All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=875388

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=5ac944070106ac92068b48088891bc0ee887e7b03215f05e4d6036d88add52b2


Bot(s) for this bug's original alert(s):

webrtc-linux-large-tests

Looks caused by https://webrtc-review.googlesource.com/93287.

-> sprang to verify this is expected.

Some interesting results here!
Most of stats are improvements. The idea with this change was to not set a too high minimum quality as this would result in increased delays and possibly need to keyframes.

It seems like by doing this we also avoid some overshooting detection from trigering and reducing overall quality - so both min and average psnr/ssim has improved substantially, like 5dB in the simulcast screenshare case!

Few dropped frames, higher overall media bitrate. With that of course comes increased network delay in some situations, but it seems well controlled.

I'll mark this as wai.