Satoshi Tamura Kento Horio Hajime Endo Satoru Hayamizu Tomoki Toda Audio-visual Voice Conversion Using Deep Canonical Correlation Analysis for Deep Bottleneck Features. 2469-2473 2018 INTERSPEECH https://doi.org/10.21437/Interspeech.2018-2286 conf/interspeech/2018 db/conf/interspeech/interspeech2018.html#TamuraHEHT18