Work-in-Progress: Adaptive Scheduling with Approximate Computing for Audio Graphs

Abstract

Interactive music systems are highly dynamic systems that combine audio processing and control in real-time. They often have to work on soft real-time platforms, where no stringent real-time guarantees can be upheld. We present here an overhead-aware online degradation algorithm that finds a tradeoff between quality and lateness for the processing nodes of a dynamic audio graph.

Publication
2017 IEEE Real-Time Systems Symposium (RTSS)