Streaming Target Sound Extraction

Streaming Target Sound Extraction is a variant of the Target Sound Extraction task, focusing on real-time audio processing. This task requires the algorithm to process audio segments shorter than 20 milliseconds at each time step, using only past data to ensure low latency and causality. Its goal is to accurately extract the target sound in a real-time streaming environment, which has significant application value, such as real-time speech enhancement, noise suppression, and audio monitoring.