Implementation of the Envelope modulation spectrum for speech rhythm analysis
Project description
RhythmoScope - Speech Rhythm Modulation Spectrum
RhythmoScope is a Python library for automatic speech rhythm modelisation. This project rely on the Envelope Modulation Spectrum (EMS) for analysing the main regularities of speech at different levels. This library is the result of my PhD thesis under the supervision of Corine Astésano and Jérôme Farinas.
🛠 Installation
Rhythmoscope should work with Python 3.9 and above. The library can be installed through the following pip
command:
pip install rhythmoscope
🔬 Basic usage
As an example, we'll extract and plot the EMS of a speech signal from a .wav file:
from rhythmoscope.ems import EMSExtractor
from rhythmoscope.envelope import LowPassEnvelope
Envelope = LowPassEnvelope(cut_frequency=10, order=3, initial_bandpass=(300, 1000)) # Define envelope extractor
Extractor = EMSExtractor(Envelope=Envelope, min_freq=0, max_freq=10) # Define EMS extractor parameters
EMS = Extractor.from_file("example.wav", start=0, end=4) # Extract EMS on the first 4 seconds of the audio
fig = EMS.plot()
fig.show()
It produce the following output which represent the Envelope Modulation Spectrum of a signal:
🔗 Related work
- Rhythm Formant Analysis from Dafydd Gibbon
- Temporal Modulation Spectrum Toolbox (Matlab code) from Léo Varnet
💬 Citation
If RhythmoScope has been useful to you, and you would like to cite, please refer to my PhD thesis:
@phdthesis{vaysse2023thesis,
TITLE = {{Caract{\'e}risation automatique du rythme de la parole : application aux cancers des voies a{\'e}ro-digestives sup{\'e}rieures et {\`a} la maladie de Parkinson}},
AUTHOR = {Vaysse, Robin},
URL = {https://theses.hal.science/tel-04198849},
NUMBER = {2023TOU30062},
SCHOOL = {{Universit{\'e} Paul Sabatier - Toulouse III}},
YEAR = {2023},
MONTH = Mar,
TYPE = {Theses},
PDF = {https://theses.hal.science/tel-04198849/file/2023TOU30062b.pdf},
HAL_ID = {tel-04198849},
HAL_VERSION = {v1},
}
📝 License
RhythmoScope is a free and open-source software licensed under the 3-clause BSD license.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for rhythmoscope-0.1.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | eff98e82988aae6d54c0c783b239f4acbdbd642cd595aa5255f40aa1dac8513a |
|
MD5 | a90ccb35479b66e3e3e1a1a8385c024d |
|
BLAKE2b-256 | 2d552c77e32c7bf173e564fe9c6c13795b0fc945cfb4da25c3f88b1f845137a1 |