Use frames instead of time for running AudioParam timelines.

Use frames instead of time for running AudioParam timelines.

This is a large change of how the AudioParam timeline is
implemented.  Previously, the timeline used floats/doubles to process
the timeline.  This poses several issues including round-off errors in
computing the timeline values and updating the timeline.  In addition,
the sampling of the automation curves was incorrect.

First, to remove the roundoff issues, use the currentFrame to step
through the timeline so roundoff is reduced.

Second, update the sampling of the automation curves correctly. These
errors are shown in crbug 534613 and 527194, which are fixed now.

The audioparam tests were also updated to sample the curves correctly
and also use the formulas from the spec to implement them instead of
trying to duplicate the implementation.

Third, a ManualLayoutTest is added for crbug 522229, which is fixed.

BUG=534613, 527194, 522229
TESTS=updated existing tests and added
audioparam-sampling.html,
audioparam-setTargetValue-value.html, and osc-exponentialRamp.html.

Committed: https://crrev.com/748834c3a5b1b0479835c4cb3ce7d9b9c8302934
Cr-Commit-Position: refs/heads/master@{#352131}

[Raymond Toy, 2015-09-29 20:31:16]

rtoy@chromium.org changed reviewers:
+ hongchan@chromium.org

[Raymond Toy, 2015-09-29 20:31:17]

PTAL.

[hongchan, 2015-09-30 22:14:18]

The first pass of review. Mostly nits.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:175: });     
audit.runTasks();
Add a line break here.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js:541: var type
= typeof value;
Let's remove this line and...

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js:542:
this._assert(type === 'number', 'value should be number for');
change this to: typeof value === 'number'

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js
(right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js:1: var
sampleRate = 44100;
Not relevant to this CL, but this test can be separated into each task inside.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/M...
File third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/M...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:12: function
updateOsc (event)
Let's be consistent with the curly brace after the function name. Also there is
a space after the name.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
File third_party/WebKit/Source/modules/webaudio/AudioParam.cpp (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:162: // equal to
AudioHandler::ProcessingSizeInFrames (the render quantum size).
Wouldn't it be better to have comments like: Using frame over second to avoid
rounding-off error.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:169: m_value =
m_timeline.valuesForTimeRange(startFrame, endFrame,
narrowPrecisionToFloat(m_value), values, numberOfValues, sampleRate,
sampleRate);
In my opinion, the method name with the word 'TimeRange' seems to expect the
arguments as second. WDYT?

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
File third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.h (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.h:58: // Given the
time range, calculates parameter values into the values buffer
This can be like: "Given the time range in frame values..."

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.h:63: float
valuesForTimeRange(size_t startTime, size_t endTime, float defaultValue, float*
values, unsigned numberOfValues, double sampleRate, double controlRate);
size_t startFrame, size_t endFrame

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.h:106: float
valuesForTimeRangeImpl(size_t startTime, size_t endTime, float defaultValue,
float* values, unsigned numberOfValues, double sampleRate, double controlRate);
Same as above.

[Raymond Toy, 2015-09-30 22:30:18]

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js:541: var type
= typeof value;
On 2015/09/30 at 22:14:18, hoch wrote:
> Let's remove this line and...

I was following the other parts of the code....

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js
(right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js:1: var
sampleRate = 44100;
On 2015/09/30 at 22:14:18, hoch wrote:
> Not relevant to this CL, but this test can be separated into each task inside.

Sorry, I don't understand this comment.  What test?  What task?

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
File third_party/WebKit/Source/modules/webaudio/AudioParam.cpp (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:162: // equal to
AudioHandler::ProcessingSizeInFrames (the render quantum size).
On 2015/09/30 at 22:14:18, hoch wrote:
> Wouldn't it be better to have comments like: Using frame over second to avoid
rounding-off error.

You want to add a comment that we're using frames instead of seconds?

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:169: m_value =
m_timeline.valuesForTimeRange(startFrame, endFrame,
narrowPrecisionToFloat(m_value), values, numberOfValues, sampleRate,
sampleRate);
On 2015/09/30 at 22:14:18, hoch wrote:
> In my opinion, the method name with the word 'TimeRange' seems to expect the
arguments as second. WDYT?

Renaming to valuesForFrameRange is fine with me.

[hongchan, 2015-09-30 22:47:59]

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js:541: var type
= typeof value;
On 2015/09/30 22:30:18, Raymond Toy wrote:
> On 2015/09/30 at 22:14:18, hoch wrote:
> > Let's remove this line and...
> 
> I was following the other parts of the code....

Ha, you're right. Let's leave it as it is for now. I will fix them all up with
another CL.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audio-testing.js:542:
this._assert(type === 'number', 'value should be number for');
On 2015/09/30 22:14:18, hoch wrote:
> change this to: typeof value === 'number'

Please disregard this comment.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js
(right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js:1: var
sampleRate = 44100;
On 2015/09/30 22:30:18, Raymond Toy wrote:
> On 2015/09/30 at 22:14:18, hoch wrote:
> > Not relevant to this CL, but this test can be separated into each task
inside.
> 
> Sorry, I don't understand this comment.  What test?  What task?

First, this file is a bit bloated now. I thought we can group/categorize things
in this file into two or three files. There is even a duplicated utility down
there. (createConstantBuffer).
Second, the utilities for the test and the actual execution of test can be
separated.

Like I mentioned, these points are not relevant to this CL, so we can deal with
them later.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
File third_party/WebKit/Source/modules/webaudio/AudioParam.cpp (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:162: // equal to
AudioHandler::ProcessingSizeInFrames (the render quantum size).
On 2015/09/30 22:30:18, Raymond Toy wrote:
> On 2015/09/30 at 22:14:18, hoch wrote:
> > Wouldn't it be better to have comments like: Using frame over second to
avoid
> rounding-off error.
> 
> You want to add a comment that we're using frames instead of seconds?

Yes, it might be confusing inside of some other contexts. Just to be clear.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:169: m_value =
m_timeline.valuesForTimeRange(startFrame, endFrame,
narrowPrecisionToFloat(m_value), values, numberOfValues, sampleRate,
sampleRate);
On 2015/09/30 22:30:18, Raymond Toy wrote:
> On 2015/09/30 at 22:14:18, hoch wrote:
> > In my opinion, the method name with the word 'TimeRange' seems to expect the
> arguments as second. WDYT?
> 
> Renaming to valuesForFrameRange is fine with me.

I think there are few other methods named with "TimeRange", so if we decide to
change this then we'll have to deal with the rest as well. Do you think it's
worth it? The word "FrameRange" is very clear to me though.

[Raymond Toy, 2015-10-01 18:05:56]

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:175: });     
audit.runTasks();
On 2015/09/30 at 22:14:18, hoch wrote:
> Add a line break here.

Done.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
File third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js
(right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/L...
third_party/WebKit/LayoutTests/webaudio/resources/audioparam-testing.js:1: var
sampleRate = 44100;
On 2015/09/30 at 22:47:59, hoch wrote:
> On 2015/09/30 22:30:18, Raymond Toy wrote:
> > On 2015/09/30 at 22:14:18, hoch wrote:
> > > Not relevant to this CL, but this test can be separated into each task
inside.
> > 
> > Sorry, I don't understand this comment.  What test?  What task?
> 
> First, this file is a bit bloated now. I thought we can group/categorize
things in this file into two or three files. There is even a duplicated utility
down there. (createConstantBuffer).
> Second, the utilities for the test and the actual execution of test can be
separated.
> 
> Like I mentioned, these points are not relevant to this CL, so we can deal
with them later.

Ok.  But I did remove createConstantBuffer in favor of the version from
audio-testing.js which did
the same thing, with args in a different order.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/M...
File third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/M...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:12: function
updateOsc (event)
On 2015/09/30 at 22:14:18, hoch wrote:
> Let's be consistent with the curly brace after the function name. Also there
is a space after the name.

Done.  Fixed inconsistent indentation too. :-)

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
File third_party/WebKit/Source/modules/webaudio/AudioParam.cpp (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:162: // equal to
AudioHandler::ProcessingSizeInFrames (the render quantum size).
On 2015/09/30 at 22:47:59, hoch wrote:
> On 2015/09/30 22:30:18, Raymond Toy wrote:
> > On 2015/09/30 at 22:14:18, hoch wrote:
> > > Wouldn't it be better to have comments like: Using frame over second to
avoid
> > rounding-off error.
> > 
> > You want to add a comment that we're using frames instead of seconds?
> 
> Yes, it might be confusing inside of some other contexts. Just to be clear.

Since we're renaming valuesForTimeRange to valuesForFrameRange, I think it's
pretty clear
we want frames now instead of seconds.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParam.cpp:169: m_value =
m_timeline.valuesForTimeRange(startFrame, endFrame,
narrowPrecisionToFloat(m_value), values, numberOfValues, sampleRate,
sampleRate);
On 2015/09/30 at 22:47:59, hoch wrote:
> On 2015/09/30 22:30:18, Raymond Toy wrote:
> > On 2015/09/30 at 22:14:18, hoch wrote:
> > > In my opinion, the method name with the word 'TimeRange' seems to expect
the
> > arguments as second. WDYT?
> > 
> > Renaming to valuesForFrameRange is fine with me.
> 
> I think there are few other methods named with "TimeRange", so if we decide to
change this then we'll have to deal with the rest as well. Do you think it's
worth it? The word "FrameRange" is very clear to me though.

I've renamed it everywhere, more or less.

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
File third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.h (right):

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.h:58: // Given the
time range, calculates parameter values into the values buffer
On 2015/09/30 at 22:14:18, hoch wrote:
> This can be like: "Given the time range in frame values..."

done

https://chromiumcodereview.appspot.com/1377903002/diff/1/third_party/WebKit/S...
third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.h:63: float
valuesForTimeRange(size_t startTime, size_t endTime, float defaultValue, float*
values, unsigned numberOfValues, double sampleRate, double controlRate);
On 2015/09/30 at 22:14:18, hoch wrote:
> size_t startFrame, size_t endFrame

Done

[hongchan, 2015-10-01 20:20:53]

Mostly nits.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:13: <div
id="console"></div>
Same comments with other newly added test files: these two DIVs are created by
js-test.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:68:
testPassed(config.desc + " passed.\n");
Do we need a new line character at the end? testPassed() outputs HTML, so there
will be no effect from it.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:70:
testFailed(config.desc + " failed.\n");
Ditto.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File
third_party/WebKit/LayoutTests/webaudio/audioparam-setTargetAtTime-sampling.html
(right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-setTargetAtTime-sampling.html:13:
<div id="console"></div>
Please remove two lines above. These divs are created by js-test dynamically.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:8: <script>
Let us move the script section to the end of body. (after the second button).
The convention is to execute JS code when all the elements in the body are
properly setup.

Even though no DOM element is being accessed by JS code except for |window|, but
let's follow the convention as possible.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:10: var sawosc
= 0;
0 is not necessary.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:13: var freq =
Math.random() * 440 + 440;
Be sure to avoid the case freq = 0. Perhaps we can add some small number to it?

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:14:
console.log(context.currentTime);
Let's remove the console print out.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:37: Press the
start button to start the test and the stop button to stop it
A missing period at the end.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.cpp (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.cpp:89: String
args = String::number(event.value()) + ", " + String::number(event.time(), 16);
Just out of curiosity, why 16?

[Raymond Toy, 2015-10-01 20:31:24]

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:68:
testPassed(config.desc + " passed.\n");
On 2015/10/01 at 20:20:52, hoch wrote:
> Do we need a new line character at the end? testPassed() outputs HTML, so
there will be no effect from it.

Look at some of the expected outputs.  There is a blank line between each
"section" of tests.  Without this \n, there is no blank line.

I find this convenient for separating out sections of tests.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:13: var freq =
Math.random() * 440 + 440;
On 2015/10/01 at 20:20:52, hoch wrote:
> Be sure to avoid the case freq = 0. Perhaps we can add some small number to
it?

How can freq be 0?  Math.random() returns numbers in the range [0,1), so freq >=
440.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.cpp (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/Source/modules/webaudio/AudioParamTimeline.cpp:89: String
args = String::number(event.value()) + ", " + String::number(event.time(), 16);
On 2015/10/01 at 20:20:52, hoch wrote:
> Just out of curiosity, why 16?

That was done in a different CL, unrelated to this.  But 16 is about the number
of significant digits in a double precision number.  Without this, event.time()
is printed with 6 or 7 digits at most.

[Raymond Toy, 2015-10-01 21:30:50]

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:13: <div
id="console"></div>
On 2015/10/01 at 20:20:52, hoch wrote:
> Same comments with other newly added test files: these two DIVs are created by
js-test.

Done

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File
third_party/WebKit/LayoutTests/webaudio/audioparam-setTargetAtTime-sampling.html
(right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-setTargetAtTime-sampling.html:13:
<div id="console"></div>
On 2015/10/01 at 20:20:52, hoch wrote:
> Please remove two lines above. These divs are created by js-test dynamically.

Done

[hongchan, 2015-10-01 22:14:48]

lgtm

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/LayoutTests/webaudio/audioparam-sampling.html:68:
testPassed(config.desc + " passed.\n");
On 2015/10/01 20:31:23, Raymond Toy wrote:
> On 2015/10/01 at 20:20:52, hoch wrote:
> > Do we need a new line character at the end? testPassed() outputs HTML, so
> there will be no effect from it.
> 
> Look at some of the expected outputs.  There is a blank line between each
> "section" of tests.  Without this \n, there is no blank line.
> 
> I find this convenient for separating out sections of tests.

Acknowledged.

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
File third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html (right):

https://chromiumcodereview.appspot.com/1377903002/diff/20001/third_party/WebK...
third_party/WebKit/ManualTests/webaudio/osc-exponentialRamp.html:13: var freq =
Math.random() * 440 + 440;
On 2015/10/01 20:31:23, Raymond Toy wrote:
> On 2015/10/01 at 20:20:52, hoch wrote:
> > Be sure to avoid the case freq = 0. Perhaps we can add some small number to
> it?
> 
> How can freq be 0?  Math.random() returns numbers in the range [0,1), so freq
>=
> 440.

Oops. Sorry. I blindly thought it is -1~1. Please ignore.

[Raymond Toy, 2015-10-02 15:23:55]

The CQ bit was checked by rtoy@chromium.org

[commit-bot: I haz the power, 2015-10-02 15:25:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1377903002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1377903002/40001

[commit-bot: I haz the power, 2015-10-02 16:47:37]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-10-02 16:47:38]

Try jobs failed on following builders:
  linux_android_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_android_r...)

[Raymond Toy, 2015-10-02 17:56:37]

The CQ bit was checked by rtoy@chromium.org

[commit-bot: I haz the power, 2015-10-02 17:58:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1377903002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1377903002/40001

[commit-bot: I haz the power, 2015-10-02 21:12:17]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-10-02 21:13:42]

Patchset 3 (id:??) landed as
https://crrev.com/748834c3a5b1b0479835c4cb3ce7d9b9c8302934
Cr-Commit-Position: refs/heads/master@{#352131}