp model: add preempahsis (or recreate model without preemphasis ) add cepstral mean normalization

isolate only the code from https://github.com/ztane/python-Levenshtein for the method Levenshtein.distance, compile and add as a static library in code.

remove hard coded sampling rate

1

for example in src.align.FeatureExtractor.FeatureExtractor._extractMFCCs it is hard coded

move this src.align.FeatureExtractor.getTimeStamps to Utilz. At the end of Utilz there are methods which repeat this code.

The code for expansion of silence word could be reduced [here](https://github.com/georgid/AlignmentDuration/blob/noteOnsets/src/for_english/CMUWord.py#L36)

Store the char, not only the pinynin in the mandarinSyllable class. See initi of the mandarinSyllable class [here](https://github.com/georgid/AlignmentDuration/blob/noteOnsets/src/for_jingju/lyricsParser.py#L221). Split chinese characters [in this method](https://github.com/georgid/AlignmentDuration/blob/noteOnsets/src/align/GenericRecording.py) and give as argument.

in file LyricsParsing

when WITH_SHORT_PAUSES = 1

1

we got error: last state for word _SAZ_ is not sp. Sorry - not implemented. The problem is it is that I removed sp from SAZ so that it is...

there is kimseye-specific code in align.LyricsAligner.LyricsAligner.alignLyricsSection remove it.

means, covars, weights

1

means, covars, weights, should not be in the constructor of _ContinuousHMM, because they are observation probabilities.