%Aigaion2 BibTeX export from Idiap Publications %Sunday 22 December 2024 04:01:44 AM @INPROCEEDINGS{Cernak_ICASSP15_2015, author = {Cernak, Milos and Potard, Blaise and Garner, Philip N.}, projects = {Idiap, armasuisse}, month = apr, title = {Phonological Vocoding Using Artificial Neural Networks}, booktitle = {IEEE 40th International Conference on Acoustics, Speech and Signal Processing (ICASSP)}, year = {2015}, pages = {4844-4848}, publisher = {IEEE}, location = {Brisbane, Australia}, doi = {10.1109/ICASSP.2015.7178891}, crossref = {Cernak_Idiap-RR-04-2015}, abstract = {We investigate a vocoder based on artificial neural networks using a phonological speech representation. Speech decomposition is based on the phonological encoders, realised as neural network classifiers, that are trained for a particular language. The speech reconstruction process involves using a Deep Neural Network (DNN) to map phonological features posteriors to speech parameters -- line spectra and glottal signal parameters -- followed by LPC resynthesis. This DNN is trained on a target voice without transcriptions, in a semi-supervised manner. Both encoder and decoder are based on neural networks and thus the vocoding is achieved using a simple fast forward pass. An experiment with French vocoding and a target male voice trained on 21 hour long audio book is presented. An application of the phonological vocoder to low bit rate speech coding is shown, where transmitted phonological posteriors are pruned and quantized. The vocoder with scalar quantization operates at 1 kbps, with potential for lower bit-rate.}, pdf = {https://publications.idiap.ch/attachments/papers/2015/Cernak_ICASSP15_2015.pdf} } crossreferenced publications: @TECHREPORT{Cernak_Idiap-RR-04-2015, author = {Cernak, Milos and Potard, Blaise and Garner, Philip N.}, keywords = {low bit rate speech coding, Parametric vocoding, phonology}, projects = {Idiap, armasuisse}, month = {2}, title = {Phonological vocoding using artificial neural networks}, type = {Idiap-RR}, number = {Idiap-RR-04-2015}, year = {2015}, institution = {Idiap}, abstract = {We investigate a vocoder based on artificial neural networks using a phonological speech representation. Speech decomposition is based on the phonological encoders, realised as neural network classifiers, that are trained for a particular language. The speech reconstruction process involves using a Deep Neural Network (DNN) to map phonological features posteriors to speech parameters -- line spectra and glottal signal parameters -- followed by LPC resynthesis. This DNN is trained on a target voice without transcriptions, in a semi-supervised manner. Both encoder and decoder are based on neural networks and thus the vocoding is achieved using a simple fast forward pass. An experiment with French vocoding and a target male voice trained on 21 hour long audio book is presented. An application of the phonological vocoder to low bit rate speech coding is shown, where transmitted phonological posteriors are pruned and quantized. The vocoder with scalar quantization operates at 1 kbps, with potential for lower bit-rate.}, pdf = {https://publications.idiap.ch/attachments/reports/2014/Cernak_Idiap-RR-04-2015.pdf} }