aminEdraki / wstmi

wSTMI: A speech intelligibility prediction algorithm for noisy and processed speech

wSTMI

This is a MATLAB implementation of wSTMI, the speech intelligibility prediction algorithm proposed here. The algorithm takes the time-aligned clean and degraded/processed speech signals as inputs. The output d is expected to have a monotonically increasing relation with speech intelligibility. Please refer to Sec. V of the reference cited below for guidance on interpreting algorithm output.

Installation

Before using the wSTMI function, the utils folder has to be added to the search path for the current MATLAB session.

Usage

The function wstmi takes three inputs:

d = wstmi(clean_speech, degraded_speech, sampling_frequency);

• clean_speech: An array containing a single-channel clean (reference) speech signal.
• degraded_speech: An array containing a single-channel degraded/processed speech signal.
• sampling_frequency: The sampling frequency of the input signals in Hz.

Note that the clean and degraded speech signals must be time-aligned and of the same length.

Citing wSTMI

If you use wSTMI, please cite the reference below:

@article{edraki2020speech,
  title={Speech Intelligibility Prediction Using Spectro-Temporal Modulation Analysis},
  author={Edraki, Amin and Chan, Wai-Yip and Jensen, Jesper and Fogerty, Daniel},
  journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing},
  volume={29},
  pages={210--225},
  year={2020},
  publisher={IEEE}
}

License

GNU GPLv3