Speech coding is an essential technology for digital cellular communications, voice over IP, and video conferencing systems. For more than 25 years, the main approach to speech these applications has been block-based analysis-by-synthesis linear predictive coding. An alternative that less successful sample-by-sample tree of speech. We reformulate this latter as a multistage reinforcement learni...

Background work in various levels of speech coding is reviewed, including unconstrained coding and recognition-synthesis approaches that assume the signal is speech. A pilot project in HMM-TTS based speech coding is then described, in which a comparison with harmonic plus noise modelling is also done. Results of the demonstration project including samples of speech under various transmission si...