Contribution
Assessing the Impact of Echo on Conversations in Real-Time Communication Systems
* Presenting author
Abstract:
Speech quality plays a crucial role in real-time communication, directly affecting user engagement and interaction. Hearing one’s own voice during a conversation can disrupt concentration and lead to discomfort, presenting significant challenges to effective communication. Especially when echo levels are high, speakers have difficulty maintaining conversation. This study introduces and discusses the specifications of a real-time communication system designed to investigate the behavioral effects of echo on speakers and its impact on conversation quality through controlled audio delays and different sound levels. We evaluated delays shorter than 10 ms, delays between 10 and 30 ms, and delays longer than 30 ms with different attenuation levels to assess their impact on the speaker’s experience. As described in ITU P.805, we conducted Short Conversion Tests (SCT) and Random Number Verification (RNV) tests at various delay and attenuation levels to analyze user responses to different echo conditions in a laboratory. Speech quality was evaluated using the Mean Opinion Score (MOS) to capture subjective user perceptions during the speaking phase and the overall quality of the conversation. The preliminary results suggest that echo delays influence the user’s speaking experience, often slowing their speech rate and lowering the volume at several delay and attenuation levels.