9 Commits

Author SHA1 Message Date
kwiberg@webrtc.org
00b8f6b364 Use base/scoped_ptr.h; system_wrappers/interface/scoped_ptr.h is going away
BUG=
R=andrew@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/36229004

Cr-Commit-Position: refs/heads/master@{#8517}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8517 4adac7df-926f-26a2-2b94-8c16560cd09d
2015-02-26 14:43:50 +00:00
andrew@webrtc.org
8328e7c44d Revert "Revert part of r7561, "Refactor audio conversion functions.""
This restores the conversion changes to AudioProcessing originally
added in r7561, with minor alterations to ensure it passes all tests.

TBR=kwiberg

Review URL: https://webrtc-codereview.appspot.com/28899004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7574 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-10-31 04:58:14 +00:00
andrew@webrtc.org
4fc4addc81 Refactor audio conversion functions.
Use a consistent naming scheme that can be understood at the callsite
without having to refer to documentation.

Remove hacks in AudioBuffer intended to maintain bit-exactness with the
float path. The conversions etc. are now all natural, and instead we
enforce close but not bit-exact output between the two paths.

Output of ApmTest.Process:
https://paste.googleplex.com/5931055831842816

R=aluebs@webrtc.org, bjornv@webrtc.org, kwiberg@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/13049004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7561 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-10-30 03:40:10 +00:00
andrew@webrtc.org
17454f79dc Add ctors to ChannelBuffer to enable copying on construction.
Also:
- Fix the constness of some parameters.
- Add more const overloads.
- Use DCHECK in place of assert.
- Removed an unnecessary memset.

R=claguna@google.com

Review URL: https://webrtc-codereview.appspot.com/24469004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7107 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-09-08 20:27:04 +00:00
kwiberg@webrtc.org
efb81d8d1f int16<->float conversions: Use size_t for array length argument, not int
size_t is more appropriate for array lengths, since int might
theoretically be too small for a really large array. But more
importantly, if the caller's value is naturally of type size_t and the
function requires an int, VC++ will trigger warning C4267
(http://msdn.microsoft.com/en-us/library/6kck0s93.aspx) because the
implicit cast might be lossy, forcing the caller to do a manual cast.
Typing the function with size_t in the first place resolves the
problem.

R=aluebs@webrtc.org, andrew@webrtc.org, minyue@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21909004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6702 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-16 08:36:52 +00:00
andrew@webrtc.org
17e40641b3 Add a deinterleaved float interface to AudioProcessing.
This is mainly to support the native audio format in Chrome. Although
this implementation just moves the float->int conversion under the hood,
we will transition AudioProcessing towards supporting this format
throughout.

- Add a test which verifies we get identical output with the float and
int interfaces.
- The float and int wrappers are tasked with conversion to the
AudioBuffer format. A new shared Process/Analyze method does most of
the work.
- Add a new field to the debug.proto to hold deinterleaved data.
- Add helpers to audio_utils.cc, and start using numeric_limits.
- Note that there was no performance difference between numeric_limits
and a literal value when measured on Linux using gcc or clang.

BUG=2894
R=aluebs@webrtc.org, bjornv@webrtc.org, henrikg@webrtc.org, tommi@webrtc.org, turaj@webrtc.org, xians@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/9179004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@5641 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-03-04 20:58:13 +00:00
turaj@webrtc.org
d4d5be8781 Minor improvement in RoundToInt16 implementation.
R=andrew@webrtc.org, bjornv@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/8959004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@5586 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-02-20 20:55:21 +00:00
andrew@webrtc.org
b159c2e3dd Reduce cost of PushSincResampler::Resample().
Ideally, PushSincResampler would have very little overhead on
SincResampler. This gets closer to that ideal.

Replace std::min/max and floor with inline functions. Add a benchmark
test to verify the improvement.

On a MacBook Retina, this results in PushSincResampler::Resample()
accounting for ~1% of CPU usage on voe_cmd_test vs the earlier ~2%
(with ISAC16 and 48 kHz audio devices).

Using the new benchmark, this results in a performance improvement of:
16 -> 44.1 : 1.7x
16 -> 48   : 1.9x
32 -> 44.1 : 1.6x
32 -> 48   : 1.7x
44.1 -> 16 : 1.5x
44.1 -> 32 : 1.7x
44.1 -> 48 : 1.7x
48 -> 16   : 1.5x
48 -> 32   : 1.5x
48 -> 44.1 : 1.8x

R=turaj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/2157005

git-svn-id: http://webrtc.googlecode.com/svn/trunk@4695 4adac7df-926f-26a2-2b94-8c16560cd09d
2013-09-06 21:15:55 +00:00
andrew@webrtc.org
50b2efef6e Add a wrapper around PushSincResampler and the old Resampler.
The old resampler is used whenever it supports the requested rates. Otherwise
the sinc resampler is enabled.

Integrated with output_mixer in order to test the change through
output_mixer_unittest. The sinc resampler will not yet be used, since we don't
feed VoE with any rates that trigger it.

BUG=webrtc:1395
R=bjornv@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/1355004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@3915 4adac7df-926f-26a2-2b94-8c16560cd09d
2013-04-29 17:27:29 +00:00