Retyped the frequency estimate of the comfort noise for the higher band to harmonize the AEC code(#4

Retyped the frequency estimate of the comfort noise for the higher band to harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort noise is computed in the different optimized ComfortNoise computation methods.

The changes have been tested for bitexactness.

BUG=webrtc:5201
Committed: https://crrev.com/99b1a32146fcc63e52fd45484958d7c8c4cf0061
Cr-Commit-Position: refs/heads/master@{#11050}

[peah-webrtc, 2015-12-03 11:54:03]

Description was changed from

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

BUG=webrtc:5201
==========

to

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

BUG=webrtc:5201
==========

[peah-webrtc, 2015-12-03 11:54:04]

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

[peah-webrtc, 2015-12-03 11:54:28]

Description was changed from

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

BUG=webrtc:5201
==========

to

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

The changes have been tested for bitexactness.

BUG=webrtc:5201
==========

[minyue-webrtc, 2015-12-04 09:49:44]

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/aec/aec_core.c (right):

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core.c:527: else {
move 527 to 526

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core.c:529: 2 * PART_LEN1 *
sizeof(comfortNoiseHband[0][0]));
I prefer use sizeof(comfortNoiseHband), since it is well defined.

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/aec/aec_core_mips.c (right):

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core_mips.c:277: if (aec->num_bands > 1
&& flagHbandCn == 1) {
Interesting, MIPS was different here, was there a reason?

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core_mips.c:321: else {
321 to 320

[hlundin-webrtc, 2015-12-04 11:01:39]

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/aec/aec_core_mips.c (right):

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core_mips.c:323: 2 * PART_LEN1 *
sizeof(comfortNoiseHband[0][0]));
... and sizeof(comfortNoiseHband)

[peah-webrtc, 2015-12-04 23:02:48]

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/aec/aec_core.c (right):

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core.c:527: else {
On 2015/12/04 09:49:43, minyue-webrtc wrote:
> move 527 to 526

Done.

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core.c:529: 2 * PART_LEN1 *
sizeof(comfortNoiseHband[0][0]));
On 2015/12/04 09:49:44, minyue-webrtc wrote:
> I prefer use sizeof(comfortNoiseHband), since it is well defined.

Me too, but it does not work for multidimensional arrays when they are passed as
method arguments. The compiler complains about that construct.

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/aec/aec_core_mips.c (right):

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core_mips.c:277: if (aec->num_bands > 1
&& flagHbandCn == 1) {
On 2015/12/04 09:49:44, minyue-webrtc wrote:
> Interesting, MIPS was different here, was there a reason?

I think it is just an error. As far as I know, num_bands are only greater than
one when the sample rate is larger than 16000.  So as long as the native sample
rates are used, the statements are equivalent.

Anyway, there cannot be any reason for these functions to behave differently.

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core_mips.c:321: else {
On 2015/12/04 09:49:44, minyue-webrtc wrote:
> 321 to 320

Done.

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core_mips.c:323: 2 * PART_LEN1 *
sizeof(comfortNoiseHband[0][0]));
On 2015/12/04 11:01:39, hlundin-webrtc wrote:
> ... and sizeof(comfortNoiseHband)

As stated in the other place for this, that does not work, the compiler
complains. N

[hlundin-webrtc, 2015-12-08 11:52:52]

lgtm

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/aec/aec_core_mips.c (right):

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core_mips.c:323: 2 * PART_LEN1 *
sizeof(comfortNoiseHband[0][0]));
On 2015/12/04 23:02:48, peah-webrtc wrote:
> On 2015/12/04 11:01:39, hlundin-webrtc wrote:
> > ... and sizeof(comfortNoiseHband)
> 
> As stated in the other place for this, that does not work, the compiler
> complains. N

Acknowledged.

[minyue-webrtc, 2015-12-08 12:33:02]

lgtm

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/aec/aec_core.c (right):

https://codereview.webrtc.org/1494133002/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/aec/aec_core.c:529: 2 * PART_LEN1 *
sizeof(comfortNoiseHband[0][0]));
On 2015/12/04 23:02:48, peah-webrtc wrote:
> On 2015/12/04 09:49:44, minyue-webrtc wrote:
> > I prefer use sizeof(comfortNoiseHband), since it is well defined.
> 
> Me too, but it does not work for multidimensional arrays when they are passed
as
> method arguments. The compiler complains about that construct.

Acknowledged.

[peah-webrtc, 2015-12-16 13:10:20]

The CQ bit was checked by peah@webrtc.org

[commit-bot: I haz the power, 2015-12-16 13:10:28]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1494133002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1494133002/20001

[commit-bot: I haz the power, 2015-12-16 14:07:27]

Description was changed from

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

The changes have been tested for bitexactness.

BUG=webrtc:5201
==========

to

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

The changes have been tested for bitexactness.

BUG=webrtc:5201
==========

[commit-bot: I haz the power, 2015-12-16 14:07:27]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-12-16 14:07:38]

Description was changed from

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

The changes have been tested for bitexactness.

BUG=webrtc:5201
==========

to

==========
Retyped the frequency estimate of the comfort noise for the higher band to
harmonize the AEC code.
-Changed the type for the frequency estimate of the comfort noise for the
 higher band to be a two dimensional float array instead of a complex_t array.
 This makes sense since all the other frequency estimate (apart from the
 coherence) use this format and doing this change allows bundling the
 IFFT operations into using the InverseFFT method.
-Moved the memset of the frequency estimate of the comfort noise to where it is
used and made it conditional so that it is only performed when used.
-Harmonized the if-statements for when the frequency estimate of the comfort
noise is computed in the different optimized ComfortNoise computation methods.

The changes have been tested for bitexactness.

BUG=webrtc:5201

Committed: https://crrev.com/99b1a32146fcc63e52fd45484958d7c8c4cf0061
Cr-Commit-Position: refs/heads/master@{#11050}
==========

[commit-bot: I haz the power, 2015-12-16 14:07:38]

Patchset 2 (id:??) landed as
https://crrev.com/99b1a32146fcc63e52fd45484958d7c8c4cf0061
Cr-Commit-Position: refs/heads/master@{#11050}