Multiple sound source tracking and utterance interval detection for planar microphone array

摘要

Problem to be solved: to provide a system and method for detecting multiple sound source tracking and multistream utterance interval in which a calculation quantity by a microphone array is reduced in order to detect and process a target speech signal in a multistream voice environment.The method isA multi tone speech signal is defined by a microphone pair from a microphone arrayStaged minimum dispersion (stmv) beamformer based on TDOA locus information constructed in multidimensional spaceThe audio track is updated based on the TDOA detection data calculated for the multi source audio signalBased on TDOA detection data, we construct utterance interval detection (VAD) data for each of a plurality of audio tracks.One or more highlighted multiple sound source audio signals are generated using the updated plurality of audio tracks and the configured VAD data.Diagram