Article details

Research area
Speech enhancement

Speech Communication; 11. ITG Symposium; Proceedings of


Timo Matheja, Markus Buck

Detection of local disturbances and simultaneously active speakers for distributed speaker-dedicated microphones in cars


For automotive hands-free and speech recognition applications, distributed microphones are often mounted in the car where each of the speakers has a dedicated microphone close to his position. To provide additional control information for further speech enhancement, it is often advantageous to distinguish between the activity of the different passengers. In this contribution speaker activity is identified by the evaluation of the ratios between the powers of multiple speaker-dedicated microphone signals while further acoustic events are differentiated from single speaker’s activity. An effective algorithm based on the exploitation of the expected range of power ratio values is presented that is able to detect local disturbances like scratch noise at the microphones. Furthermore, situations are identified where multiple passengers speak at the same time. Besides some examples, it can be shown that the proposed method for local disturbance detection is able to identify such noise within realistic driving conditions.

Read/download now