msb-public
/
PaddleSpeech
mirror of https://github.com/PaddlePaddle/PaddleSpeech
1
0
Code Issues Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'develop'
${ noResults }
PaddleSpeech/third_party/python_kaldi_features
History
Jackwaterveg 2bccde3def
update the version of ctcdecoders and feat,test=doc_fix (#1155) 		3 years ago
..
docs E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
python_speech_features E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
test E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
.gitignore E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
LICENSE E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
MANIFEST E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
README.rst E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
english.wav E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
example.py E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
requirements.txt E2E/Streaming Transformer/Conformer ASR (#578) 3 years ago
setup.py update the version of ctcdecoders and feat,test=doc_fix (#1155) 3 years ago

README.rst
Escape

This file contains invisible Unicode characters!

This file contains invisible Unicode characters that may be processed differently from what appears below. If your use case is intentional and legitimate, you can safely ignore this warning. Use the Escape button to reveal hidden characters.




forked from `<https://github.com/jameslyons/python_speech_features>`_



check the readme therein for the usages



It has been modified to produce the same results as with the compute-mfcc-feats and compute-fbank-feats (check their default parameters first) commands in Kaldi.

 

-------------------------------



The compute-mfcc-feats pipeline:



src/featbin/Compute-mfcc-feats.cc

    

    Mfcc mfcc(mfcc_opts)  --> src/feat/Feature-mfcc.h

    

                                 struct MfccOptions

                                 

                                 typedef OfflineFeatureTpl<MfccComputer> Mfcc --> src/feat/Feature-common.h

           

                                 MfccComputer()  --> src/feat/Feature-mfcc.cc

                                 

                                                         ComputeDctMatrix()  --> src/matrix/Matrix-functions.cc

                                                         

                                                         ComputeLifterCoeffs()  --> src/feat/Mel-computations.cc

  

    

    for each utterance:

    mfcc.ComputeFeatures()



src/feat/Feature-common-inl.h



    OfflineFeatureTpl<F>::ComputeFeatures()

    

        Compute()

        

            ExtractWindow()  --> src/feat/Feature-window.cc

                                     

                                     ProcessWindow()

                                         

                                         Dither, remove_dc_offset, log_energy_pre_window, Preemphasize, window

            

            computer_.Compute() --> src/feat/Feature-mfcc.cc

               

                                      MfccComputer::Compute()

                                      

                                          const MelBanks &mel_banks --> Mel-computations.cc

                                          

                                          srfft_

                                        

                                          ComputerPowerSpectrum()

                                          

                                          mel_banks.Compute()

                                          

                                          mel_energies_.ApplyLog()

                                          

                                          dct, cepstral_lifter