Speech Processing

Speech coding
Whisper Labs has been involved in speech coding research for more than 10 years. In particular we have investigated Waveform Interpolation (WI) speech coding for low bit rate, perfect reconstruction, scalable and wideband speech coding.

Speech Signal Separation and Enhancement
Our research into the separation of mixed speech signals has led to the development of new techniques derived from the TIFROM and TIFCORR algorithms. Our approach combines speech modelling techniques with Blind Source Separation (BSS) techniques and can perform separation in real time.