Add a reliability term to the echo detector.

Add a reliability term to the echo detector.

This will ensure that the estimated likelihood starts at a low value and prevents initial spikes.

BUG=webrtc:6525
Committed: https://crrev.com/fbb374d8ed961f055d9ec71a25e0bb77991e6ec3
Cr-Commit-Position: refs/heads/master@{#15131}

[ivoc, 2016-11-15 15:18:05]

Description was changed from

==========
Add a reliability term to the echo detector.

This will ensure that the estimated likelihood starts at a low value and
prevents initial spikes.

BUG=webrtc:6525
==========

to

==========
Add a reliability term to the echo detector.

This will ensure that the estimated likelihood starts at a low value and
prevents initial spikes.

BUG=webrtc:6525
==========

[ivoc, 2016-11-15 15:18:05]

ivoc@webrtc.org changed reviewers:
+ peah@webrtc.org

[ivoc, 2016-11-15 15:26:50]

Hi Per,

I noticed some initial spikes in the estimated echo likelihood, so I added a
reliability term to dampen the echo likelihood initially. PTAL.

[ivoc, 2016-11-16 09:26:36]

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

[ivoc, 2016-11-16 09:27:50]

Adding Henrik as reviewer since Per is busy travelling. Henrik: PTAL.

[hlundin-webrtc, 2016-11-16 09:46:31]

https://codereview.webrtc.org/2503843004/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/residual_echo_detector.cc (right):

https://codereview.webrtc.org/2503843004/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/residual_echo_detector.cc:104: reliability_ =
(1.0f - kAlpha) * reliability_ + kAlpha * 1.0f;
How long until this converges to close enough to 1.0? I'm wondering if it might
make sense to enclose those two lines in 
if (reliability_ < 1.0 - some_epsilon) {
  ...
}

[ivoc, 2016-11-16 12:30:38]

https://codereview.webrtc.org/2503843004/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/residual_echo_detector.cc (right):

https://codereview.webrtc.org/2503843004/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/residual_echo_detector.cc:104: reliability_ =
(1.0f - kAlpha) * reliability_ + kAlpha * 1.0f;
On 2016/11/16 09:46:30, hlundin-webrtc wrote:
> How long until this converges to close enough to 1.0? I'm wondering if it
might
> make sense to enclose those two lines in 
> if (reliability_ < 1.0 - some_epsilon) {
>   ...
> }

After about a minute the reliability is at 0.9975 (16 seconds to 0.8). I'm not
sure if a branch would be more or less efficient than 5 floating point
operations though.

[hlundin-webrtc, 2016-11-17 08:14:45]

lgtm

https://codereview.webrtc.org/2503843004/diff/1/webrtc/modules/audio_processi...
File webrtc/modules/audio_processing/residual_echo_detector.cc (right):

https://codereview.webrtc.org/2503843004/diff/1/webrtc/modules/audio_processi...
webrtc/modules/audio_processing/residual_echo_detector.cc:104: reliability_ =
(1.0f - kAlpha) * reliability_ + kAlpha * 1.0f;
On 2016/11/16 12:30:38, ivoc wrote:
> On 2016/11/16 09:46:30, hlundin-webrtc wrote:
> > How long until this converges to close enough to 1.0? I'm wondering if it
> might
> > make sense to enclose those two lines in 
> > if (reliability_ < 1.0 - some_epsilon) {
> >   ...
> > }
> 
> After about a minute the reliability is at 0.9975 (16 seconds to 0.8). I'm not
> sure if a branch would be more or less efficient than 5 floating point
> operations though.

Neither am I...

[ivoc, 2016-11-17 08:43:28]

The CQ bit was checked by ivoc@webrtc.org

[commit-bot: I haz the power, 2016-11-17 08:43:40]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[commit-bot: I haz the power, 2016-11-17 08:54:29]

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

[commit-bot: I haz the power, 2016-11-17 08:54:31]

Try jobs failed on following builders:
  mac_rel on master.tryserver.webrtc (JOB_FAILED,
http://build.chromium.org/p/tryserver.webrtc/builders/mac_rel/builds/20222)

[ivoc, 2016-11-17 13:43:13]

To prevent the slow-start from causing issues in the unittests, I added a
function to set it directly, which should only be used from tests. Does that
sound reasonable?

[hlundin-webrtc, 2016-11-17 13:59:48]

lgtm

[ivoc, 2016-11-17 14:03:03]

The CQ bit was checked by ivoc@webrtc.org

[commit-bot: I haz the power, 2016-11-17 14:03:12]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[commit-bot: I haz the power, 2016-11-17 14:19:49]

Description was changed from

==========
Add a reliability term to the echo detector.

This will ensure that the estimated likelihood starts at a low value and
prevents initial spikes.

BUG=webrtc:6525
==========

to

==========
Add a reliability term to the echo detector.

This will ensure that the estimated likelihood starts at a low value and
prevents initial spikes.

BUG=webrtc:6525
==========

[commit-bot: I haz the power, 2016-11-17 14:19:50]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-11-17 14:20:04]

Description was changed from

==========
Add a reliability term to the echo detector.

This will ensure that the estimated likelihood starts at a low value and
prevents initial spikes.

BUG=webrtc:6525
==========

to

==========
Add a reliability term to the echo detector.

This will ensure that the estimated likelihood starts at a low value and
prevents initial spikes.

BUG=webrtc:6525

Committed: https://crrev.com/fbb374d8ed961f055d9ec71a25e0bb77991e6ec3
Cr-Commit-Position: refs/heads/master@{#15131}
==========

[commit-bot: I haz the power, 2016-11-17 14:20:06]

Patchset 2 (id:??) landed as
https://crrev.com/fbb374d8ed961f055d9ec71a25e0bb77991e6ec3
Cr-Commit-Position: refs/heads/master@{#15131}