Refactored the comfort noise generation code in the AEC.

Refactored the comfort noise generation code in the AEC.
This CL will be followed with other CLs that break apart
the application of the comfort noise from the comfort
noise generation.

The changes in the CL are very close to bitexaxt. The
bitinexactness is caused by differences in numerical
behavior when bundling the spectral band power and the
noise scaling based on the NLP gain.

BUG=webrtc:5201, webrtc:5298
Committed: https://crrev.com/5df729489f4c27e6325f08671dc5bf10d8c14a31
Cr-Commit-Position: refs/heads/master@{#12713}

[peah-webrtc, 2016-05-09 06:33:48]

Description was changed from

==========
Refactored the comfort noise generation code in the AEC.
This CL will be followed with other CLs that break apart the application of the
comfort noise from the comfort noise generation.

The changes in the CL are very close to bitexaxt. The bitinexactness is caused
by differences in numerical behavior when bundling the spectral band power and
the noise scaling based on the NLP gain.

BUG=webrtc:5201, webrtc:5298
==========

to

==========
Refactored the comfort noise generation code in the AEC.
This CL will be followed with other CLs that break apart
the application of the comfort noise from the comfort 
noise generation.

The changes in the CL are very close to bitexaxt. The 
bitinexactness is caused by differences in numerical 
behavior when bundling the spectral band power and the 
noise scaling based on the NLP gain.

BUG=webrtc:5201, webrtc:5298
==========

[peah-webrtc, 2016-05-09 06:33:48]

peah@webrtc.org changed reviewers:
+ henrik.lundin@webrtc.org

[hlundin-webrtc, 2016-05-12 11:01:11]

LGTM with nits.

https://codereview.webrtc.org/1958933002/diff/20001/webrtc/modules/audio_proc...
File webrtc/modules/audio_processing/aec/aec_core.cc (right):

https://codereview.webrtc.org/1958933002/diff/20001/webrtc/modules/audio_proc...
webrtc/modules/audio_processing/aec/aec_core.cc:527: for (int i = 1; i <
PART_LEN1; i++) {
I think the correct thing is to have size_t as index in for loops. But don't
bother if the definition of PART_LEN1 causes mismatch problems.

https://codereview.webrtc.org/1958933002/diff/20001/webrtc/modules/audio_proc...
webrtc/modules/audio_processing/aec/aec_core.cc:558: int start_avg_band =
PART_LEN1 >> 1;
Change to /2 while you are at it.

[peah-webrtc, 2016-05-13 06:13:27]

https://codereview.webrtc.org/1958933002/diff/20001/webrtc/modules/audio_proc...
File webrtc/modules/audio_processing/aec/aec_core.cc (right):

https://codereview.webrtc.org/1958933002/diff/20001/webrtc/modules/audio_proc...
webrtc/modules/audio_processing/aec/aec_core.cc:527: for (int i = 1; i <
PART_LEN1; i++) {
On 2016/05/12 11:01:11, hlundin-webrtc wrote:
> I think the correct thing is to have size_t as index in for loops. But don't
> bother if the definition of PART_LEN1 causes mismatch problems.

Done.

https://codereview.webrtc.org/1958933002/diff/20001/webrtc/modules/audio_proc...
webrtc/modules/audio_processing/aec/aec_core.cc:558: int start_avg_band =
PART_LEN1 >> 1;
On 2016/05/12 11:01:11, hlundin-webrtc wrote:
> Change to /2 while you are at it.

Done.

[peah-webrtc, 2016-05-13 06:14:21]

The CQ bit was checked by peah@webrtc.org

[peah-webrtc, 2016-05-13 06:14:21]

The patchset sent to the CQ was uploaded after l-g-t-m from
henrik.lundin@webrtc.org
Link to the patchset: https://codereview.webrtc.org/1958933002/#ps60001 (title:
"Rebase")

[commit-bot: I haz the power, 2016-05-13 06:14:30]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1958933002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1958933002/60001

[commit-bot: I haz the power, 2016-05-13 07:13:52]

Description was changed from

==========
Refactored the comfort noise generation code in the AEC.
This CL will be followed with other CLs that break apart
the application of the comfort noise from the comfort 
noise generation.

The changes in the CL are very close to bitexaxt. The 
bitinexactness is caused by differences in numerical 
behavior when bundling the spectral band power and the 
noise scaling based on the NLP gain.

BUG=webrtc:5201, webrtc:5298
==========

to

==========
Refactored the comfort noise generation code in the AEC.
This CL will be followed with other CLs that break apart
the application of the comfort noise from the comfort 
noise generation.

The changes in the CL are very close to bitexaxt. The 
bitinexactness is caused by differences in numerical 
behavior when bundling the spectral band power and the 
noise scaling based on the NLP gain.

BUG=webrtc:5201, webrtc:5298
==========

[commit-bot: I haz the power, 2016-05-13 07:13:53]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-05-13 07:14:04]

Description was changed from

==========
Refactored the comfort noise generation code in the AEC.
This CL will be followed with other CLs that break apart
the application of the comfort noise from the comfort 
noise generation.

The changes in the CL are very close to bitexaxt. The 
bitinexactness is caused by differences in numerical 
behavior when bundling the spectral band power and the 
noise scaling based on the NLP gain.

BUG=webrtc:5201, webrtc:5298
==========

to

==========
Refactored the comfort noise generation code in the AEC.
This CL will be followed with other CLs that break apart
the application of the comfort noise from the comfort
noise generation.

The changes in the CL are very close to bitexaxt. The
bitinexactness is caused by differences in numerical
behavior when bundling the spectral band power and the
noise scaling based on the NLP gain.

BUG=webrtc:5201, webrtc:5298

Committed: https://crrev.com/5df729489f4c27e6325f08671dc5bf10d8c14a31
Cr-Commit-Position: refs/heads/master@{#12713}
==========

[commit-bot: I haz the power, 2016-05-13 07:14:05]

Patchset 4 (id:??) landed as
https://crrev.com/5df729489f4c27e6325f08671dc5bf10d8c14a31
Cr-Commit-Position: refs/heads/master@{#12713}