In this paper, we investigated the effect of competing noise source on the intelligibility of target speech. We are currently investigating on the possibility of a network-based conference system in which individual users participating from stand-alone PCs share common virtual space. We are currently focusing on the acoustic aspects of this conferencing system. The intelligibility improvement of the primary speech when the competing sources are localized away on the horizontal plane was investigated. We used an impulse response measured with the KEMAR dummy head, as well as impulse response measured for each subject to localize both target speech and competing multi-speaker noise at various angles in the horizontal plane. The speech intelligibility of the target speech was evaluated using the Diagnostic Rhyme Test (DRT). The target speech was placed directly in front of the listener, and a single competing source was placed on the horizontal plane at various azimuths and distances. We also compared the intelligibility when these sound sources were actually generated from loud speakers located at target and competing source positions in actual acoustic space. Individual HRTFs show slightly better intelligibility than with KEMAR HRTFs, but the difference is insignificant if target speech to competing source ratio is high. Intelligibility in actual acoustic space still clearly outperforms both types of HRTFs especially when the competing sound source is significant.
展开▼
