I am working on a WebRTC project using the jsSIP library. One user connects from their browser, and the other connects by making a real phone call. While reviewing the audio recordings, I noticed a problem. If there is high background noise on the phone user's side, the voice of the other user is suppressed for 5-6 seconds and is not understandable. However, after that period, the browser user's voice can be heard clearly and the background noise is being suppressed. I tested that on both Chrome and Firefox.
I suspect that the problem may be related to constraints such as noise suppression, echo cancellation, and auto gain control. When I turned off echo cancellation, I saw that the problem of voice suppression in the first few seconds disappeared, but in this case, the situation of the user's voice suppressing background noise did not occur. I am looking for a solution that does not involve turning off the echo cancellation constraint since I don't want to lose this feature.
I also tried disabling other constraints I mentioned but that didn't make any difference.
Is there a known way to shorten the 5-6 second delay at the beginning of the call? Where else should I look?
