The AEC3 delay estimator sometime is slow at tracking audio path changes |
||||||
Issue descriptionOn some systems with audio path changes, the delay estimator does not track the changes as fast as would be desired which may cause echo leakage for up to 10 seconds.
,
Feb 13 2018
The following revision refers to this bug: https://webrtc.googlesource.com/src.git/+/fdd4400ef48f17ed2a8bcf841938c30f1d89588e commit fdd4400ef48f17ed2a8bcf841938c30f1d89588e Author: Per Åhgren <peah@webrtc.org> Date: Tue Feb 13 14:29:23 2018 Removed hysteresis in the delay estimation offset Bug: chromium:811658 , webrtc:8879 Change-Id: I9e67fd9aaae4b85e344b9b40ca6bcf9a8fe1eec1 Reviewed-on: https://webrtc-review.googlesource.com/52480 Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org> Commit-Queue: Per Åhgren <peah@webrtc.org> Cr-Commit-Position: refs/heads/master@{#22002} [modify] https://crrev.com/fdd4400ef48f17ed2a8bcf841938c30f1d89588e/modules/audio_processing/aec3/render_delay_controller.cc
,
Feb 15 2018
,
Feb 15 2018
This bug requires manual review: M65 has already been promoted to the beta branch, so this requires manual review Please contact the milestone owner if you have questions. Owners: cmasso@(Android), cmasso@(iOS), bhthompson@(ChromeOS), govind@(Desktop) For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Feb 15 2018
Before we approve merge to M65, could you pls confirm followings? Is this M65 regression and critical to merge? Is the change well baked/verified in Canary, having enough automation tests coverage and safe to merge? Any other imp details to justify the merge Please note M65 is already promoted to Beta so merge bar is very high. Thank you.
,
Feb 15 2018
The merge is aims to reduce the recently identified effects of audio playout glitches (among other places reported in https://bugs.chromium.org/p/chromium/issues/detail?id=798690) which are causing echo issues in VoIP calls. The change has been baked in Canary for 1.5 days and the metrics looks good. Apart from that it has been extensively tested on the devices in our test lab. Furthermore, the code resides beneath an experimental flag which makes the merge safe.
,
Feb 15 2018
What I meant with the experimental flag is that it can be used to turn off the merged code if anything would go wrong with the merge.
,
Feb 15 2018
Approving merge to M65 branch based on comments #6 and #7. Please merge ASAP. Also pls mark bug 798690 as fixed after merge if nothing is pending. Thank you.
,
Feb 18 2018
The following revision refers to this bug: https://webrtc.googlesource.com/src.git/+/fc67520ded384429039770ec0b449aa3e601a249 commit fc67520ded384429039770ec0b449aa3e601a249 Author: Per Åhgren <peah@webrtc.org> Date: Sun Feb 18 23:00:40 2018 Merge to M65: Leveraging the skew in API call order to a boost AEC3 ... This CL resets the AEC3 realignment functionality when a significant and persistent skew in the number of render and capture API calls is detected. TBR=gustaf@webrtc.org,henrik.lundin@webrtc.org (cherry picked from commit 4712776bf43f9dd3cf93368d0f4bfd892e32aebe) Bug: chromium:811658 , webrtc:8879 Change-Id: Ib5c727b38f427da2a7d25eac7c939a17bdaabe74 Reviewed-on: https://webrtc-review.googlesource.com/52260 Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org> Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org> Cr-Original-Commit-Position: refs/heads/master@{#21997} Reviewed-on: https://webrtc-review.googlesource.com/54312 Reviewed-by: Per Åhgren <peah@webrtc.org> Cr-Commit-Position: refs/branch-heads/65@{#21} Cr-Branched-From: 3ac67a736bb200ecf7c116a88b2f8d5c542973c8-refs/heads/master@{#21637} [modify] https://crrev.com/fc67520ded384429039770ec0b449aa3e601a249/modules/audio_processing/aec3/block_processor.cc [modify] https://crrev.com/fc67520ded384429039770ec0b449aa3e601a249/modules/audio_processing/aec3/echo_path_delay_estimator.cc [modify] https://crrev.com/fc67520ded384429039770ec0b449aa3e601a249/modules/audio_processing/aec3/echo_path_delay_estimator.h [modify] https://crrev.com/fc67520ded384429039770ec0b449aa3e601a249/modules/audio_processing/aec3/mock/mock_render_delay_controller.h [modify] https://crrev.com/fc67520ded384429039770ec0b449aa3e601a249/modules/audio_processing/aec3/render_delay_controller.cc [modify] https://crrev.com/fc67520ded384429039770ec0b449aa3e601a249/modules/audio_processing/aec3/render_delay_controller.h
,
Feb 18 2018
The following revision refers to this bug: https://webrtc.googlesource.com/src.git/+/53941ef41f884ea13c0593e1c3645af3ab50c2e3 commit 53941ef41f884ea13c0593e1c3645af3ab50c2e3 Author: Per Åhgren <peah@webrtc.org> Date: Sun Feb 18 23:45:59 2018 Merge to M65: Removed hysteresis in the delay estimation offset (cherry picked from commit fdd4400ef48f17ed2a8bcf841938c30f1d89588e) TBR: gustaf@webrtc.org,henrik.lundin@webrtc.org Bug: chromium:811658 , webrtc:8879 Change-Id: I9e67fd9aaae4b85e344b9b40ca6bcf9a8fe1eec1 Reviewed-on: https://webrtc-review.googlesource.com/52480 Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org> Commit-Queue: Per Åhgren <peah@webrtc.org> Cr-Original-Commit-Position: refs/heads/master@{#22002} Reviewed-on: https://webrtc-review.googlesource.com/54313 Reviewed-by: Per Åhgren <peah@webrtc.org> Cr-Commit-Position: refs/branch-heads/65@{#22} Cr-Branched-From: 3ac67a736bb200ecf7c116a88b2f8d5c542973c8-refs/heads/master@{#21637} [modify] https://crrev.com/53941ef41f884ea13c0593e1c3645af3ab50c2e3/modules/audio_processing/aec3/render_delay_controller.cc
,
Feb 18 2018
Is anything pending for M65?
,
Feb 19 2018
Not for this issue. There is, however, a related issue for which a merge is requested (https://bugs.chromium.org/p/chromium/issues/detail?id=812524) which addresses a delay buildup which has been shown to happen due to the audio path glitches. But the fixing CL for that addresses the size of the look window rather than the speed of tracking for the delay estimator so therefore it merits a different issue.
,
Feb 19 2018
|
||||||
►
Sign in to add a comment |
||||||
Comment 1 by bugdroid1@chromium.org
, Feb 13 2018