Add UMA reporting for audio hardware parameters.

Add UMA reporting for audio hardware parameters.

Records bits per channel, channel layout, and sample rate.

BUG=147572
TEST=build/run

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=156296

[scherkus (not reviewing), 2012-09-11 16:53:31]

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/audio_o...
File media/audio/audio_output_resampler.cc (right):

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/audio_o...
media/audio/audio_output_resampler.cc:87: 40);
not worth enumifying a la sample rate?

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/audio_o...
media/audio/audio_output_resampler.cc:89: "Media.HardwareAudioChannelLayout",
output_params.channel_layout(),
I thought we only got hardware channel *count* and not layout?


If we do indeed track layout, to preserve accuracy of UMA data over time you'll
have to assign values to ChannelLayout + include documentation to not re-use
values (see PipelineStatus for an example)

[DaleCurtis, 2012-09-11 17:35:59]

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/audio_o...
File media/audio/audio_output_resampler.cc (right):

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/audio_o...
media/audio/audio_output_resampler.cc:87: 40);
On 2012/09/11 16:53:31, scherkus wrote:
> not worth enumifying a la sample rate?

Copy/pasted from AudioDecoderConfig. You tell me.

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/audio_o...
media/audio/audio_output_resampler.cc:89: "Media.HardwareAudioChannelLayout",
output_params.channel_layout(),
On 2012/09/11 16:53:31, scherkus wrote:
> I thought we only got hardware channel *count* and not layout?
> 
> 
> If we do indeed track layout, to preserve accuracy of UMA data over time
you'll
> have to assign values to ChannelLayout + include documentation to not re-use
> values (see PipelineStatus for an example)

This is all a copy paste of what's in AudioDecoderConfig? Why would I need to
add more for ChannelLayout here but not there?

We really don't have either. This code will track the input parameters
currently. If we want to track the actual hardware value I'll need to #if
WINDOWS this as well as muck with audio manager to expose that information in a
platform agnostic manner.

(We don't want to put this in GetPreferredLowLatencyBlahBlahBlah right now since
it's being handled at the output device instead of here).

[scherkus (not reviewing), 2012-09-11 17:38:33]

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/sample_...
File media/audio/sample_rates.h (right):

https://chromiumcodereview.appspot.com/10914203/diff/2001/media/audio/sample_...
media/audio/sample_rates.h:23: k192000Hz = 8,
FYI my mac lists 88.2 as well

[DaleCurtis, 2012-09-12 09:50:57]

http://codereview.chromium.org/10914203/diff/2001/media/audio/audio_output_re...
File media/audio/audio_output_resampler.cc (right):

http://codereview.chromium.org/10914203/diff/2001/media/audio/audio_output_re...
media/audio/audio_output_resampler.cc:89: "Media.HardwareAudioChannelLayout",
output_params.channel_layout(),
On 2012/09/11 16:53:31, scherkus wrote:
> I thought we only got hardware channel *count* and not layout?
> 
> 
> If we do indeed track layout, to preserve accuracy of UMA data over time
you'll
> have to assign values to ChannelLayout + include documentation to not re-use
> values (see PipelineStatus for an example)

Done. I'm now pulling the actual hardware channel count for Windows.

http://codereview.chromium.org/10914203/diff/2001/media/audio/sample_rates.h
File media/audio/sample_rates.h (right):

http://codereview.chromium.org/10914203/diff/2001/media/audio/sample_rates.h#...
media/audio/sample_rates.h:23: k192000Hz = 8,
On 2012/09/11 17:38:33, scherkus wrote:
> FYI my mac lists 88.2 as well

Done. Plus a few more from SoundFlower on my mac.

[scherkus (not reviewing), 2012-09-12 13:34:50]

one q -- fyi your overlapping CLs are starting to confuse me :s

http://codereview.chromium.org/10914203/diff/5002/media/audio/audio_output_re...
File media/audio/audio_output_resampler.cc (right):

http://codereview.chromium.org/10914203/diff/5002/media/audio/audio_output_re...
media/audio/audio_output_resampler.cc:30: "Media.HardwareAudioBitsPerChannel",
output_params.bits_per_sample(),
don't you have a WASAPI version of this?

http://codereview.chromium.org/10914203/diff/5002/media/base/channel_layout.h
File media/base/channel_layout.h (right):

http://codereview.chromium.org/10914203/diff/5002/media/base/channel_layout.h...
media/base/channel_layout.h:11: // Do not change the order of these values.
can you elaborate / mention UMA?

Instead of copying the comment from sample_rates.h I'd use the one from
pipeline_status.h as it's much more informative (it's not so much the order but
rather reuse / always adding new values)

Bonus points for updating the one in sample_rates.h at the same time :)

[scherkus (not reviewing), 2012-09-12 13:35:13]

http://codereview.chromium.org/10914203/diff/5002/media/audio/audio_output_re...
File media/audio/audio_output_resampler.cc (right):

http://codereview.chromium.org/10914203/diff/5002/media/audio/audio_output_re...
media/audio/audio_output_resampler.cc:33: // TODO(dalecurtis): Since channel
mixing is handle by the output device right
s/handle/handled/

[DaleCurtis, 2012-09-12 14:03:52]

Done! If you reviewed faster I wouldn't have to keep iterating on top of
existing code reviews :)

http://codereview.chromium.org/10914203/diff/5002/media/audio/audio_output_re...
File media/audio/audio_output_resampler.cc (right):

http://codereview.chromium.org/10914203/diff/5002/media/audio/audio_output_re...
media/audio/audio_output_resampler.cc:30: "Media.HardwareAudioBitsPerChannel",
output_params.bits_per_sample(),
On 2012/09/12 13:34:50, scherkus wrote:
> don't you have a WASAPI version of this?

Not sure I'm landing that. I don't think it will work w/o supporting 24-bit
interleave on ToInterleaved() first.

http://codereview.chromium.org/10914203/diff/5002/media/audio/audio_output_re...
media/audio/audio_output_resampler.cc:33: // TODO(dalecurtis): Since channel
mixing is handle by the output device right
On 2012/09/12 13:35:13, scherkus wrote:
> s/handle/handled/

Done.

http://codereview.chromium.org/10914203/diff/5002/media/base/channel_layout.h
File media/base/channel_layout.h (right):

http://codereview.chromium.org/10914203/diff/5002/media/base/channel_layout.h...
media/base/channel_layout.h:11: // Do not change the order of these values.
On 2012/09/12 13:34:50, scherkus wrote:
> can you elaborate / mention UMA?
> 
> Instead of copying the comment from sample_rates.h I'd use the one from
> pipeline_status.h as it's much more informative (it's not so much the order
but
> rather reuse / always adding new values)
> 
> Bonus points for updating the one in sample_rates.h at the same time :)

Done.

[scherkus (not reviewing), 2012-09-12 14:09:22]

lgtm