All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=722920

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgmuLS5AkM


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

chromium-rel-mac11

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8979417692104012736

=== Auto-CCing suspected CL author uzair.jaleel@samsung.com ===

Hi uzair.jaleel@samsung.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 : uzair.jaleel
  Commit : f5978920bd00bc61544631be138f042d29aecbb4
  Date   : Tue May 09 06:40:08 2017
  Subject: Below change safely converts video color space enums directly to gfx color

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : media_perftests
  Metric       : audio_bus_to_interleaved/int8_t
  Change       : 38.78% | 20.8885333333 -> 28.9883666667

Revision             Result                   N
chromium@470228      20.8885 +- 0.124618      6      good
chromium@470244      21.0689 +- 0.297273      6      good
chromium@470248      21.0552 +- 0.382915      6      good
chromium@470249      21.0504 +- 0.25105       6      good
chromium@470250      29.0141 +- 0.310409      6      bad       <--
chromium@470252      28.9559 +- 0.306784      6      bad
chromium@470260      29.07 +- 0.297728        6      bad
chromium@470292      28.9884 +- 0.213026      6      bad

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

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8979417692104012736

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4951430319833088


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

That CL cannot possibly affect audio performance.

I think this graph is just semi-randomly bimodal.
(Like, maybe the buffer that is used in the test sometimes gets allocated over a page boundary, and sometimes not?  Or maybe it's sometimes better aligned?)
Either way, I don't think there is a regression here.