Phonological features

Phonological features

Created on Jun 24 2020

@author: J. C. Vasquez-Correa

class phonological.Phonological

Compute phonological features from continuous speech files.

18 descriptors are computed, bases on 18 different phonological classes from the phonet toolkit https://phonet.readthedocs.io/en/latest/?badge=latest

It computes the phonological log-likelihood ratio features from phonet

Static or dynamic matrices can be computed:

Static matrix is formed with 108 features formed with (18 descriptors) x (6 functionals: mean, std, skewness, kurtosis, max, min)

Dynamic matrix is formed with the 18 descriptors computed for frames of 25 ms with a time-shift of 10 ms.

Script is called as follows

>>> python phonological.py <file_or_folder_audio> <file_features> <static (true or false)> <plots (true or false)> <format (csv, txt, npy, kaldi, torch)>

Examples command line:

>>> python phonological.py "../audios/001_ddk1_PCGITA.wav" "phonologicalfeaturesAst.txt" "true" "true" "txt"
>>> python phonological.py "../audios/001_ddk1_PCGITA.wav" "phonologicalfeaturesUst.csv" "true" "true" "csv"
>>> python phonological.py "../audios/001_ddk1_PCGITA.wav" "phonologicalfeaturesUdyn.pt" "false" "true" "torch"

>>> python phonological.py "../audios/" "phonologicalfeaturesst.txt" "true" "false" "txt"
>>> python phonological.py "../audios/" "phonologicalfeaturesst.csv" "true" "false" "csv"
>>> python phonological.py "../audios/" "phonologicalfeaturesdyn.pt" "false" "false" "torch"
>>> python phonological.py "../audios/" "phonologicalfeaturesdyn.csv" "false" "false" "csv"

Examples directly in Python

>>> phonological=Phonological()
>>> file_audio="../audios/001_ddk1_PCGITA.wav"
>>> features1=phonological.extract_features_file(file_audio, static=True, plots=True, fmt="npy")
>>> features2=phonological.extract_features_file(file_audio, static=True, plots=True, fmt="dataframe")
>>> features3=phonological.extract_features_file(file_audio, static=False, plots=True, fmt="torch")
>>> phonological.extract_features_file(file_audio, static=False, plots=False, fmt="kaldi", kaldi_file="./test")
extract_features_file(audio, static=True, plots=False, fmt='npy', kaldi_file='')

Extract the phonological features from an audio file

Parameters:
  • audio – .wav audio file.
  • static – whether to compute and return statistic functionals over the feature matrix, or return the feature matrix computed over frames
  • plots – timeshift to extract the features
  • fmt – format to return the features (npy, dataframe, torch, kaldi)
  • kaldi_file – file to store kaldi features, only valid when fmt==”kaldi”
Returns:

features computed from the audio file.

 
>>> phonological=Phonological()
>>> file_audio="../audios/001_ddk1_PCGITA.wav"
>>> features1=phonological.extract_features_file(file_audio, static=True, plots=True, fmt="npy")
>>> features2=phonological.extract_features_file(file_audio, static=True, plots=True, fmt="dataframe")
>>> features3=phonological.extract_features_file(file_audio, static=False, plots=True, fmt="torch")
>>> phonological.extract_features_file(file_audio, static=False, plots=False, fmt="kaldi", kaldi_file="./test")

>>> phonological=Phonological()
>>> path_audio="../audios/"
>>> features1=phonological.extract_features_path(path_audio, static=True, plots=False, fmt="npy")
>>> features2=phonological.extract_features_path(path_audio, static=True, plots=False, fmt="csv")
>>> features3=phonological.extract_features_path(path_audio, static=False, plots=True, fmt="torch")
>>> phonological.extract_features_path(path_audio, static=False, plots=False, fmt="kaldi", kaldi_file="./test.ark")
extract_features_path(path_audio, static=True, plots=False, fmt='npy', kaldi_file='')

Extract the phonological features for audios inside a path

Parameters:
  • path_audio – directory with (.wav) audio files inside, sampled at 16 kHz
  • static – whether to compute and return statistic functionals over the feature matrix, or return the feature matrix computed over frames
  • plots – timeshift to extract the features
  • fmt – format to return the features (npy, dataframe, torch, kaldi)
  • kaldi_file – file to store kaldifeatures, only valid when fmt==”kaldi”
Returns:

features computed from the audio file.

 
>>> phonological=Phonological()
>>> path_audio="../audios/"
>>> features1=phonological.extract_features_path(path_audio, static=True, plots=False, fmt="npy")
>>> features2=phonological.extract_features_path(path_audio, static=True, plots=False, fmt="csv")
>>> features3=phonological.extract_features_path(path_audio, static=False, plots=True, fmt="torch")
>>> phonological.extract_features_path(path_audio, static=False, plots=False, fmt="kaldi", kaldi_file="./test.ark")