The PhonNet is end-to-end neural network (NN) based very low bit rate speech coder.  It is designed as a composition of deep and spiking neural networks: a bank of phonological analysers at the transmitter, and a phonological synthesizer at the receiver, both realised as deep NNs, and a spiking NN as an incremental and robust encoder of syllable boundaries for coding of continuous fundamental frequency.  The PhonNet operates real-time on bit rates about 360 bits/sec, and achieves quality of 2.3 mean opinion score.

The PhonVoc is a cascaded deep neural network composed of speech analyser and synthesizer that use a shared phonetic and/or phonological speech representation. The PhonVoc is available as open-source software available at https://github.com/idiap/phonvoc, distributed under a BSD 3-Clause License.

The PhonVoc is a cascaded deep neural network composed of speech analyser and synthesizer that use a shared phonetic and/or phonological speech representation. The PhonVoc is available as open-source software available at https://github.com/idiap/phonvoc, distributed under a BSD 3-Clause License.