generic-library/openh264
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
3,485 Commits 23 Branches 9 Tags 55 MiB
e9916c7592
Commit Graph

2150 Commits

Author SHA1 Message Date
sijchen
d7af7cedf9 Merge pull request #1845 from ruil2/rc_check 
use each spatial layer qp instead of global qp and update traces
 2015-03-12 17:27:24 +08:00
ruil2
fd7cd76b60 use each spatial layer qp instead of global qp and update traces 2015-03-12 15:05:51 +08:00
huili2
a93e4647df remove unused as264_common.h in decoder 2015-03-10 15:48:37 +08:00
Martin Storsjö
0ef69e55e2 Set pfMeCost to NULL when it isn't initialized 
This avoids using uninitialized data in the check for bSatdInMdFlag.
 2015-03-05 10:29:40 +02:00
huili2
4048b9791a Merge pull request #1834 from sijchen/check_param 
complete parameter checking related to loop-filter-idc
 2015-03-05 09:39:46 +08:00
huili2
3022d4f655 Merge pull request #1832 from sijchen/imp_mc 
[Encoder] replace conditional judgements with faster operations
 2015-03-05 09:39:31 +08:00
huili2
06b8e1abb7 allow slice-level data come in for parse only 2015-03-04 00:57:43 -08:00
dongzha
5b12578960 Merge pull request #1829 from syureyi/wp_p 
Decoder WP support
 2015-03-04 14:58:40 +08:00
zhuiling
9ef0c74fdd Decoder WP suppor 
missed some part

fix clip value

update according to wayne's suggestion
 2015-03-04 09:49:32 +08:00
Sijia Chen
94d1976a4e complete parameter checking related to loop-filter-idc 2015-03-03 18:31:37 +08:00
sijchen
5a1aedb622 Merge pull request #1827 from dongzha/cleanlog 
modify decoder log
 2015-03-03 11:01:46 +08:00
dong zhang
153dd0d3f5 format the files 2015-03-03 09:14:27 +08:00
Sijia Chen
63926a3d10 put a flag calculation in higher-level to save calculation in MB level 2015-03-02 17:36:55 +08:00
Sijia Chen
303bf77e42 replace conditional judgements with function pointer for acceleration 2015-03-02 17:11:11 +08:00
dong zhang
d28970d625 modify decoder log 2015-03-02 16:55:37 +08:00
Haibo Zhu
afbb499c77 Add Bs Lenght File Input in Console 2015-03-01 18:22:40 -08:00
Martin Storsjö
e8cdbd2ea7 Don't downsample to anything smaller than 4x4 pixels 
This makes for a chroma plane of 2x2. The SIMD versionf of generic
downscalers assume that the width and height is at least 2, since
it does an unconditional loop for the body of the image, and a
separate step for the last pixel and last row. The SIMD versions
assume that (width-1) and (height-1) are larger than zero.

This fixes spurious crashes in EncodeDecodeTestAPI.SetOptionEncParamExt.
 2015-02-28 14:58:53 +02:00
sijchen
2294b786cf Merge pull request #1774 from huili2/dec_flowchart_ec 
modify decoder flowchart for non vcl AU check
 2015-02-27 17:44:03 +08:00
Martin Storsjö
861f54887d Avoid trying to downscale layers to dimensions below 2 
When calculating what resolution to actually downscale to,
it can end up smaller than what the caller set. When scaling
down to resolutions close to the limit of allowed values,
this can end up setting values lower than the limit.

Previously, e.g. a downscale from 2266x8 to 566x2 will end
up as 566x1 after this calculation. When scaling to a
566x1, the chroma plane gets a height of 0, which doesn't
make sense, and which breaks e.g. the SSE2 scaler. Therefore,
make sure none of the dimensions end up set below 2.
 2015-02-19 14:53:02 +02:00
ruil2
e29ceca392 Merge pull request #1813 from sijchen/savc61 
[Encoder] complete logic of simulcastavc with sps_pps_listing
 2015-02-13 16:31:18 +08:00
Sijia Chen
15723e1204 add comments for write para set functions 2015-02-13 15:48:17 +08:00
HaiboZhu
1641ce9ee5 Merge pull request #1812 from huili2/cropping_error 
return error for cropping data error
 2015-02-12 11:27:48 +08:00
Sijia Chen
f9080ac090 complete logic of simulcastavc with sps_pps_listing 2015-02-11 17:31:17 +08:00
huili2
4e9e5f5c36 Merge pull request #1794 from syureyi/monochroma 
to support monochroma
 2015-02-11 16:10:07 +08:00
huili2
5683393f38 return error for cropping data error 2015-02-10 22:31:24 -08:00
Sijia Chen
943981abc4 fix a nal generation for simulcast avc and add nal-type check for simulcast avc UT 2015-02-09 13:50:31 +08:00
sijchen
bea361a1f0 Merge pull request #1805 from syureyi/fix_android_demo 
Maintain android demo of encoder, fix a typo error
 2015-02-09 11:05:17 +08:00
sijchen
9008b15d94 Merge pull request #1799 from mstorsjo/copy-spatial-downsample 
Allow forcing copying in downscale
 2015-02-06 14:22:18 +08:00
ruil2
8fc2af760f Merge pull request #1804 from sijchen/fix_over3 
[Encoder] fix potential overflow under large bit rate
 2015-02-06 14:22:01 +08:00
ruil2
296ab4d490 Merge pull request #1802 from sijchen/savc50 
[Encoder] fix for in-middle changing simulcast avc setting
 2015-02-06 14:21:47 +08:00
Sijia Chen
431bcee310 1, update the max-nal-size setting in UT and param check since we are using a larger input check 
2, fix potential overflow (will change bs but little impact on bs)
 2015-02-06 13:24:20 +08:00
ruil2
8246225eed Merge pull request #1791 from sijchen/fix_nal1 
[Encoder] add a para check for max_nal_size
 2015-02-06 12:17:40 +08:00
sijchen
3f19534dd6 Merge pull request #1790 from mstorsjo/simplify-vcproj 
Remove the intermediate SW folder in the decoder vcproj files
 2015-02-06 09:49:49 +08:00
Martin Storsjö
0a2cd1e59a Allow forcing copying in downscale 
Normally, the DownsamplePadding skips scaling if the target
size is the same as the source size, assuming that the caller
will use the source data pointer in that case. This is true
for the base layer (the first call to DownsamplePadding in
SingleLayerPreprocess), but when downsampling the other layers,
there is no special handling for the case when the target
is the same size as the source.

Previously, the encoding of such spatial layers will use
completely uninitialized data, encoding complete garbage.

Instead force DownsamplePadding to make a copy if no scaling
is required, for the dependency layers. The base layer still
avoids a copy unless scaling of that layer is required.

Whether it actually makes sense to have lower spatial layers
the same size as the original one is a different question
though - currently the code allows it, and
EncodeDecodeTestAPI.SetOptionEncParamExt will try to use it.
 2015-02-05 13:10:39 +02:00
Sijia Chen
bc1b28b5dd fix for in-middle changing simulcast avc setting 2015-02-05 16:34:28 +08:00
zhuiling
0a8a4f2284 to support monochroma 
add mbtype judge logic
 2015-02-04 15:00:54 +08:00
HaiboZhu
442230a91d Merge pull request #1789 from huili2/bugfix_checkspsactive 
bug fix for multi-layer active sps check
 2015-02-04 09:13:18 +08:00
Sijia Chen
2211324e2b add a para check for max_nal_size 2015-02-03 15:14:36 +08:00
Martin Storsjö
87b73257cb Remove the intermediate SW folder in the decoder vcproj files 
This makes the folder structure in these project files match all
the other ones.
 2015-02-03 09:09:58 +02:00
huili2
1734133012 bug fix for multi-layer active sps check 2015-02-02 22:31:50 -08:00
Martin Storsjö
4f237b51ef Move parts of wels_const.h into a shared header 2015-02-02 12:41:11 +02:00
Martin Storsjö
713beb7df7 Fix typos in header include guards 2015-02-02 12:41:11 +02:00
sijchen
5fdd01ec0c Merge pull request #1787 from mstorsjo/remove-stray-semicolon 
Remove accidental double semicolons
 2015-02-02 18:15:02 +08:00
sijchen
e7a7a35611 Merge pull request #1779 from mstorsjo/share-memalign 
Move the memory allocation/deallocation routines to the common library
 2015-02-02 18:14:55 +08:00
Martin Storsjö
a3063531c4 Remove accidental double semicolons 2015-02-02 09:20:35 +02:00
zhuiling
ee515e8b35 fix it 2015-02-02 10:17:19 +08:00
zhilwang
fb5c60b6fc Merge pull request #1782 from mstorsjo/avoid-yasm-warnings 
Move DEFAULT REL into the x86_64 cases
 2015-02-01 17:38:59 -08:00
zhilwang
7629543e3e Merge pull request #1784 from mstorsjo/force-arm-mode 
Force armv7/neon within the arm assembly header file
 2015-02-01 17:36:36 -08:00
Martin Storsjö
5a78735802 Force armv7/neon within the arm assembly header file 
This avoids having to add extra compiler flags to be able to build
them.
 2015-02-02 00:53:39 +02:00
Martin Storsjö
3243a78959 Move DEFAULT REL into the x86_64 cases 
This fixes warnings when building for x86_32 using yasm, which says
the "DEFAULT REL" is ignored for non-64-bit targets.
 2015-02-02 00:49:37 +02:00
Martin Storsjö
2356eb1ff6 Move the memory allocation/deallocation routines to the common library 
They are still used slightly differently in the encoder and decoder;
the decoder uses plain functions while the encoder uses one object
keeping track of the number of allocated bytes, and keeping track
of the requested alignment.
 2015-01-30 11:30:59 +02:00
Martin Storsjö
078d96bef7 Don't overwrite the git-tracked version.h 
When generating a new version of the header, that includes the
actual git hash, don't overwrite the file that is tracked by git.
Instead create a new file, and include this only if the build system
indicates that it exists (by setting a define). This allows the
untouched source tree to be built from within an IDE even if make
has not been run.

This reduces the hassle with a file that needs to be ignored in the
git configuration.

The downside is that the generated file isn't used if building
from within an IDE, if the header has been updated by calling make
before (since the IDE configuration doesn't know whether the user
actually has run make). Since users of the IDE might not build via
make in the command line at all (in the same source checkout at least),
this should not be an issue in practice. The previous way things worked,
the version hash (generated by make) when used in an IDE could actually
be outdated and misleading.
 2015-01-30 10:55:16 +02:00
Martin Storsjö
869870e670 Rename the decoder WelsMalloc to WelsMallocz 
This function actually zero-initializes the allocated memory, thus
make this clear in the function name.

This makes the function name match the same behaviour in the encoder.
 2015-01-30 10:37:48 +02:00
Sijia Chen
a4f09f4d76 update the version info to sync with recent change of API 2015-01-29 11:14:08 +08:00
dongzha
4f4adcec49 Merge pull request #1772 from mstorsjo/mc-unify 
Move the MC implementation to the common library
 2015-01-29 09:28:57 +08:00
Martin Storsjö
f03edd79c2 Merge all MC unit tests into one file 
Since both encoder and decoder use the same MC implementation now,
one set of tests is enough.
 2015-01-28 23:59:37 +02:00
Martin Storsjö
9a0663620a Move the MC routines to the common library 
Use the decoder versions of the functions (which are capable
of handling widths 4/8/16 for luma, not only 16 as in the
encoder). By using the more generic versions, there may be a small
performance loss since the functions need to check the width
in every call. Actual measurements show that the actual change is
very small (and the shared routines turn out to actually be faster
than the existing ones in ARM NEON setups).
 2015-01-28 15:32:46 +02:00
Martin Storsjö
1a7d0ab831 Hook up McChromaWidthEq8_ssse3 into use in the decoder as well 2015-01-28 15:20:43 +02:00
Martin Storsjö
5d67df4adb Remove leftover references to expand_pic in the welsdec xcode project 
These files are part of the common library (and the common xcode
project) now instead.
 2015-01-28 15:13:50 +02:00
Martin Storsjö
9a9fc4c489 Change pfSampleAveraging to be a single function with internal width handling 
This makes it match the behaviour of pMcLumaFunc and pMcChromaFunc.
 2015-01-28 13:59:58 +02:00
Martin Storsjö
8b6bf0499f Rename the encoder MC chroma/luma functions to match the naming in the decoder 2015-01-28 13:59:58 +02:00
Martin Storsjö
1cf68cad13 Move the qpel function table into the luma function itself 
This unifies the luma MC interface to match the decoder side.
 2015-01-28 13:59:58 +02:00
huili2
b6e28495dd modify decoder flowchart for non vcl AU check 2015-01-28 00:22:48 -08:00
Martin Storsjö
7354a8cbae Handle width==2 for the neon version of McCopy 
Fall back on the C version of the algorithm here, just as in
McCopy_sse2.

This fixes 2 of the McCopy tests on arm/aarch64.
 2015-01-28 09:01:00 +02:00
ruil2
c4c73ee5d6 Merge pull request #1764 from sijchen/fix_br3 
[Encoder] fix potential overflow and save unnecessary calculation
 2015-01-28 11:45:47 +08:00
ruil2
051a16e02e Merge pull request #1762 from sijchen/fix_sta 
[Encoder] fix statistics: updating should be independent with log interval
 2015-01-28 10:51:51 +08:00
Sijia Chen
5752888c81 1, fix a potential overflow 
2,    remove some unnecessary multiplier and add input check of spatial BR
 2015-01-28 10:51:27 +08:00
Martin Storsjö
23b20fb14c Simplify code in HorFilterInput16bit in MC 
This avoids a gcc optimizer bug (which seems to be present in some
gcc 4.6 and 4.7 versions) at the -O3 level.
 2015-01-27 09:41:00 +02:00
Martin Storsjö
279e14b34e Add const to some inline functions within MC 2015-01-27 09:41:00 +02:00
Martin Storsjö
478af8f00c Unify the pfChromaMc function signature with the decoder version 
Instead of passing a struct with x/y, pass them as two separate
parameters.
 2015-01-27 09:40:59 +02:00
Martin Storsjö
9fc9acfd14 Avoid a duplicate local typedef 2015-01-27 09:40:59 +02:00
Martin Storsjö
7d9fc35235 Unify the encoder MC init function with the decoder 
This simplifies the code a little, by passing a pointer to the sub-struct
SMcFunc instead of to the full SWelsFuncPtrList, which isn't necessary.
 2015-01-27 09:40:59 +02:00
ruil2
5b27407993 Merge pull request #1757 from sijchen/sps_list4 
[Encoder] add input checking and fix a bs length under new strategy
 2015-01-27 10:43:20 +08:00
ruil2
1b48e6c88c Merge pull request #1756 from sijchen/savc4 
[Encoder] Implementation of bSimulcastAVC and UT
 2015-01-27 10:42:50 +08:00
Sijia Chen
d557578be3 fix statistics: updating should be independent with log interval 2015-01-27 10:12:51 +08:00
Martin Storsjö
4ca8ad2612 Initialize pCodecInstance in the constructor 
This avoids using uninitialized memory if SetCodecInstance
isn't ever called (as in DecUT_ParseSyntax).
 2015-01-26 13:16:09 +02:00
Sijia Chen
deccd1eadb fix profileidc 2015-01-26 16:36:42 +08:00
Sijia Chen
e2fb48d404 change the trace to warning 2015-01-26 15:42:09 +08:00
Sijia Chen
8d5ec6759d add support of SimulcastAVC functions 2015-01-26 15:25:09 +08:00
Sijia Chen
f13624316e fix level-idc related input checking and a statistics calculation 
will change bs but tested there is little impact on RD
being reviewed at: https://rbcommons.com/s/OpenH264/r/1084/
 2015-01-26 15:20:58 +08:00
Sijia Chen
45e3424364 add input checking and fix a bs length under new strategy 2015-01-26 11:12:24 +08:00
sijchen
7f967f6fc4 Merge pull request #1750 from syureyi/crqp 
to support different chroma qp
 2015-01-23 13:30:21 +08:00
zhuiling
a45c661525 to support different chroma qp 
fix travis ci error
 2015-01-23 09:06:09 +08:00
dong zhang
6dcaf2f890 fix crash in SGE test 2015-01-22 09:10:28 +08:00
ruil2
e7fff10d91 Merge pull request #1745 from sijchen/fix_ut 
[UT] fix a random failure of UT
 2015-01-21 17:31:34 +08:00
Sijia Chen
d74234d975 enhance max-br checking logic 2015-01-21 16:54:43 +08:00
Sijia Chen
6664593ef7 Merge branch 'master' of https://github.com/cisco/openh264 into fix_max_br 2015-01-21 16:35:31 +08:00
Sijia Chen
5b10c1ff8d fix random failure of UT under randon seed 1421732719 2015-01-20 15:47:05 +08:00
dong zhang
bd7fabd892 modify decoder console log level to WELS_LOG_WARNING 
change the postion of setting EC method
 2015-01-20 13:18:33 +08:00
Sijia Chen
ee40ba6392 improve the para checking of max-br 2015-01-20 10:31:13 +08:00
sijchen
8bef6bc942 Merge pull request #1738 from huili2/init_EC_modify 
modify EC init position
 2015-01-19 16:36:29 +08:00
Sijia Chen
d9502aa71d rename the strategy to enum 2015-01-19 10:49:38 +08:00
huili2
50ce93e31c modify EC init position 2015-01-18 18:07:27 -08:00
Martin Storsjö
9e00bf89aa Use VerifyVersionInfo instead of GetVersionEx 
The GetVersionEx function has been deprecated since Windows 8.1.
VerifyVersionInfo exists at least since Windows 2000.

This fixes warnings about use of deprecated functions in MSVC.
 2015-01-16 13:38:21 +02:00
Martin Storsjö
b83b19254d Add a cast when converting from uint64_t to int 
This silences warnings with MSVC.
 2015-01-16 13:38:07 +02:00
ruil2
aea1017562 Merge pull request #1732 from sijchen/sps_list2 
[Encoder] add new SpsPpsStrategy and UT
 2015-01-16 10:53:45 +08:00
Martin Storsjö
d6b4b5365d Declare a float constant as float instead of as double 
This avoids warnings about converting from double to float
in MSVC.
 2015-01-15 12:38:42 +02:00
Martin Storsjö
75bbc286e1 Add explicit casts when converting from double to int32_t 
This fixes build warnings with MSVC.
 2015-01-15 12:38:39 +02:00
Martin Storsjö
c8972f1c32 Consistently use the right type for log prints 
These fields were originally int64_t. Even though long long is
the same size as int64_t in all setups, some compilers treat
it as a different type than long long (in 64 bit environments,
int64_t is a long, not a long long, even if they are the same
size).

%"PRId64" is used for printing int64_t, while %lld is used
for printing long long.

This fixes build warnings with gcc.
 2015-01-15 12:33:44 +02:00
Sijia Chen
98ed302990 add new SpsPpsStrategy and UT 2015-01-15 18:04:39 +08:00