Hello grunell@, can you provide more detail in the bug about what's going on? Is this a regression? What prompted filing the bug? What's the severity/implications of it not being fixed?

This was prompted by https://buganizer.corp.google.com/issues/13976602, which makes hangouts painful to use due to the frequent echo. 

The UMA stats reported in the link in the report shows that there are glitches and we should investigate this further. Glitches can cause echo due to delay jumps, and may need to be compensated for.

One piece of this investigation is  issue 610967 . We also need to take a deeper look at the stats.

What's the min time for glitches to be noticeable to the user? Looking at Media.Audio.Capture.LargestGlitchMs there's a mode at 7-9ms. And do you have any theories on the large number of glitches at 60s and above? Just wondering if some other event is being captured by mistake (e.g. some error in input stream setup rather than an actual 60s+ gap in getting data).

Any glitch size *can* be noticeable due to the discontinuity that comes with it. Whether it actually is and to what extent depends on the audio data when it happens, and if there's just one or several following each other. 5-10 ms is very likely to be noticeable.

Very large glitches could be going to sleep and wake up.

Moving this nonessential bug to the next milestone.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

what milestone are you aiming for? M54, M55?

M55.

[Re-triage] Still valid.

Windows counterpart in issue 826668.