All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=775291

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=dac84b332a6bca0c2c1e895174a364b106b0bfbc185509a01235f65748667106


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

chromium-rel-mac11

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8965532817270894368

=== Auto-CCing suspected CL author julien.isorce@chromium.org ===

Hi julien.isorce@chromium.org, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : Julien Isorce
  Commit : 6c83d8de501b554f57128b285c5f91275b85ab08
  Date   : Thu Oct 12 13:11:29 2017
  Subject: Move VideoRotation from DemuxerStream to VideoDecoderConfig

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media_perftests
  Metric       : audio_bus_from_interleaved/int8_t
  Change       : 9.16% | 10.6741833333 -> 11.65195

Revision             Result                    N
chromium@508285      10.6742 +- 0.0887706      6      good
chromium@508315      10.7122 +- 0.205215       6      good
chromium@508323      10.6927 +- 0.0604792      6      good
chromium@508327      10.6786 +- 0.0947677      6      good
chromium@508328      11.7692 +- 0.479453       6      bad       <--
chromium@508329      11.7516 +- 0.166282       6      bad
chromium@508330      11.7696 +- 0.198882       6      bad
chromium@508344      11.652 +- 0.134814        6      bad

To Run This Test
  ./src/out/Release/media_perftests --single-process-tests

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8965532817270894368


For feedback, file a bug with component Speed>Bisection

Hi, thx for reporting this, I will have a look. Can you confirm it happens on Linux too  ?

I cannot reproduce the issue on Linux at least. Also my CL https://chromium-review.googlesource.com/c/chromium/src/+/684740 does not change AudioBus and the test is only about to AudioBus. Unless I am missing something I believe this is a false alarm or something wrong with the bisection.

Adding Chrome Cunningham for thoughts ?

Could easily be noise. I don't see how your CL could affect this metric. Zooming out on the graph, it looks like the current performance is generally in line with past values. 

I'll restart the bisect to see if the tool *reliably* points to your CL... 

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8965350875427763680

=== Auto-CCing suspected CL author servolk@google.com ===

Hi servolk@google.com, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : Sergey Volk
  Commit : 4c1fafef4501996b4346d75c51d0e4ce4714ea9c
  Date   : Wed Oct 11 01:50:52 2017
  Subject: Fix initial track selection in ffmpeg demuxer

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media_perftests
  Metric       : audio_bus_from_interleaved/int8_t
  Change       : 9.82% | 10.6891833333 -> 11.738975

Revision             Result                    N
chromium@507308      10.6892 +- 0.081273       6      good
chromium@507715      10.734 +- 0.162731        6      good
chromium@507817      10.7484 +- 0.253599       6      good
chromium@507843      10.6728 +- 0.0833583      6      good
chromium@507856      10.8115 +- 0.637625       6      good
chromium@507858      10.6678 +- 0.127134       6      good
chromium@507859      11.8111 +- 0.140914       6      bad       <--
chromium@507862      11.7773 +- 0.47933        6      bad
chromium@507868      11.8713 +- 0.655324       6      bad
chromium@507919      11.7777 +- 0.189628       6      bad
chromium@508122      11.7766 +- 0.402003       6      bad
chromium@508936      11.739 +- 0.236708        6      bad

To Run This Test
  ./src/out/Release/media_perftests --single-process-tests

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8965350875427763680


For feedback, file a bug with component Speed>Bisection

Nope - that culprit also looks wrong (no changes to audio bus code paths). Closing as noise. +Dale for FYI