13 Commits

Author SHA1 Message Date
aluebs
4896aaa9e4 Limit the maximum and minimum gain the IntelligibilityEnhancer can apply
Although the algorithm should generate the optimal gains, these limits are in place so that it doesn't go bananas in extreme noise situations or if there are some errors on the noise estimate. The limits were chosen as the extreme values from recordings with -3dB SNR, which is the minimum where the POLQA metric increases when processed.

Review-Url: https://codereview.webrtc.org/1925933002
Cr-Commit-Position: refs/heads/master@{#12550}
2016-04-28 23:09:03 +00:00
aluebs
0a00759780 Fix the stereo support in IntelligibilityEnhancer
Review URL: https://codereview.webrtc.org/1729753003

Cr-Commit-Position: refs/heads/master@{#11795}
2016-02-27 01:17:44 +00:00
Alejandro Luebs
18fcbcf48c Use VAD to get a better speech power estimation in the IntelligibilityEnhancer
R=henrik.lundin@webrtc.org, turaj@webrtc.org

Review URL: https://codereview.webrtc.org/1693823004 .

Cr-Commit-Position: refs/heads/master@{#11713}
2016-02-22 23:57:45 +00:00
Alejandro Luebs
32348192cc Fix and simplify the power estimation in the IntelligibilityEnhancer
R=henrik.lundin@webrtc.org, turaj@webrtc.org

Review URL: https://codereview.webrtc.org/1685703004 .

Cr-Commit-Position: refs/heads/master@{#11663}
2016-02-18 04:04:25 +00:00
Peter Kasting
dce40cf804 Update a ton of audio code to use size_t more correctly and in general reduce
use of int16_t/uint16_t.

This is the upshot of a recommendation by henrik.lundin and kwiberg on an original small change ( https://webrtc-codereview.appspot.com/42569004/#ps1 ) to stop using int16_t just because values could fit in it, and is similar in nature to a previous "mass change to use size_t more" ( https://webrtc-codereview.appspot.com/23129004/ ) which also needed to be split up for review but to land all at once, since, like adding "const", such changes tend to cause a lot of transitive effects.

This was be reviewed and approved in pieces:
https://codereview.webrtc.org/1224093003
https://codereview.webrtc.org/1224123002
https://codereview.webrtc.org/1224163002
https://codereview.webrtc.org/1225133003
https://codereview.webrtc.org/1225173002
https://codereview.webrtc.org/1227163003
https://codereview.webrtc.org/1227203003
https://codereview.webrtc.org/1227213002
https://codereview.webrtc.org/1227893002
https://codereview.webrtc.org/1228793004
https://codereview.webrtc.org/1228803003
https://codereview.webrtc.org/1228823002
https://codereview.webrtc.org/1228823003
https://codereview.webrtc.org/1228843002
https://codereview.webrtc.org/1230693002
https://codereview.webrtc.org/1231713002

The change is being landed as TBR to all the folks who reviewed the above.

BUG=chromium:81439
TEST=none
R=andrew@webrtc.org, pbos@webrtc.org
TBR=aluebs, andrew, asapersson, henrika, hlundin, jan.skoglund, kwiberg, minyue, pbos, pthatcher

Review URL: https://codereview.webrtc.org/1230503003 .

Cr-Commit-Position: refs/heads/master@{#9768}
2015-08-24 21:52:45 +00:00
ekmeyerson
60d9b332a5 Integrate Intelligibility with APM
- Integrates intelligibility into audio_processing.
    - Allows modification of reverse stream if intelligibility enabled.
- Makes intelligibility available in audioproc_float test.
    - Adds reverse stream processing to audioproc_float.
- (removed) Makes intelligibility toggleable in real time in voe_cmd_test.
- Cleans up intelligibility construction, parameters, constants and dead code.

TBR=pbos@webrtc.org

Review URL: https://codereview.webrtc.org/1234463003

Cr-Commit-Position: refs/heads/master@{#9713}
2015-08-14 17:35:58 +00:00
ekmeyerson
3ab2f14d56 Remove C++11 calls from intelligibility_utils
The C++11 here was overkill. This replaces it with simpler logic that
covers all cases encountered so far in practice.

The problem was previously brought up here: https://codereview.webrtc.org/1250663007/

BUG=427718, 487341, webrtc:4866
R=andrew@webrtc.org

Review URL: https://codereview.webrtc.org/1242943008

Cr-Commit-Position: refs/heads/master@{#9627}
2015-07-23 19:15:32 +00:00
pkasting
b297c5a01f Miscellaneous changes split from https://codereview.webrtc.org/1230503003 .
These are mostly trivial changes and are separated out just to reduce the
diff on that change to the minimum possible.

Note explanatory comments on patch set 1.

BUG=none
TEST=none

Review URL: https://codereview.webrtc.org/1235643003

Cr-Commit-Position: refs/heads/master@{#9617}
2015-07-22 22:17:26 +00:00
ekm
35b72fbceb Add new variance update option and unittests for intelligibility
- New option for computing variance that is more adaptive with lower complexity.
- Fixed related off-by-one errors.
- Added intelligibility unittests.
- Do not enhance if experiencing variance underflow.

R=andrew@webrtc.org, henrik.lundin@webrtc.org

Review URL: https://codereview.webrtc.org/1207353002 .

Cr-Commit-Position: refs/heads/master@{#9567}
2015-07-10 21:11:57 +00:00
ekm
db4fecfb01 Attempt to reland: Allow intelligibility to compile in apm (https://codereview.webrtc.org/1182323005/)
Revert of original: https://codereview.webrtc.org/1187033005/

Changes in original:
- Added files to gyp and BUILD
- Made minor fixes to get everything to compile
    and intelligibility_proc to run
- Added comments
- Auto-reformatting

New Changes:
- Added <numeric> header to intelligibility_enhancer.cc to address buildbot errors
- Switched to use WAV for i/o in intelligibility_proc.cc to address windows errors
- clean up

Note: Patch 1 duplicates Patch 7 of https://codereview.webrtc.org/1182323005/

R=andrew@webrtc.org

Review URL: https://codereview.webrtc.org/1190733004.

Cr-Commit-Position: refs/heads/master@{#9486}
2015-06-23 00:49:14 +00:00
aluebs
c555b99c13 Revert of Allow intelligibility to compile in apm (patchset #1 id:1 of https://codereview.webrtc.org/1182323005/)
Reason for revert:
Breaking the build bots: http://build.chromium.org/p/client.webrtc/builders/Mac32%20Release%20%5Blarge%20tests%5D/builds/4544

Fails to compile with this error:

../../webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.cc:218:25: error: no member named 'accumulate' in namespace 'std'
    power_target = std::accumulate(clear_variance_.variance(),

Original issue's description:
> Allow intelligibility to compile in apm
>
> - Added files to gyp and BUILD
> - Made minor fixes to get everything to compile
>     and intelligibility_proc to run
> - Added comments
> - Auto-reformatting
>
> Original cl is at: https://webrtc-codereview.appspot.com/57579004/
>
> TBR=aluebs@webrtc.org
>
> Committed: b7553dfdbb

TBR=ekmeyerson@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.webrtc.org/1187033005

Cr-Commit-Position: refs/heads/master@{#9455}
2015-06-17 03:26:20 +00:00
ekm
b7553dfdbb Allow intelligibility to compile in apm
- Added files to gyp and BUILD
- Made minor fixes to get everything to compile
    and intelligibility_proc to run
- Added comments
- Auto-reformatting

Original cl is at: https://webrtc-codereview.appspot.com/57579004/

TBR=aluebs@webrtc.org

Review URL: https://codereview.webrtc.org/1182323005.

Cr-Commit-Position: refs/heads/master@{#9454}
2015-06-17 01:57:37 +00:00
ekm
030249dd24 Initial SIE commit: migrating existing code
Moved exact existing intelligibility enhancement implementation into new
repository for reference when making further changes.

Note: this cl does not add these files to any gyp.

Original cl is at https://webrtc-codereview.appspot.com/52719004/ .

TBR=aluebs@webrtc.org

Review URL: https://codereview.webrtc.org/1177953006.

Cr-Commit-Position: refs/heads/master@{#9441}
2015-06-15 20:02:33 +00:00