Modify OLA to use window size in seconds instead of bytes.

Modify OLA to use window size in seconds instead of bytes.
BUG=16011
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=20416

[kylep, 2009-07-10 01:17:33]

Trybots are failing because they haven't picked up my previous patch yet. Please
review regardless; I'll re-try in the morning. Thanks, -Kyle

[scherkus (not reviewing), 2009-07-10 01:23:50]

Can you add BUG=16011 to the changelist description?

[scherkus (not reviewing), 2009-07-10 01:30:19]

few tiny nits, and don't forget to attach the BUG= number

http://codereview.chromium.org/157001/diff/1/5
File media/filters/audio_renderer_algorithm_ola.cc (right):

http://codereview.chromium.org/157001/diff/1/5#newcode16
Line 16: const double kDefaultXfadeLength = 0.008;
nit: could you use "Crossfade" instead of "Xfade" (for the non-audio-savvy)

http://codereview.chromium.org/157001/diff/1/5#newcode101
Line 101: window_size_ = static_cast<size_t>(sample_rate()
add a quick comment about what's going on here

what about precision or overflow?  add comments saying if you don't care about
precision or why we wouldn't overflow etc...

http://codereview.chromium.org/157001/diff/1/2
File media/tools/wav_ola_test.cc (right):

http://codereview.chromium.org/157001/diff/1/2#newcode110
Line 110: * wav.bit_rate
nit: may want to do this:
* (wav.bit_rate / 8)

so it's more clear that it's a conversion from bits to bytes

[kylep, 2009-07-10 18:06:07]

Frank - I think this patch makes it sound better even on the inputs I thought
sounded ok before. If you have some free time can you run a few wavs thru this?
I think your ears are better trained than mine.

http://codereview.chromium.org/157001/diff/1/5
File media/filters/audio_renderer_algorithm_ola.cc (right):

http://codereview.chromium.org/157001/diff/1/5#newcode16
Line 16: const double kDefaultXfadeLength = 0.008;
On 2009/07/10 01:30:19, scherkus wrote:
> nit: could you use "Crossfade" instead of "Xfade" (for the non-audio-savvy)

Done.

http://codereview.chromium.org/157001/diff/1/5#newcode101
Line 101: window_size_ = static_cast<size_t>(sample_rate()
On 2009/07/10 01:30:19, scherkus wrote:
> add a quick comment about what's going on here
> 
> what about precision or overflow?  add comments saying if you don't care about
> precision or why we wouldn't overflow etc...

Done.

http://codereview.chromium.org/157001/diff/1/2
File media/tools/wav_ola_test.cc (right):

http://codereview.chromium.org/157001/diff/1/2#newcode110
Line 110: * wav.bit_rate
On 2009/07/10 01:30:19, scherkus wrote:
> nit: may want to do this:
> * (wav.bit_rate / 8)
> 
> so it's more clear that it's a conversion from bits to bytes

Done.

[scherkus (not reviewing), 2009-07-10 18:41:45]

awesome!

lgtm!

http://codereview.chromium.org/157001/diff/3003/3006
File media/filters/audio_renderer_algorithm_base.cc (right):

http://codereview.chromium.org/157001/diff/3003/3006#newcode37
Line 37: DCHECK(callback);
nit: move this DCHECK to end of your DCHECKs so that all the sample_bits ones
are together