Yet Another x264 Builder x264 builds by JEEB

revision 1703

Wee, had to check out how quickly L-SMASH people could decide which would be better to use, the repo or the diff. Went for the 26th diff in the end. Not like they should have much difference, but still~ Also, ffmpegsource should now officially support those header compressed Matroska files better.

Download 32bit -- Download 64bit

Unpatched 64bit

(ffmpeg: git-a426aeb, libswscale: git-12463fa, ffms2: revision 328)

Currently patched into the patched builds:

• mp4muxer.diff
• x264_film_grain_optimization_1703.patch
• x264_fade_compensation_1703.patch

revision 1698

And another one down the tracks~. Also, congrats for wipple getting his lavf mov-related fixes into x264 . Also switched to the original LZMA as LZMA2 really didn't help that much in reducing file size >_>.

Patched files updated at 16:25 GMT+2 2010/08/16 to use L-SMASH's git repo.

Download 32bit -- Download 64bit

Unpatched 64bit

(ffmpeg: git-1045d70, libswscale: git-809d77e, ffms2: revision 322, L-SMASH: git-2d99fcd)

Currently patched into the patched builds:

• 0001-Fix-compilation-to-support-L-SMASH.patch
• L-SMASH's git repo's files in /output + L-SMASH's updated mp4.c
• x264_fade_compensation_1629.patch
• x264_film_grain_optimization_1680.patch

revision 1683

Count me flabbergasted. That FGO patch seems to be completely fine, and the original culprit seems to have been the newest Haali splitter that I was using to load the source M2TS file. But there's a catch that made me think it's a problem on x264's side -- a fine huffy avi with the same segment failed at least two times during the testing period. But then again, virtualdub failed at encoding lossless from the .avs script afterwards...

Which lead me to install an older Haali's splitter (from before the December 2009 release). And voila, it worked. Now I'm hurfing and durfing and testing it otherwise. Will post these builds now, and comment on it as I get this stuff ready. As an addition, the Haali release before the end-of-May one works fine for me... only the newest one seems to bork on this system.

Edit: As far as I can see it, it is the newest Haali splitter that is the culprit. Testing with a lossless H.264 matroska file with ffms2 input works mighty fine at the moment. Sorry for the hurf and durf, but I was really getting weird stuff with the 1680 build + Haali/DSS2 when loading that avs via avs2yuv Also, re-making 32bit builds because it seems I installed the 32bit zlib in a way that makes stuff actually use the non-static library (I needed to re-build it because building ffmpegsource with MSVS would fail without doing that). Will edit this post when that is done.

Edit2: 32 & 64bit builds updated, 32bit doesn't need zlib1.dll any more.

Download 32bit -- Download 64bit

Unpatched 64bit

(ffmpeg: git-096a163, libswscale: git-f3a12db, ffms2: revision 320)
patched builds: git-77dea3d, git-f3a12db, 320

Currently patched into the patched builds:

• mp4muxer.diff
• x264_fade_compensation_1629.patch
• x264_film_grain_optimization_1680.patch

revision 1680

Building Aegisub and getting Avisynth 2.6 to build straight from the CVS with Visual Studio 2010 took most of my power until last morning, but now I've gotten some sleep (Well, hello thar sleeping until 1PM) so x264 updated it is. And the patched builds are in 7z archives now because it makes them roughly 3-4 times smaller. What's inside is still x264.exe as that's the default binary name. And, last but not least, congrats to Dark Shikari for getting that libswscale win64 fix in. Makes my life just that one little tiny bit easier.

Edit: uploaded some new patched builds with the up-to-date Film Grain Optimization patch. You can check if you have them by doing x264 --version and checking if you have 1680+2 or 1680+3 written there.

Edit2: From the 1680+3 builds I've started using LZMA2 in 7zip. If you have an older 7zip/WinRAR you might have problems with the extraction of the archive. Updating your archiving application is recommended.

Download 32bit -- Download 64bit

Unpatched 64bit

(ffmpeg: git-a984b10, libswscale: git-eb5482e, ffms2: revision 320)
for the patched 1680+3 builds: git-91df001

Currently patched into the patched builds:

• mp4muxer.diff
• x264_fade_compensation_1629.patch
• x264_film_grain_optimization_1680.patch

revision 1677

[07/17 0800 GMT] New builds up, should fix the resize crash bug on win64.

! A bug in win64 libswscale confirmed, 64bit build might crash in some cases.

Filtering and various fixes. Do enjoy, and tell me if something breaks. Special thanks to Ramiro for providing a box to build stuff on, real kindness from one of the troll project developers. Also, I kind of meant that "help" on the last post as "work / a job" more than "giev me money". I just got lazy and never edited it to be like that.

Download 32bit -- Download 64bit

Unpatched 64bit

(ffmpeg: git-b7ee51b, libswscale: git-f8f339a, ffms2: revision 318)

Currently patched into the patched builds:

• mp4muxer.diff (VFR Maniac's and felidlabo's own "little" GPAC replacement mp4 muxer.)
• x264_fade_compensation_1629.patch

revision 1649

Ohi. Tenshi came over to say hi. Otherwise mbtree optimizations and some regression fixes.

Download 32bit -- Download 64bit

Unpatched 64bit

(ffmpeg: git-b5bfc9b, libswscale: git-0eac904, ffms2: revision 312)

revision 1643

Some other bugfixes and some speedups on the faster presets -- as well as the welcoming of a new stable branch (which won't get used here, but might get used by some people).

Download 32bit -- Download 64bit

Unpatched 64bit

(ffmpeg: git-dac8369, libswscale: git-0eac904, ffms2: revision 312)