This priority refers to new AI methods for audio separation and voice identification, voice synthesis and speech enhancement including that in noisy and reverberant environments. This priority extends beyond voice to musical, textural, synthesis and environmental sounds, e.g. extracting accurate music/singing information and converting it into appropriate control protocols, such as MIDI. The use of biosensors that can collect the voice generation information directly on the source (e.g. vocal folds), in combination with AI techniques, can lead to much better understanding of the oldest musical instrument, the human voice, significantly novel solutions in the concept of human-computer interaction.