Smart Channel strip using Neural audio processing
The proposed PhD focuses on developing novel techniques for a smart channel strip using deep learning techniques. The smart channel strip would allow musicians with no audio engineering knowledge to use a complex mixing console using their own musical terms aided by Artificial Intelligence. The concept is to offer audio effects typically included in mixer channel strips such as Level, Pan, Gate, Compressors, EQ, as well as Env Shaper, Saturator, Limiter, etc., which could be controlled using musical parameters instead of technical parameters. The approach can make use of the content of the current track associated with a particular mixer channel, but may also rely on the whole project (cross-adaptive processing across tracks) to understand its context. The research is not targeting fully automatic mixing, but it aims to develop a technology that would be assisting a user toward improving the mix, without removing the feeling of being rewarded for the work. Potentially, this research could be extended to include techniques for audio quality improvement or restoration purposes.
C4DM theme affiliation: