Abstract
Network latency coupled with the lack of audiovisual cues in multi-party online conferencing can lead users to experience speech collisions when two or more parties start speaking at roughly the same time. Attempts to repair speech collisions typically result in a period of silence and attempts to break the silence may result in further speech collisions. This disclosure describes techniques to automatically detect and correct speech collisions in a multi-party conference by breaking ties among simultaneous speakers by modulating the amplitude and/or shifting the time of the audio forward for any user who starts speaking after the first speaker. As the first speaker continues speaking without pausing to repair the speech collision, the slight forward time shift added to the utterances of the subsequent speakers can ensure correct ordering of the utterances and match the expectations of the subsequent speakers.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Shin, D, "Avoiding Speech Collisions in Multi-party Conferencing via Modulation and Time Shifting", Technical Disclosure Commons, (March 29, 2023)
https://www.tdcommons.org/dpubs_series/5767