cea0c5d | Ashok Kumar Mishra | 21 February 2018, 08:55:56 UTC | release: Release notes for v2.7 | 21 February 2018, 08:55:56 UTC |
aa121f7 | Ashok Kumar Mishra | 20 February 2018, 10:32:32 UTC | Merge with default; prep for v2.7 | 20 February 2018, 10:32:32 UTC |
11f839e | Mythreyi P | 15 February 2018, 10:21:26 UTC | x86: Split ipfilter16 kernals part3 Port vertical 4tap kernals from ipfilter16.asm to new source file, v4-ipfilter16.asm to reduce compile time of x265. | 15 February 2018, 10:21:26 UTC |
11aab4c | Mythreyi P | 13 February 2018, 08:28:25 UTC | x86: Split ipfilter16 kernals part2 Port horizonal 4tap kernals from h-ipfilter16.asm to a new source file, h4-ipfilter16.asm to improve build time. | 13 February 2018, 08:28:25 UTC |
d5404ef | Mythreyi P | 15 February 2018, 08:56:16 UTC | x86: Split ipfilter16 kernals into two different source files. Port horizonal kernals from ipfilter16.asm to a new source file, h-ipfilter16.asm to improve build time. | 15 February 2018, 08:56:16 UTC |
058f4d4 | Radhakrishnan | 31 January 2018, 09:43:30 UTC | x86: Split ipfilter8 kernals part3 This patch adds infra and moves the vertical 4tap ipfilter kernals from ipfilter8.asm to new source file v4-ipfilter8.asm to reduce compile time of x265. | 31 January 2018, 09:43:30 UTC |
9675729 | Aruna Matheswaran | 27 January 2018, 13:15:21 UTC | rc: fix inconsistency in pass 2 when weightp and cutree are enabled. | 27 January 2018, 13:15:21 UTC |
aa9af2a | Radhakrishnan | 24 January 2018, 10:48:24 UTC | x86: Split ipfilter8 kernals into two different source files, part2 This patch port the horizontal 4tap and 8tap kernals from ipfilter8.asm to a different source file h-ipfilter8.asm. It improves the build time by 20%. | 24 January 2018, 10:48:24 UTC |
62f96c6 | Radhakrishnan | 20 February 2018, 09:29:18 UTC | x86: Remove header file for horizontal ipfilters Since all ipfilter kernals share the ipfilter8.h file, a separate header file isn't needed for horizontal ipfilter kernals. So this patch removes the duplicate header file | 20 February 2018, 09:29:18 UTC |
b74734e | Praveen Tiwari | 19 January 2018, 06:34:23 UTC | x86: split ipfilter8 kernels into two different source file This patch implements infrastructure to split ipfiletr8 asm source file into two different files in order to avoid longer build time. It moves interp_8tap_horizontal kernels to the newly created file. | 19 January 2018, 06:34:23 UTC |
c0f65e1 | Bhavna Hariharan | 16 February 2018, 06:10:59 UTC | dhdr: Skip comments in the json file while parsing it. Comments of the type /* */ OR // are ignored. | 16 February 2018, 06:10:59 UTC |
59a3fab | Bhavna Hariharan | 16 February 2018, 06:07:45 UTC | dhdr: comply to HDR10+ LLC spec This patch does the following - 1) Consider Json files without Bezier Curve data as valid and inject them correctly without errors. 2) The numWindows values is now taken from the numWindows JSON value, instead of taking it from 1+ellipsesNum. Add numWindows to the dictionary. 3) Changed hardcoded applicationVersion value to 1 due to LLC requirement. | 16 February 2018, 06:07:45 UTC |
0fee5dc | Bhavna Hariharan | 14 February 2018, 10:15:29 UTC | limitTU: Save intra CU's TU depth when analysis save/load is enabled This patch will cause a output mismatch between analysis save and load when limit-tu 3/4 is enabled. This change is expected as the load run will have only the best mode's TU information. For CUs where the neighbour's TU depth is unavailable, load encode will evaluate all TU detphs. | 14 February 2018, 10:15:29 UTC |
6c3f4a2 | Ma0 | 10 February 2018, 05:16:45 UTC | fix output to pipe on Windows | 10 February 2018, 05:16:45 UTC |
6d467f9 | David Christenson | 03 February 2018, 05:18:18 UTC | CMake: fix generation of version info from .hg_archival.txt When building from a commit more recent than the latest tag, version.cmake wouldn't set X265_LATEST_TAG and X265_TAG_DISTANCE resulting in their defaults (0.0 and 0) being used. Also truncate the revision ID to 12 characters to match the length used when parsing from the hg executable. | 03 February 2018, 05:18:18 UTC |
3ba4c6a | Bhavna Hariharan | 31 January 2018, 14:14:19 UTC | csv: Bypass luma calculatations when --max-cll is OFF or when csv-log-level < 2 The maxFall and maxCll values are calculated from the min, max and average luma values. The luma values were being calculated even when --max-cll is disabled. This patch bypasses the luma calculations when --max-cll is OFF or when csv-log-level < 2 | 31 January 2018, 14:14:19 UTC |
0107f23 | Aruna Matheswaran | 31 January 2018, 13:07:13 UTC | slicetype: fix hanging issue due to dropping of BREF frames While forcing slicetypes through qp file, few BREF frames were not copied into lookahead's output queue due to incorrect check condition. | 31 January 2018, 13:07:13 UTC |
943c070 | Ricardo Constantino | 16 January 2018, 17:39:53 UTC | CMake: blacklist mingw implicit link libraries These also aren't meant to be in pkg-config's Libs.Private. | 16 January 2018, 17:39:53 UTC |
4a6bc8f | Ashok Kumar Mishra | 30 January 2018, 10:27:08 UTC | Merge with stable | 30 January 2018, 10:27:08 UTC |
abdc2d1 | Santhoshini Sekar | 24 January 2018, 07:41:11 UTC | fix warning in getRefFrameList | 24 January 2018, 07:41:11 UTC |
61245cc | Ashok Kumar Mishra | 19 January 2018, 09:16:54 UTC | fix for unable to open input file for gcc 32 bit compiler | 19 January 2018, 09:16:54 UTC |
a4d520d | Ma0 | 13 January 2018, 01:47:30 UTC | input: change from ifstream to stdio stream | 13 January 2018, 01:47:30 UTC |
ad3140d | Ashok Kumar Mishra | 16 January 2018, 13:55:26 UTC | Change type of bufSize variable to size_t to avoid warning. | 16 January 2018, 13:55:26 UTC |
e8cbdbc | Ma0 | 12 January 2018, 23:37:31 UTC | change date to 2018 in output file header | 12 January 2018, 23:37:31 UTC |
5be095e | Ma0 | 12 January 2018, 23:25:44 UTC | param2string: increase buffer size, do not store file names | 12 January 2018, 23:25:44 UTC |
36fdc05 | Vignesh Vijayakumar | 04 January 2018, 07:07:01 UTC | Fix Mac OS build warnings 1. due to SEA motion search 2. due to condition check for CTUInfo | 04 January 2018, 07:07:01 UTC |
6cead48 | Aarthi Thirumalai | 04 January 2018, 09:46:06 UTC | api: change signature of x265_csvlog_encode() replace x265_encoder* with x265_param* and int padx, int pady as function arguments. | 04 January 2018, 09:46:06 UTC |
9414c71 | Pradeep Ramachandran | 04 January 2018, 08:26:07 UTC | doc: Fix invalid UTF-8 characters in docs to enable linux gen | 04 January 2018, 08:26:07 UTC |
8612237 | Pradeep Ramachandran | 02 January 2018, 11:11:36 UTC | Merge with stable | 02 January 2018, 11:11:36 UTC |
730d621 | Pradeep Ramachandran | 02 January 2018, 04:26:04 UTC | Merge with stable | 02 January 2018, 04:26:04 UTC |
f9b6ce1 | Aasaipriya Chandran | 27 December 2017, 13:35:56 UTC | Reconfigure csvfpt if encoder param is null | 27 December 2017, 13:35:56 UTC |
6080fd6 | Aruna Matheswaran | 22 December 2017, 12:53:24 UTC | Restore macros defining analysis options | 22 December 2017, 12:53:24 UTC |
579e2ac | Aruna Matheswaran | 22 December 2017, 09:20:08 UTC | Add support for RADL pictures | 22 December 2017, 09:20:08 UTC |
0489642 | Ashok Kumar Mishra | 22 December 2017, 09:44:57 UTC | Merge with stable | 22 December 2017, 09:44:57 UTC |
08a4b22 | Santhoshini Sekar | 22 December 2017, 07:18:37 UTC | Fix possible NULL pointer dereferencing in cudata init | 22 December 2017, 07:18:37 UTC |
6ba2fe7 | Aruna Matheswaran | 21 December 2017, 09:56:55 UTC | analysis: Enable analysis save and load in a single run. This patch does the following, 1. Deprecates the cli option --analysis-reuse-mode and introduces two new cli options --analysis-save <filename> and --analysis-load <filename>. 2. Deprecates the param analysisReuseMode and introduces two new param options analysisSave and analysisLoad. | 21 December 2017, 09:56:55 UTC |
744a6bd | Divya Manivannan | 14 December 2017, 11:23:44 UTC | analysis: avoid copying lowres vbv data when vbv is disabled | 14 December 2017, 11:23:44 UTC |
b7115c1 | Santhoshini Sekar | 21 December 2017, 03:39:46 UTC | warn out saying that limitTU=3 or 4 with AVCINFO produces inconsistent output. Set mincusize to 8 when it is not 8 for MVtype=AVCINFO as AVCINFO expects so | 21 December 2017, 03:39:46 UTC |
5b9a0ad | Santhoshini Sekar | 19 December 2017, 10:03:30 UTC | fix bugs in analysis-reuse-level=7 and refine-mv-type=AVC | 19 December 2017, 10:03:30 UTC |
706d82c | Divya Manivannan | 14 December 2017, 10:27:01 UTC | analysis: avoid creating analysis file when useanalysisFile option is disabled | 14 December 2017, 10:27:01 UTC |
28351e1 | Santhoshini Sekar | 12 December 2017, 12:34:04 UTC | modify api x265_get_ref_frame_list to provide POC lists for L0 and L1 references | 12 December 2017, 12:34:04 UTC |
e1e294a | Ma0 | 07 December 2017, 08:19:11 UTC | remove unnecessary sqrt from MVTHRESHOLD checking Thanks MonoS Diablo for pointing this. | 07 December 2017, 08:19:11 UTC |
7efcffd | Aruna Matheswaran | 07 December 2017, 05:29:12 UTC | rc: Fix inconsistency in --const-vbv (issue #381) VBV intializations during ABR-reset is removed. Inconsistent rowTotalBits in rc-update with --const-vbv is fixed. | 07 December 2017, 05:29:12 UTC |
1b57db6 | Ashok Kumar Mishra | 06 December 2017, 08:03:00 UTC | Merge with stable | 06 December 2017, 08:03:00 UTC |
9b61205 | Santhoshini Sekar | 04 December 2017, 10:43:45 UTC | remove unnecessary event m_reconEncoded from Frame class | 04 December 2017, 10:43:45 UTC |
556810f | Santhoshini Sekar | 04 December 2017, 10:38:19 UTC | wait until the last row of recon is complete. Waiting on m_reconEncoded doesn't ensure full recon generation. | 04 December 2017, 10:38:19 UTC |
c30643f | Aruna Matheswaran | 04 December 2017, 09:30:04 UTC | test: Fix gop-lookahead cli error in regression test | 04 December 2017, 09:30:04 UTC |
e85149c | Ma0 | 30 November 2017, 17:29:19 UTC | cli: add new option '--fullhelp' | 30 November 2017, 17:29:19 UTC |
5bf15bd | Ashok Kumar Mishra | 04 December 2017, 07:58:10 UTC | Merge with stable | 04 December 2017, 07:58:10 UTC |
a300cb9 | Mahesh Pittala | 30 November 2017, 13:39:45 UTC | Install symbol files in MinGW In debug and RelWithDebInfo configuration, .pdb files are installed. | 30 November 2017, 13:39:45 UTC |
ebeebed | Santhoshini Sekar | 30 November 2017, 12:06:32 UTC | deblock: set reference frame to NULL if refIdx < 0. Dereferencing refFrameList when refIdx < 0 is not valid | 30 November 2017, 12:06:32 UTC |
8f96e0e | Pradeep Ramachandran | 29 November 2017, 03:22:18 UTC | Added tag 2.6 for changeset 0e9ea76945c8 | 29 November 2017, 03:22:18 UTC |
9b774c2 | Pradeep Ramachandran | 29 November 2017, 03:21:40 UTC | release: Release notes for v2.6 | 29 November 2017, 03:21:40 UTC |
468d25b | Praveen Tiwari | 22 November 2017, 08:11:20 UTC | encoder.cpp: fix encoder crash for --analysis-reuse-level=10 | 22 November 2017, 08:11:20 UTC |
4fe185f | Aruna Matheswaran | 21 November 2017, 05:18:24 UTC | Extend gop boundary by doing gop lookahead | 21 November 2017, 05:18:24 UTC |
24abfd9 | Ma0 | 30 November 2017, 17:16:40 UTC | fix ambiguous call to overloaded function 'sqrt' (in old MSVC) | 30 November 2017, 17:16:40 UTC |
84e38f4 | Pradeep Ramachandran | 30 November 2017, 04:36:49 UTC | doc: Update build/readme.txt for detailed instructions with nasm | 30 November 2017, 04:36:49 UTC |
8e0cb19 | Pradeep Ramachandran | 21 November 2017, 04:20:45 UTC | Merge with default; prep for v2.6 | 21 November 2017, 04:20:45 UTC |
83d010f | Pradeep Ramachandran | 21 November 2017, 04:10:16 UTC | warn: Fix warning from analysis.cpp | 21 November 2017, 04:10:16 UTC |
1b62499 | Ma0 | 20 November 2017, 09:01:22 UTC | doc: Update an outdated help message for --bframes | 20 November 2017, 09:01:22 UTC |
bbc294e | Praveen Tiwari | 17 November 2017, 13:53:14 UTC | analysis: use AVC CU analysis-info for HEVC mode analysis This patch work implements the functionality for anlysis-reuselevel 7, here we want to use AVC analysis-info for HEVC mode decision and use the depth from offload for AVC sizes | 17 November 2017, 13:53:14 UTC |
ce84ea5 | Mythreyi P | 17 November 2017, 09:26:12 UTC | Fix install symbol files .pdb files are installed only in Windows. | 17 November 2017, 09:26:12 UTC |
c42434c | Mythreyi P | 20 November 2017, 04:21:55 UTC | Fix Build fails for MacOS Duplicate uninstall targets were removed. | 20 November 2017, 04:21:55 UTC |
9cabe8d | Bhavna Hariharan | 17 November 2017, 06:56:48 UTC | analysis: evaluate CUs when inter refinement is enabled and analysis data from analysis save mode is unavailable. Earlier, smaller CU sizes were evaluated only when the current block is decided as "skip" by the save mode. This patch changes the behaviour of inter refinement by evaluating smaller CUs irrepective of the mode of the current block. | 17 November 2017, 06:56:48 UTC |
e89042d | Praveen Tiwari | 17 November 2017, 08:46:31 UTC | analysis: use AVC full CU analysis-info for analysis-reuselevel 8 This patch work implements the functionality for anlysis-reuselevel 8, here we want to incorpated AVC analysis data from offload in our enocde, HEVC speific sizes (32x32 and 64x64) CU analysis still will be caluated as per x265 CLI option provided. | 17 November 2017, 08:46:31 UTC |
82293d8 | Praveen Tiwari | 13 November 2017, 08:24:43 UTC | analysis: update analysisReuseLevel 7 for analysis sharing | 13 November 2017, 08:24:43 UTC |
d9248ad | Pradeep Ramachandran | 16 November 2017, 14:53:14 UTC | doc: Cleanup api.rst, x265cli.h, update comments in x265.h | 16 November 2017, 14:53:14 UTC |
3b0d0fb | Praveen Tiwari | 13 November 2017, 06:50:50 UTC | API: 'x265_set_analysis_data' to capture analysis information | 13 November 2017, 06:50:50 UTC |
ab572d4 | Pradeep Ramachandran | 16 November 2017, 04:55:50 UTC | param: Disable opt-qp-pps and opt-ref-list-length-pps by default The bits saved by these options is very miniscule, and not observable in most tests focusing on low bitrates, and low resolutions; this was the goal of these two options. Also, since these options change the PPS, the muxer has to ensure that the headers are sent every frame which seems to trouble some muxers and users. | 16 November 2017, 04:55:50 UTC |
e6dd38c | Kavitha Sampath | 14 November 2017, 05:30:09 UTC | Add CLI option to enable or disable picture copy to internal frame buffer | 14 November 2017, 05:30:09 UTC |
7ed3f81 | Ashok Kumar Mishra | 10 November 2017, 13:32:48 UTC | Encoder: avoid blocking in getEncodedPicture() for zero latency when multiple slices are enabled Improved performance by 10 - 15 fps depending upon preset and number of slices. | 10 November 2017, 13:32:48 UTC |
ff7eb83 | Mythreyi P | 13 November 2017, 05:17:48 UTC | Add support for uninstall in windows Files installed in the INSTALL target are uninstalled. | 13 November 2017, 05:17:48 UTC |
753903d | Mythreyi P | 10 November 2017, 06:16:13 UTC | Install symbol files In debug and RelWithDebInfo configuration, .pdb files are installed. | 10 November 2017, 06:16:13 UTC |
e501131 | indumathi@multicorewareinc.com | 09 November 2017, 08:20:03 UTC | Fix Visual Studio version check | 09 November 2017, 08:20:03 UTC |
3eaa1cd | Ma0 | 14 November 2017, 04:48:56 UTC | x265.h: change references into pointers for C compatibility Commit 563cbe1 'api: move csv and dither functions into api structure' inserts to x265.h C++ only specific code (references) that broke compatibility with C compilers. This patch fixes this by changing code style only. | 14 November 2017, 04:48:56 UTC |
ea72994 | Divya Manivannan | 08 November 2017, 11:38:18 UTC | api: move csv and dither functions into api structure csv is now moved into libx265. So, we can remove x265-extras.h | 08 November 2017, 11:38:18 UTC |
63d87cb | Pradeep Ramachandran | 08 November 2017, 10:48:29 UTC | test: Fix cli error in regression test | 08 November 2017, 10:48:29 UTC |
2800d1f | Santhoshini Sekar | 08 November 2017, 10:34:27 UTC | add help information for refine-mv-type | 08 November 2017, 10:34:27 UTC |
c87e5ec | hribeiro | 08 November 2017, 05:28:47 UTC | Disable QP limitations for lowpass subband dct | 08 November 2017, 05:28:47 UTC |
86a95b2 | Santhoshini Sekar | 07 November 2017, 07:13:23 UTC | rename function pointer names for x265_get_slicetype_poc_and_scenecut() and x265_get_ref_frame_list() in x265_api | 07 November 2017, 07:13:23 UTC |
305e240 | Santhoshini Sekar | 07 November 2017, 07:02:07 UTC | export API functions x265_get_slicetype_poc_and_scenecut() and x265_get_ref_frame_list() to x265.def.in | 07 November 2017, 07:02:07 UTC |
0ddfa15 | Vignesh Vijayakumar | 02 November 2017, 04:10:41 UTC | x86: Change assembler from YASM to NASM Supports NASM versions 2.13 and greater | 02 November 2017, 04:10:41 UTC |
5a81c7b | Vignesh Vijayakumar | 02 November 2017, 04:09:58 UTC | x86: Modify asm codes for NASM compatibility | 02 November 2017, 04:09:58 UTC |
00c60f4 | Pradeep Ramachandran | 29 November 2017, 03:25:44 UTC | Merge with stable | 29 November 2017, 03:25:44 UTC |
75fdd73 | Ma0 | 24 November 2017, 16:23:59 UTC | nits: remove unnecessary parentheses | 24 November 2017, 16:23:59 UTC |
1362fae | Pradeep Ramachandran | 22 November 2017, 16:30:48 UTC | Merge with stable | 22 November 2017, 16:30:48 UTC |
81c899f | Pradeep Ramachandran | 05 September 2017, 05:51:24 UTC | lic: Update license info in source files of HDR10+ | 05 September 2017, 05:51:24 UTC |
c2c6924 | Santhoshini Sekar | 09 August 2017, 10:43:19 UTC | Do not wait until all the frames in a stream are flushed, return to API as soon as one frame is flushed. In few cases it is better to flush only one frame in x265-lib and return to API. The application should flush the remaining frames in the stream in these situations. This is especially useful when the application wants to retrieve information from the lib. | 09 August 2017, 10:43:19 UTC |
d80ca30 | Santhoshini Sekar | 09 August 2017, 10:32:20 UTC | wait for the event of copying MV data if refine-mv-type is enabled | 09 August 2017, 10:32:20 UTC |
b5536d8 | Santhoshini Sekar | 06 November 2017, 08:22:22 UTC | add new CLI refine-mv-type | 06 November 2017, 08:22:22 UTC |
8d5226e | Santhoshini Sekar | 09 August 2017, 06:08:58 UTC | add function to copy analysis data after z scan The pointer to x265_analysis_data sent as parameter to this function has data for 16x16 blocks stored continuously. To store this data in bigger CTU sizes(32x32 or 64x64) we need to perform z-scan of constituent 16x16 blocks before copying. | 09 August 2017, 06:08:58 UTC |
1bab351 | Aruna Matheswaran | 07 November 2017, 13:45:31 UTC | vbv-end: nits | 07 November 2017, 13:45:31 UTC |
0de0134 | Aruna Matheswaran | 06 November 2017, 06:02:04 UTC | Add vbv-end to denote VBV emptiness after inserting all the frames into it. | 06 November 2017, 06:02:04 UTC |
110b8c5 | Pradeep Ramachandran | 06 November 2017, 04:13:36 UTC | doc/test: Add lowpass test, remove/add experimental labels to some features. | 06 November 2017, 04:13:36 UTC |
555e14a | hribeiro | 06 November 2017, 04:24:56 UTC | Implementation of low-pass subband dct approximation. | 06 November 2017, 04:24:56 UTC |
8de7a2c | Praveen Tiwari | 31 October 2017, 10:43:49 UTC | api: 'x265_get_ref_frame_list' to get forward and backward refrence list | 31 October 2017, 10:43:49 UTC |
0493efd | Praveen Tiwari | 31 October 2017, 08:27:37 UTC | api: 'x265_get_slicetype_poc_and_scenecut' to fetch slicetype, poc and scenecut information | 31 October 2017, 08:27:37 UTC |
a30f6e2 | Mahesh Pittala | 02 November 2017, 06:47:29 UTC | cross compile visual studio builds from msys bash shell x265 already supports native visual studio builds but if some one wants to integrate visual studio builds to the ffmpeg, they can generate from MinGw. Here, it sets all required environment variables. Build procedure: For 32 bit - launch msys bash shell from 'visual studio command prompt' and run make-Makefiles.sh For 64 bit - launch 'visual studio command prompt' and run '..\vcvarsall.bat amd64 | x86_amd64 | x86_arm | amd64_x86' and then launch msys bash shell and run ./make-Makefiles-64bit.sh | 02 November 2017, 06:47:29 UTC |
8fc1e3f | indumathi@multicorewareinc.com | 30 October 2017, 09:18:45 UTC | Add Visual Studio 2017 build folders | 30 October 2017, 09:18:45 UTC |
35ffeb6 | Aruna Matheswaran | 09 October 2017, 12:12:44 UTC | lowres: re-init plannedType | 09 October 2017, 12:12:44 UTC |