History log of /frameworks/av/include/media/stagefright/OMXCodec.h
Revision Date Author Comments (<<< Hide modified files) (Show modified files >>>)
60b1c0e79d12a1c70758bc8d060156924635f8ba 07-Aug-2014 Lajos Molnar <lajos@google.com> stagefright: rework media codec list and infos

This is in preparation of serving the codec list and codec infos
from the mediaserver

Bug: 11990470
Change-Id: Ib8e2708679c9ce461a4ba179974a740cdcdf2731
3b5a6b9fa6c6825a1d0b441429e2bb365b259827 20-May-2014 Harish Mahendrakar <harish.mahendrakar@ittiam.com> stagefright: added support, and SoftOMX decoder, for HEVC

also fixed QuerySupportedProfileLevels in

Bug: 14571712
Change-Id: Ifa3793f7c1b18ac5e8c0a096848c998cabd9e777
d3c079ae9859011d118f94616d0069c2987013ed 28-Oct-2013 Changwan Ryu <changwan@google.com> Support TS + AC3 for ATSC standard

Change-Id: I141667f3f54b242bafdf0ab9db86852c56f49ffa
2ee14000a38683220fb250d9e7300e1d71ccdaa0 07-Aug-2013 Lajos Molnar <lajos@google.com> Check adaptive playback support via CodecCapabilities

Added flags field to CodecCapabilities, so that applications
can query whether codecs support various features. For now
added one video-decoder feature:


Media playback applications can query it to see if a codec
supports seamless resolution changes, such as by supporting
dynamic output buffers.

Signed-off-by: Lajos Molnar <lajos@google.com>
Change-Id: I09da46e8ab6b8645fa2749e33128e49eda2e865b
Related-to-bug: 7093648
8fa80e3be63ffde5ef029afa713877440856b663 08-Sep-2012 James Dong <jdong@google.com> Allow OMXCodec to specify an output color format for OMX decoder component

The output color format is specified via the meta argument in OMXCodec::Create()

o related-to-bug: 7122195

Change-Id: Id3247686b893af25cc190685201e53ad34b0399c
aa8879d381918335c9d829fc279b76d53a15a550 24-Aug-2012 James Dong <jdong@google.com> Fixed a native crash due to unexpected state at destroy time when encoder component's source fails to start

o Encoder component initializes to be in the state of EXECUTING before its source gets started, because we wanted to be able
to configure the source to use the advertised number of input buffers. However, if the source fails to start, then the encoder
ends up in the state of EXECUTING when OMXCodec object gets destroyed. As a result, the assertion on the expected state in
OMXCodec's constructor fails. The fix is to stop the video encoder component right way when its source fails to start so to
bring the state of the encoder component back to the expected state.

o related-to-bug: 7045494

Change-Id: I6d4a221eb809d7137f53e58098a04816998f7a25
e671207115fac3914134c61b336d5fa0242c68ca 26-Jun-2012 Andreas Huber <andih@google.com> Make sure codec names and corresponding quirks are returned together.

Previously they were returned in separate vectors and only one of them was sorted if
software codecs were preferred, leaving the quirks no longer matching the codec name
at the same index.

Change-Id: Id3f1e6f9f7f8c9cc4b6ebfb86a203b4d59de8604
related-to-bug: 6737884
a6ee7bc5f9524c220377e37f6525e2d46c69b75b 07-Aug-2012 James Dong <jdong@google.com> Remove obsolete hack for TI specific code

o related-to-bug: 6566886

Change-Id: I39aad214cbf7b748a95a9d22db50cd8f421931e6
8b71241ce7353731ab75322c46e090ee35014a33 27-Apr-2012 Marco Nelissen <marcone@google.com> Fix mono gapless playback for decoders that output stereo

The old AAC decoder always outputs stereo, even for mono source material, so we
need to use the number of channels of the output when calculating the number of
bytes to skip, not the number of channels in the source.
This makes OMXCodec skip the right amount of data, and prevents NuPlayer from
writing half a frame and then asserting when the AudioSink doesn't accept it.
Also move use of the SkipCutBuffer from NuPlayer to ACodec, so that it also
works when using the new Java APIs, and make SkipCutBuffer derive from RefBase.

Change-Id: I34df9fea3e6730617eae559afaa556f4085ef0a0
aeb8fd460ed87d032b3fb8bb61e21eb542ce0f5b 19-Apr-2012 Dave Burke <daveburke@google.com> Add Fraunhofer AAC encoder with AAC-ELD support.

Change-Id: I6cd499d257d72f50a5b508bed97796a591a51506
c1d8115e8a0bdaeb2b723d395b9a85a02c90c933 24-Apr-2012 Andreas Huber <andih@google.com> Make sure OMXCodec and ACodec both accept more than 2 channels of audio

and fill in the OMX channel mask properly.

Change-Id: I915950a0b252142b9eb3277cf7c6e0d9f5875305
69829f3bd09ee4e6be49fee8795c5df24c4da70e 30-Mar-2012 Andreas Huber <andih@google.com> Add a few more APIs to MediaCodecList.

Change-Id: I5ac193cd40c82bbcd87c1e55003b78102e8d4674
ed3e3e046840d5bf1ca84a8c0cc097425e89d6d6 26-Mar-2012 Andreas Huber <andih@google.com> Provisional support for secure decryption of media streams.

Change-Id: Ib3982a9c960bfdb0cb7e1b174440b141b194cfbe
a98478bfbcc0f7fb4b164d3dce40ca96df75667d 21-Mar-2012 Marco Nelissen <marcone@google.com> Support gapless playback for mp3 and m4a

Gapless playback for appropriately tagged mp3 and m4a files.
Currently this is implemented in OMXCodec, which most players
use, but should be easy to support in other players as well by
using the SkipCutBuffer utility class.

Change-Id: I748c669adc1cfbe5ee9a7dea2fad945d48882551
afc16d667afa23f5aa00154ccad62f8c45cf5419 01-Mar-2012 Andreas Huber <andih@google.com> Instead of hardcoding OMX component names in our code, support

a config file instead.

Change-Id: I5835903ab9f1c4a22ccc605ca99ed966767adf57
000e18370baae60ffd9f25b509501dd8c26deabf 07-Feb-2012 James Dong <jdong@google.com> Enable B frame support in MPEG4Writer

This patch allows us to automatically detect whether ctts box is needed in MPEG4Writer.
MPEG4Writer uses ctts version 0 (non-negative offset value) store the composition time
offset on a needed basis.

Currently, the size of the ctts box is not optimized. Optimization will be addressed
in a subsequent patch.

o also changed the private method retrieveDecodingTime(bool) in OMXCodec
to getDecodingTime()

o related-to-bug: 4232183

Change-Id: Ic6dc7b25ecd258c2506ca4b9c25156e922456e51
2b1892db4a2ab47bfc09e59a4c11751aea99e4ea 26-Jan-2012 Andreas Huber <andih@google.com> The software AAC encoder is now an OMX component.


Change-Id: I74938a20b4e0a622836ea5184d3761180eb0f5de
71b61a960ae30df990976ca5dbe780d85a15b0d1 18-Nov-2011 Andreas Huber <andih@google.com> Report a runtime error instead of asserting on malformed avc configuration data.

Change-Id: Ibcd9c3fb5b6532eba843ed80ecdcdacaf50d8845
related-to-bug: 5641069
1765d6e0c962d2b89897c0278f969084f632d181 25-Oct-2011 Jamie Gennis <jgennis@google.com> Stagefright: push blank buffers when tearing down

This change makes OMXCodec push RGB 565 buffers filled with black to an
ANativeWindow when tearing down after decoding to protected gralloc
buffers. This allows the OMX tear down to zero out any protected
buffers that were used without the possibility that the buffer is still
being used by SurfaceFlinger or HWComposer.

Bug: 5483222
Change-Id: I8acedd81a7bb67dfdc2fd15733e3375b6ce8d560
d20bf0676d6db8f12edcd2307f82409c076c7015 30-Sep-2011 Lakshman Gowda <lakshman79@ti.com> Set crop params to ANative Window during Initialization.

The native_window_set_crop() is called when port reconfig event callback comes from decoder's and
crop parameters are changed from default getconfig() OMX_IndexConfigCommonOutputCrop values.
Since the default crop params are same as port reconfig crop params, the native_window_set_crop()
is not called, hence resulting in displaying the whole frame(paddedWidth x paddedHeight).
By calling native_window_set_crop() during initilaization of output port of decoder ensures
in setting up ANative window to crop region.

Change-Id: I68926464a1f5c7e6053804615c8b9bd32ea85688
Signed-off-by: Lakshman Gowda <lakshman79@ti.com>
d3523f89a867afa111bb332887bf006ec3ae93e6 16-Sep-2011 Gilles-Arnaud Bleu-Laine <gilles@ti.com> Gracefuly return on detecting wrong AAC format from corrupted files

Return BAD_VALUE error upon detection of wrongly formatted files.
The client should abort the initialization upon error detection.

The current CHECK() interrupts the configurecodec() preventing a graceful

Change-Id: Ic79313fa76a63284897df5d91635de87d06f3100
Signed-off-by: Gilles-Arnaud Bleu-Laine <gilles@ti.com>
9a5e04438df742c297c5dbcdb60e6fd4d8f8321d 13-Sep-2011 James Dong <jdong@google.com> Change an assert failure due to unsupported level by HW AVC decoder and report the error to applications

Change-Id: I522a994981d86329ccc86c3b536b561dc6e40eeb
related-to-bug: 5300618
17741baf43818ca66c2a88db6b6e252d7abfff49 10-Aug-2011 Andreas Huber <andih@google.com> Revert "Notify the OMX component that it's going to be used in "secure" mode."

This reverts commit 7616178271e95f009452a21ea45e7225997dc27a.
a1f10e8959cd4656aedb2613e855342102e59555 09-Aug-2011 Andreas Huber <andih@google.com> Notify the OMX component that it's going to be used in "secure" mode.

Change-Id: Id87c4b295eb38f7d24045918e73df298d7b842f9
related-to-bug: 5137212
be1cc822304d3a29c106ad5b6f330aed3e8c5658 24-Jul-2011 James Dong <jdong@google.com> Merge "Do not wait forever for output buffers in OMXCodec.cpp and error out in case time out happens"
d35924d9928f29dcee6f5666b5bbd084640c7b34 22-Jul-2011 Jean-Michel Trivi <jmtrivi@google.com> QueryCodecs() signature change

Restore QueryCodecs() signature exactly as used by third-party libs.
Add an alternative function to select HW only codecs.

Change-Id: Ibb2cb4dd36fc0c6599eb93aa5751c216397e3b6f
db4d3717b4677b3239546a248dc0a7d82db4ba50 18-Jul-2011 Jean-Michel Trivi <jmtrivi@google.com> Merge "Add option to query hardware decoders only"
df57a0d6d3b0609e69ff79afa63fe44dfa91f6f5 18-Jul-2011 Jean-Michel Trivi <jmtrivi@google.com> Add option to query hardware decoders only

Add an option to OMXCodec::QueryCodecs() to filter out
software codecs.

Update stagefright command line tool to list the roles of codecs.

Change-Id: Icddb79118c30fe13cc3aea20f340174cc082c4e0
a57a9a491272aa884494b2ec7854960827a73742 16-Jul-2011 James Dong <jdong@google.com> Do not wait forever for output buffers in OMXCodec.cpp and error out in case time out happens

o Deal with vendor codec hang bug

Change-Id: Ic8449afd43045f09a9e0bd3d1be9a320e59ccabe
457116d3a01618acf9a875020ca5860551ba03a6 11-Jul-2011 James Dong <jdong@google.com> Support profile and level query for SW video decoders

Change-Id: I9c20db128be96cd36cf2083e08e8c21f5e6d1fdf
0bb6b4c5da2451ee0973372b0a3858b15c742689 30-Jun-2011 Andreas Huber <andih@google.com> Support a "secure input buffer" mode for use by Widevine DRM in OMXCodec.

Change-Id: Ie03b285265099dbac9b12f3f41d9ad758d0b50ad
1065b3f17d3048948e7d522049d1980b90df3dc1 28-Jun-2011 Andreas Huber <andih@google.com> Multiple changes to ACodec/codec tools:

Make sure sf2 does not coalesce input buffers, generalize ACodec's codec
instantiation based on OMXCodec's list of eligible component names.

Some changes/additions to the "sf2" commandline tool

Make surface options consistent with stagefright tool, i.e. use '-S' instead of '-s'
New option '-R' renders surface-allocated buffers.

Also fixes a longstanding bug introduced when generalizing from surfaces to native windows that never used surface-allocated buffers in sf2 even when the option was specified.

Change-Id: I59fd533f0f6ef0337ebe2806ddc81a46878eb3ae
965e4239ca1cf6c824c1f8ce23116f9ba8cf6ebd 08-Jun-2011 James Dong <jdong@google.com> Add B frame support for MPEG4Writer

o requires the support of negative ctts duration values (ctts version 1)

Change-Id: Ib14130c9359c3bff3c76f20a7380d468a065dcaf
bbba88cb1bdc34705d1477208990a06904c022e7 11-May-2011 Andreas Huber <andih@google.com> Squashed commit of the following:

commit c80992e419ed567abef451042f09c4958534b90d
Author: Andreas Huber <andih@google.com>
Date: Wed May 11 14:00:07 2011 -0700

Support for the mp3 audio decoder as a software OMX component.

Change-Id: I66e10c4d0be4c3aecdef1c21b15a2c7359ceb807

commit a358d0e1bf2a88897887445f42ccdda0f5f2f528
Author: Andreas Huber <andih@google.com>
Date: Wed May 11 13:11:23 2011 -0700

Support for G.711 alaw and mulaw decoders as software OMX components

Change-Id: Ia5c76c02cb83a9f94ce39a27b2251e5880218f03

commit 79088b9c9a5c8b8c97ea66cb4f90a2b0f0d34553
Author: Andreas Huber <andih@google.com>
Date: Thu May 5 15:43:32 2011 -0700

Instead of using an RGB surface and conversion yuv420->rgb565

convert from OMX_COLOR_FormatYUV420Planar to HAL_PIXEL_FORMAT_YV12 instead.

Change-Id: I8c4fc3c54c963f0d4ba6377f3c4ab4e0013152e5
related-to-bug: 4394005

commit 69469d3bd84425777b11b9fc938c5e0c61af26a7
Author: Andreas Huber <andih@google.com>
Date: Tue May 10 15:46:42 2011 -0700

voip mustn't link against libstagefright.so

Change-Id: I4d0ba9a8b9dc9380b792a1bd04bcda231964862c

commit 2a9a9eeeeeb36ae3a9e680469c3016d509ff08c3
Author: Andreas Huber <andih@google.com>
Date: Tue May 10 14:37:10 2011 -0700

Remove most non-OMX software decoders by default

Change-Id: Ic56514bc1b56b8fa952e8c4a164ea7379ecb69d0

commit a4de62c37b335c318217765403a9fb282b20a216
Author: Andreas Huber <andih@google.com>
Date: Mon May 9 16:50:02 2011 -0700

Conditionally build the old-style software decoders.

Change-Id: I5de609e1d76c92d26d6eb81d1551462258f3f15f

commit 5d8b039f9449dc3dad1e77c42c80cc0b54b0c846
Author: Andreas Huber <andih@google.com>
Date: Mon May 9 16:13:12 2011 -0700

Support for MPEG4 and H.263 video decoders as soft OMX components.

Change-Id: I5e3a4835afab89f98e3aa128d013628f5830eafe

commit b25a1bfbeb0ff6e62e1cc694ce2599c91489c7d0
Author: Andreas Huber <andih@google.com>
Date: Mon May 9 11:49:10 2011 -0700

Boost Soft OMX thread priority, fix timestamp handling in vorbis Soft OMX decoder.

Change-Id: I68d26d4999f06fcc451d69e5303663fab0cba9e8

commit c0574362f8dc3319ce84d981097867062a698527
Author: Andreas Huber <andih@google.com>
Date: Mon May 9 11:28:53 2011 -0700

Support for the AMR decoders (NB and WB) as Soft OMX components.

Change-Id: Ia565f59833fb52653e23f26536e7e41fc329a754

commit 3e5575a8f0e27a490cb7bde77bd9456087837f08
Author: Andreas Huber <andih@google.com>
Date: Wed May 4 13:41:25 2011 -0700

Signal an error if the aac decoder failed to initialize from codec specific data.

Change-Id: I01da7831bdf722edd7d6dc5974486daa2cf2b209
related-to-bug: 4272179

commit f94aeaa9886e772ff4823e671ed237096649f4af
Author: Andreas Huber <andih@google.com>
Date: Tue May 3 13:07:38 2011 -0700

Software OMX nodes don't (yet?) support native_window mode.

Change-Id: I7d9ca9164ef4abf66b573ca21dba12d672f8b12d

commit eefdfabac8dc659e00daa56da69aea705c49cb67
Author: Andreas Huber <andih@google.com>
Date: Tue May 3 12:57:16 2011 -0700

Fixing the OMX tests to refer to appropriate files from test content.

Change-Id: I5b61c3498749bfb876abbd3946a5132356e3f6ff

commit f31b7326aef14b6a1b7946520a9688f092e844d5
Author: Andreas Huber <andih@google.com>
Date: Tue May 3 11:08:38 2011 -0700

Soft OMX components are now dynamiclly loaded/unloaded, not directly linked against.

Change-Id: I1e2ecfbfab67a8869886f738eaf0c7b3c948b6d9

commit b7f0343879e4df06f0a1c9bfece24df557954e2f
Author: Andreas Huber <andih@google.com>
Date: Mon May 2 15:58:36 2011 -0700

Support for the AVC software decoder as an OMX component.

Change-Id: I13c12df435ba4afbd968a9fc659f66b91c818bc2

commit 5bb9e616d6c8e1b13d531fe996b9a9affdfb2977
Author: Andreas Huber <andih@google.com>
Date: Fri Apr 29 12:05:37 2011 -0700

Fix Vorbis OMX decoder's component role.

Change-Id: I5e871e5e11b3f951c93590210e63fd7987c467b5

commit 089c91f2333062e196c7afd5fb0ca914878aa474
Author: Andreas Huber <andih@google.com>
Date: Fri Apr 29 12:05:18 2011 -0700

Support vorbis_decoder OMX testing.

Change-Id: I1985be178a12ae3f8768bc72067d9236238be170

commit 56e241fa36fc37219bc536b823bdc2ab82dc1fad
Author: Andreas Huber <andih@google.com>
Date: Fri Apr 29 12:01:46 2011 -0700

SoftVorbis OMX component now respects the number of valid frames per page.

Change-Id: I82a117a064d9b083fc58a54ad900a987a763ef03

commit fcd618ec520c376fdb78f4cbb44b8d9f5d213e2b
Author: Andreas Huber <andih@google.com>
Date: Fri Apr 29 10:59:38 2011 -0700

Support for the vorbis audio decoder as a soft OMX component.

Change-Id: Iaeb057e58ca306d3dce205c0445b74d5aefef492

commit d1fcc3203fc8003ad79c6e96b3a1fc4261743f16
Author: Andreas Huber <andih@google.com>
Date: Fri Apr 29 10:07:50 2011 -0700

VPX decoder now properly resizes buffers after a port settings change.

Change-Id: I110749a31b6cba087891d8e5dfe420830bdbf831

commit 35c7168243cb69849d88911144a2c7fdfed5c54e
Author: Andreas Huber <andih@google.com>
Date: Thu Apr 28 13:23:34 2011 -0700

Support for the VPX video decoder as a Software OMX component.

Change-Id: Ic345add2d6d768d4af631160153f2e9b97fcea71

commit 923b2534b4211fc5405377b5190bfa6f2dd27f32
Author: Andreas Huber <andih@google.com>
Date: Thu Apr 28 11:34:40 2011 -0700

Table-based registration of soft omx components.

Change-Id: I7f45f0fa5b3a7950776e69c66349731f7674e937

commit 04a88f3edb2266a463da9c4481b80178be460902
Author: Andreas Huber <andih@google.com>
Date: Thu Apr 28 11:22:31 2011 -0700

Apparently OMX_GetParameter is valid in any state other than OMX_StateInvalid

OMX_SetParameter is still constrained to OMX_StateLoaded or a disabled port.

Change-Id: I1032d7cf4011982d306aa369d4158a82830d26fb

commit 9d70ca68445e7c40f5c9b2d12466e468f514de88
Author: Andreas Huber <andih@google.com>
Date: Wed Apr 27 15:03:18 2011 -0700

Use the new soft OMX aac decoder for HTTP live playback.

Change-Id: Ifbcfb732a9edb855cb46b49f6d0ac942170ee28f

commit 213fe4a10ea93cce08e8622dc3908053f29878a1
Author: Andreas Huber <andih@google.com>
Date: Tue Apr 12 16:39:45 2011 -0700

Foundation for supporting software decoders as OMX components

Change-Id: I7fdab256563b35d1d090617abaea9a26b198d816

Change-Id: I83e9236beed4af985d10333c203f065df9e09a42
2eb62955eb84b97695e8a7e56e14310cbb86412b 28-Jan-2011 Glenn Kasten <gkasten@google.com> Bug 3012968 DRM output control

Modifies Stagefright to verify that there is a hardware-protected path
to video sink for DRM content.

Change-Id: I18b8741390e803a05a88c7f180b860a24ba88a10
9c0096378820e5a61db26e52a7e6df50ba9c872d 08-Feb-2011 Andreas Huber <andih@google.com> Handle some edge cases when seeking while starting up OMXCodec

These were exposed by the new preview-seekframe while paused code.
In particular, the codec may have been in state RECONFIGURING when attempting
to seek, or we may have initiated flushing of the output port and this may not
have completed yet by the time we want to reconfigure the output port.

Change-Id: Id7640ade11dbc7205a22f648ea0b5e3e9b49cf4b
related-to-bug: 3392259
7757f5010a771fb8824b6fdf9788f588a1577e3f 26-Jan-2011 James Dong <jdong@google.com> Make sure that key frame is generated for timelapse video recording if there are at least two input video frames from camera source.

This will fix the stop failure issue where we have to wait n * time_interval before a key frame can be received by the file writer, where
o n is the actual number of buffers advertised by the video encoder
o time_interval is the interval settings for timelapse video recording
specifying the time distance between neighboring input video frames

The fix includes two parts:
o OMXCodec will not submit all n buffers at one time, but instead submit one input
frame at one time if it become available.

o Timelapse camera source made available the first two input frames and do not skip
them so that the first compressed output frame data can be received regardless
the specified time_interval

bug - 3367659

Change-Id: Ia68cc2cb0d71aa7dc54540e9ad82fae911ad530b
7fac331b39ca49ce49a67e425dcc031a3cb9e97f 21-Jan-2011 Andreas Huber <andih@google.com> Properly rotate video that's marked as such and decoded to a surface.

Change-Id: I1e9144db3447e58c99aac3f47702ad471678789c
related-to-bug: 3378148
14c858e80dfe2030c9f343dc0c6e2048e030731b 10-Jan-2011 James Dong <jdong@google.com> Avoid deadlock in OMX::freeNode by making sure OMXCodecObserver does not hold the last reference of OMXCodec object

bug - 3336424

Change-Id: I4c79b66a900c527e3ae6a833f76d5da1b75c5a89
79e23b41fad961008bfde6e26b3c6f86878ca69d 11-Dec-2010 James Dong <jdong@google.com> Revert "Allows the authoring engine to skip frame."

o Skipping frames could lead to a lot of issues such as I frames is lost etc.
It is not being used anyway.

This reverts commit 53d4e0d58e2d5c18f6e026c705af833b9bdd7aba.



Change-Id: I3abba1647de48db25bdc369066eb2a7ae4dedec2
92bf2f96c53d24adc1ace362439e82ca2cf6b856 07-Dec-2010 Andreas Huber <andih@google.com> Better buffer status management and verification in OMXCodec.

Change-Id: I90410f2ac0d8ff86076a239dc6e281b1bc4d8643
b50a8033f1c7c2b58913212825f9200f1a9e5652 11-Nov-2010 James Dong <jdong@google.com> Add color format query support to QueryCodecs()

Change-Id: Ic8589649cd09392a1b969a30082b4c9c4e6cc6a7
f23c4f92c3b0202435cf87db2642156fabc46f02 11-Nov-2010 James Dong <jdong@google.com> Fix memory leak when 0-memcpy quirk is used

o minor change to the original patch by Andreas:
call restorePatchedDataPointer() method only if kAvoidMemcopyInputRecordingFrames is turned on

Change-Id: Idf3710e6f759d37d28e866613f98d39215722cb9
8480835b4bc1350646376aa7f3ae33742a7adeb1 02-Nov-2010 James Dong <jdong@google.com> Use meta data in the media recording framework

o This patch allows us to do 720p video recording

Change-Id: I2ea37e80a59630145396b08ebcdc6ee71df53333
ed45fe0730a7e7fc9944741428f5a484350acc8a 02-Nov-2010 Jamie Gennis <jgennis@google.com> Fix a MediaBuffer leak in Stagefright.

Change-Id: I548e60b07cf1676476874b156cfbc4ffefdfa2b9
d332a72e8f71caea0d3dc898db6a9d7a929fad32 23-Oct-2010 James Dong <jdong@google.com> Add two creation flags to OMXCodec::Create()

o This allows to force to use software codecs or hardware codecs
o If request cannot be fullfilled, Create() returns NULL.

Change-Id: I02b56a9229abb56d49703fe80ac18571d33f3748
6a9da9fc558263548ebfbae2cbf177eb7454a41b 07-Oct-2010 Jamie Gennis <jgennis@google.com> Add decode-to-ANativeWindow support to Stagefright.

This change adds support to Stagefright for doing OMX video decoding directly
into buffers dequeued from an ANativeWindow. It does this by registering the
dequeued buffers with the OMX component using an Android-specific OMX
extension, and then exchanging buffers between the OMX component and the

Change-Id: Ida66f836503255a68d378c6903d96dfe9747ce87
f3712f026aad1fc46b1df18d1dba718281e39726 18-Oct-2010 Andreas Huber <andih@google.com> For thumbnail extraction make sure we instantiate a decoder that allows access to the framebuffer. Implement the samsung workaround to support this by reconfiguring the decoder.

related-to-bug: 3106534
Change-Id: Ie28d72af2f9e93818d1840ac83aa7bc11fa57b3b
2a4d22d79e927f2245537921e10fc5fda1c47a29 08-Sep-2010 Andreas Huber <andih@google.com> Work to support switching transport streams mid-stream and signalling discontinuities to the decoder.

Change-Id: I7150e5e7342e1117c524856b204aadcb763e06ed
related-to-bug: 2368598
2b82e9652ba049e754c2cc74e381282f231d5fbf 07-Oct-2010 Andreas Huber <andih@google.com> On this particular device the hardware video decoder spits out buffers that don't actually contain our video data, so we cannot use them to restore the video frame after suspend/resume.

Change-Id: I1b8fe68c1766299844fe84ebbff49cb8b3e4cc7c
related-to-bug: 3070094
824c9ff4a55e86faae4f8f158977329909cbfaf6 27-Aug-2010 James Dong <jdong@google.com> Workaround for a QCOM issue where the output buffer size advertised by the AVC encoder
is occasionally too small.

bug - 2882917

Change-Id: Id59d8529084c5689a26f272e0cd3b1e955fd8a30
299766cdbf93fba14634e364b177b7f4d5811453 04-Aug-2010 James Dong <jdong@google.com> Use the target color format from the camera source if possible

- If the target color format is unavailable, the default
platform-dependent color format will be used.

- Also add some logic to prevent looping forever if the
omx component is buggy supporting color format enumeration.

Change-Id: I119a78f0d6201b4c3621235cca2f523ec14e24e3
542db5d438988360d491a5add1040a2df9aa90c9 21-Jul-2010 James Dong <jdong@google.com> Allows the authoring engine to skip frame.

This is 1st part of the work to allow audio and video resync if
we found out that audio and video are out of sync during authoring

- also fixed a problem in AACEncoder::read() where the buffer acquired
from the buffer group does not release when error out at
reading from source.

Change-Id: I8a2740097fcfdf85e6178869afeb9f3687a99118
abd1f4f870925d6776dbe4b930b759a1ab6595ca 21-Jul-2010 Andreas Huber <andih@google.com> Support finer seek control on MediaSources.

related-to-bug: 2858448

Change-Id: Ifb4b13b990fd5889113e47e2c62249ac43391fa1
145bfe5eb3e08c9689c28f6bf3287a979438b04b 02-Jul-2010 James Dong <jdong@google.com> Allow application to set two more encoding paramters: video profile and level

Change-Id: I673e681cefe184d5c556c612c54600a24a2143e5
d329e21495eda9dbc531fdd0c26c77f1593ac3f4 30-Jun-2010 James Dong <jdong@google.com> Added encoding parameters set up for H263 video encoder

- Allowed start() call when encoder already starts and stop() call when encoder has not started yet
- Handled default value for audio/video sources/encoders and file output format

Change-Id: I03b2f7d3cf570baa0fd011a8c0ad200f2f2a5da1
d35bd5fb4e09c2cd8608497c279cbb2ef9c3a029 10-Jun-2010 Andreas Huber <andih@google.com> Support "pausing" of MediaSources with the effect that they no longer pull on their upstream source until a subsequent read-with-seek.

Change-Id: Ie4153a10ab36c1135f5fcfb572958129d886bcc3
ddcc4a66d848deef6fb4689e64e30cd9bd2684fe 08-Jun-2010 James Dong <jdong@google.com> Remove some hard-coded encoding parameters

Change-Id: I7a8ccd5d57891a6a585c8da2ee53acb094955913
d07139e2e817a9b3ae9c87ba4e1e8d65d3e549da 07-Jun-2010 James Dong <jdong@google.com> Second part of speeding up video recording frame rate

1. Avoid copying the input recording frames to the encoder via OMX interface
for TI video encoder
This is a missing change for part one which help reduces the CPU load.

2. Release output buffers as early as possible. This is a little bit helpful, but not critical.

We should save the underlying pointers allocated by the OMX component before we replace them
and restore them before we call OMX_FreeBuffer()!

Change-Id: Ib3a88978f4c3b1153808872eaa7ac4c265a811ff
956c553ab0ce72f8074ad0fda2ffd66a0305700c 15-May-2010 James Dong <jdong@google.com> Initial software encoder checkins

Change-Id: I27f387db23594e46384c4eb3a0093ce220bb6b60
7f498b90a4300ef9badf14d202b0a67c26e20931 02-Apr-2010 Andreas Huber <andih@google.com> Coalesce multiple encoded AAC frames into a single input buffer on this particular OMX codec to increase throughput significantly.

Change-Id: I90c7db6656a53339c5d454336548c4f00d0d9064
related-to-bug: 2548426
2a3847ee1cbdaa8a65eee397a0173bb02211c459 16-Mar-2010 Andreas Huber <andih@google.com> Properly handle errors during the codec configuration phase, attempt to revert to next available (likely software-) codec if configuration fails.

Change-Id: Id1c699711e30139c9cc29df972254b5ba026e6fb
related-to-bug: 2517098
5295c0c55d41a2906ea7f65a3f22e6278cb17d4b 23-Feb-2010 Andreas Huber <andih@google.com> Propagate errors all the way through the MediaSources and send either MEDIA_PLAYBACK_COMPLETE or MEDIA_ERROR depending on the final reason for running out of buffers to play back.

related-to-bug: 2463749
86559de0275b46a9bdba06b749fbd549ad2022d1 22-Feb-2010 Andreas Huber <andih@google.com> This hardware video decoder lies about its required input buffer sizes allocating 2.7 MB of memory instead of the required 176 KB... Added another quirk.

related-to-bug: 2281327
78d529eb330ab6c04b5c694403f5a7e7de4b702f 01-Feb-2010 Andreas Huber <andih@google.com> The TI MP3 decoder lies about the number of channels it outputs, add a quirk for that.
213addfaf4b359c69da4e9b4490c511d116845bb 25-Jan-2010 Andreas Huber <andih@google.com> The qcom OMX video decoders do not allocate output buffer memory at the time OMX_AllocateBuffer is called, wait until we received the first FILL_BUFFER_DONE notification until we rely on the buffer data ptr.
570a3cb7582daa030cb38eedc5eb6a06f86ecc7f 21-Jan-2010 Andreas Huber <andih@google.com> Fix no-copy-overhead OMXCodec implementation to actually work.
7eaa9c9385535b651064e02d05a8ffa4b2359281 16-Jan-2010 Andreas Huber <andih@google.com> Avoid unnecessary buffer copying if at all possible, detect if running in the mediaserver process.
bf697e37550d9e8376089b0b5e498613bede798c 07-Dec-2009 Andreas Huber <andih@google.com> Change OMXCodec::Create to return an sp<MediaSource> instead of an sp<OMXCodec>, this is more general and does not sacrifice any functionality as the remaining OMXCodec APIs are not meant to be public anyway.
c297fccffc4ab1cb3b9f5c6a5b0802be057f3e0f 02-Dec-2009 Andreas Huber <andih@google.com> A small sample tool to encode pcm audio data to amr, decode it again and play it. Some changes to OMXCodec to properly configure the AMR decoder(s).
30ab66297501757d745b9ae10da61adcd891f497 17-Nov-2009 Andreas Huber <andih@google.com> Squashed commit of the following:

commit 333057b355f8c260c549553b9a0634755c838b6a
Author: Andreas Huber <andih@google.com>
Date: Fri Nov 13 15:35:48 2009 -0800

Some more tweaks to AVC encoding on sholes.

commit 9981d0ee52ec5b8b0182aae733d1571e3ebb8390
Author: Andreas Huber <andih@google.com>
Date: Thu Nov 12 16:36:57 2009 -0800

Support for avc encoding, including sholes specific tweaks to pick the right colorspace for the camera to not require transcoding.

commit 5ba0ebbbd4efca51f3ae1f60e2ca31e7d2cf136d
Author: Andreas Huber <andih@google.com>
Date: Wed Nov 11 09:50:03 2009 -0800

Enable actual (camera) video-only recording using h.263 or mpeg4 encoding.

commit 3fd59c3526a37fe7c696f4a978925d1831c09313
Author: Andreas Huber <andih@google.com>
Date: Tue Nov 10 14:57:48 2009 -0800

Allow switching between the PV recorder implementation and one supported by stagefright.

This is controlled through the property "media.stagefright.enable-record".
888f72a4628eed3c136ff2e346a15d129da0d520 29-Oct-2009 Andreas Huber <andih@google.com> Revive support for video encoding in OMXCodec.
ba229b3faa98b3bb6567cf95040062be2d9c2b3f 22-Oct-2009 Andreas Huber <andih@google.com> Prefer software decoders over hardware for thumbnail extraction.

While our hardware decoders clearly outperform the software decoders in terms
of raw throughput, their startup latency makes them less suitable for thumbnail
c6323f9d72e0f08dd1e813550a56c8d2f0fd76a2 21-Oct-2009 Andreas Huber <andih@google.com> Some cleanup of OMXCodec code, make sure to ignore 0-length output buffers in the metadata retriever.
7f281f87670c63775ea1ae8b24af31822d5ad416 09-Oct-2009 Andreas Huber <andih@google.com> OMXCodec now notifies the reader of changes in the output format by returning a special result/error code.
9749025189acc6c737be22007e90ed8bbbcdbc21 08-Oct-2009 Andreas Huber <andih@google.com> The vendor is about to check in a fix for this issue, the workaround is no longer needed
ec430b158df17b2c766f7750a96c35d3902f5f49 07-Oct-2009 Andreas Huber <andih@google.com> Workaround for avc decoder misreporting output buffer size requirements if the content is not a multiple-16 width/height.
18291bc20e55e8f3fd5feb786771a8ed32c19c59 10-Sep-2009 Andreas Huber <andih@google.com> Media MIME types now have corresponding constants in MediaDefs.h, new codec enumeration API.
456db75516efc889e1ee4e5e16021e77c03b0941 08-Sep-2009 Andreas Huber <andih@google.com> Support encoding amr-wb content in stagefright.
7a6b9e2eca7d20457ace3538c689640e5bfda4f3 03-Sep-2009 Andreas Huber <andih@google.com> Setting the component role appears to be mandatory now for all mime types.

Also using CODEC_LOGV instead of vanilla LOGV to include the component name the message related to.
e5adbeee3401915ff8e1a983396ce3554436b11c 02-Sep-2009 Andreas Huber <andih@google.com> Some work to make audio encoding work.
37940eefcba982836b579fe1ffec6cada72b0974 02-Sep-2009 Andreas Huber <andih@google.com> Squashed commit of the following:

commit c45bfbb97ccd05982008df47181f9c73abaf0497
Author: Andreas Huber <andih@google.com>
Date: Tue Sep 1 15:58:12 2009 -0700

This quirk should not be enabled by default in order to make the bug reproducible by the vendor.

commit 21d72e80e795fcae53d9c3bcc8ba6312b081e420
Author: Andreas Huber <andih@google.com>
Date: Tue Sep 1 15:55:45 2009 -0700

Undoing the hack to temporarily give up the lock to facilitate reading from the buffer source.

This simply causes too many issues, there need to be independent threads providing input buffers and dequeuing output buffers.

commit 84d507def8999c146ce124cc8edfe106c9ca70c2
Author: Andreas Huber <andih@google.com>
Date: Tue Sep 1 15:16:23 2009 -0700

The AAC components appear to output stereo data even if the input data is mono...
284f513a4ce2c5aa0e250c07c873731aedb0be26 27-Aug-2009 Andreas Huber <andih@google.com> An attempt to fix a deadlock between OMXCodec::start and upstream ::read
8297cedd12f689167d74de5ddacb514e9f6896ea 27-Aug-2009 Andreas Huber <andih@google.com> Another quirk, some OMX components fail to properly flush before shutdown and we must do it for them...
ea7d15629752256f6ea1d5c6ea796e59aefd030f 26-Aug-2009 Andreas Huber <andih@google.com> Keep up with latest changes to overlay handling on that unnamable platform.
1beb760d920561679862ded945a04e370368c7f7 25-Aug-2009 Andreas Huber <andih@google.com> Make sure not to ask for more buffers when we know that there won't be any, added a quirk for the aac decoder.
7530e9c708275c273c134c36c68179f511c1940e 20-Aug-2009 Andreas Huber <andih@google.com> Dynamically allocate a pair of MemoryHeaps according buffer count/sizes required by the OMX component, respect JPEG compressed size.
dbc03445db2bbf83b64f0c0a5dc62e61408864d7 19-Aug-2009 Andreas Huber <andih@google.com> The PV AVC software decoder expects each fragment of a NAL unit to be handed to the decoder unframed.
693d271e62a3726689ff68f4505ba49228eb94b2 14-Aug-2009 Andreas Huber <andih@google.com> Squashed commit of the following:

commit 5bb012f0065f7ffaaeb4f569d71f0e3a8d6b19c3
Author: Andreas Huber <andih@google.com>
Date: Fri Aug 14 10:40:08 2009 -0700

An attempt at fixing export using the qcom encoders. More quirks.

commit 0690e76bfa48118a68287ccf1bbfa82febaa620c
Author: Andreas Huber <andih@google.com>
Date: Fri Aug 14 09:08:28 2009 -0700

Callbacks are now dispatched from a separate thread in OMX.

commit c6571a039526df29b6343f9a1971dbc019088c61
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 13 15:42:25 2009 -0700

Massive API changes throughout stagefright, smart pointers everywhere.

commit 900612af6a0555664d9ba195112cd859491265f4
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 13 13:33:12 2009 -0700

OMXCodecs now properly shutdown.

commit 96732f05e1b0603dcd1b11f16a23512592eeb4f5
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 13 12:04:04 2009 -0700

More work on JPEG decoding using the hardware OMX component.

commit 63839a073ac393e3a130434ba467969053b694ad
Author: Andreas Huber <andih@google.com>
Date: Wed Aug 12 13:13:31 2009 -0700

An attempt to drive the JPEG decoder OMX node.

commit 3ac2fe5ab2926eda81b2123610b2434c645294ff
Author: Andreas Huber <andih@google.com>
Date: Tue Aug 11 16:38:21 2009 -0700

Renamed StateMachine to OMXCodec and put it in its proper place.

commit 247da75a96bf8881956413023dd49a84d5b4f5b2
Author: Andreas Huber <andih@google.com>
Date: Tue Aug 11 16:06:19 2009 -0700

Statemachine is now a full-fledged MediaSource.

commit 045244f6771fa0b9b329495c953afda900a84b71
Author: Andreas Huber <andih@google.com>
Date: Fri Aug 7 09:16:54 2009 -0700

Properly setup the input format when exporting to AMR audio.

commit 271b984cb32c5cd9e46e3f90ae121f334e4b8da9
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 6 09:59:38 2009 -0700

Added some code to test audio encoding to the OMX harness.

commit 79af4748e4af33bd66d3fbac606e332a69741cf4
Author: Andreas Huber <andih@google.com>
Date: Wed Aug 5 14:36:22 2009 -0700

Merge the old OMXDecoder and the new, shiny, StateMachine code.

commit 91cf5dd77a8762bc10a0b2ffce35e3bbeb262231
Author: Andreas Huber <andih@google.com>
Date: Tue Aug 4 17:41:43 2009 -0700

A new harness to test OMX node compliance (and quirks).