Vocal Fry and other Metrics
Resemble Speech Tools
This package contains collated tools for quantifying issues in speech clips.
The repo currently has support for detecting vocal fry (H2H1 metric proposed by Kane-Drugman):
from speechtools.features import get_creak_features
get_creak_features(waveform, fs, f0_min=20, f0_max=500, window=1600, hop=160, use_fixed_windows=False, return_frames=False
use_fixed_frames: This needs to be
Trueif user wants to override internally used
False, the default window and hop sizes specified by the KaneDrugman papers are used for analysis.
Falsethe resulting metric is interpolated to be of
There are also a few private functions bundled in for signal processing under
- Splitting up the signal into chunks of
- Obtaining frame level LPC residuals
- Kane, J., Drugman, T., Gobl, C., (2013) "Improved automatic detection of creak", 27(4), pp. 1028-1047, Computer Speech and Language.
- Drugman, T., Kane, J., Gobl, C. (2012) "Resonator-based creaky voice detection", Proceedings of Interspeech.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Hashes for speech-analysis-tools-0.2.tar.gz