Source code for selected algorithms developed by
the CMU Robust Speech Group
- Robust
Speech Group, Carnegie Mellon University
- Miscellaneous Audio Utilities (Pyaudlib)
for reading, writing, and manipulating wav files in
Python
- Magnitude-Spectrum
Recurrent Neural Networks (MRNN) for ASR in
noise
- The STRFNet
system
- SNR-Based Wiener Enhancement usng
Recurrent Neural Networks (SWRNN) for ASR in noise
- Non-Negative Matrix Factorization (NMF)
for ASR in reverberation
- Phase Difference Channel
Weighting (PDCW) spatial separation
- Power Normalized Cepstral Coefficients
(PNCC) for robust feature extraction
- Compensation for the effects of demodulation
frequency mismatch
- in single-sideband
(SSB) communication
- Suppression of Slowly-Varying Components and
the Falling Edge (SSF)
- for onset enhancment in reverberation
- Vector Taylor Series (VTS) for compensation for additive noise
and linear filtering
- Waveform Amplitude Distribution Analysis (WADA) for SNR
estimation