Current Patches, Where to get them, How they affect speed/output

[komisar]

x264 GIT:969 source
http://komisar.gin.by/test/x264_969_src.7z

[stanjr]

yes, going to snv would be a lot better.. at least for linux users, since no one are building latest snapshot+patch for *nix :(

BHHI could build this for you for 64bit Ubuntu Linux if you want.

[buzzqw]

i actually can compile git versione without problem (./configure --enable-pthread --enable-mp4-output)
but.. i miss the patches

anyway.. on last git b-adapt is committed, just missing the psy-rd and hrd_pulldown

and.. btw i use 32bit debian sid

thanks a lot for your offer stanjr!

BHH

[kemuri-_9]

x264 GIT:969 source
http://komisar.gin.by/test/x264_969_src.7z

thanks for providing!
since it's CRLF, makes things more of a pain though

and stanjr should still be able to compile for x86 linux on his x64 one, by using
(iirc) --host=i686-pc-linux-gnu
which may or may not need --extra-cflags=-m32 as well

going to need the gcc multilib packages to do that tho.

[ImmortAlex]

Hey, GIT is on again and psy-rd is commited! Waiting for fresh builds...

[martino]

yes, going to snv would be a lot better.. at least for linux users, since no one are building latest snapshot+patch for *nix :(

BHH
Actually, I started after switching to Ubuntu.

[buzzqw]

i am compiling too (for anyone speaking italian http://forum.divxmania.it/x264-linux-build-14092008/t11427/n.html )

@ImmortAlex
git is out for me... are you sure you can git clone git://git.videolan.org/x264.git ?

BHH

[Ranguvar]

Git is back online :)

[Comatose]

Dead for me ><

[LoRd_MuldeR]

Doesn't work for me either, I still get:

LoRd_MuldeR@MULDER_NEU /c/downloads/x264_git
$ ./build.sh
Initialize x264/.git
Initialized empty Git repository in c:/downloads/x264_git/x264/.git/
git.videolan.org[0: 91.121.111.144]: errno=No such file or directory
fatal: unable to connect a socket (No such file or directory)

[burfadel]

GIT back online, seems nice and fast now unlike a couple of hours ago. The downside is that bobor's site x264.nl seems to have missed the last two git commits, which is now Psy-rd (on by default) and psy-trellis (optional) :) and some hadamard stuff. Its just unfortunate GIT when down when these goodies were committed!

[LoRd_MuldeR]

GIT back online, seems nice and fast now unlike a couple of hours ago. The downside is that bobor's site x264.nl seems to have missed the last two git commits, which is now Psy-rd (on by default) and psy-trellis (optional) :) and some hadamard stuff. Its just unfortunate GIT when down when these goodies were committed!

No, it's not back online. At least I cannot check out. Or something has changed and I need to adjust my command? :confused:

git clone git://git.videolan.org/x264.git

[komisar]

Git on web online... But "git clone" not work. Alternatively, you can take snapshot from http://git.videolan.org/gitweb.cgi?p=x264.git;a=shortlog

[kemuri-_9]

yeah but the snapshots also don't have any of the git repository information within them,
so the ./configure in x264 will go :mad: from not having the information to write in the version build number

[komisar]

Of course. The option for "impatient". :-)
As we have seen already 971 available

[Inventive Software]

This is the build I've been waiting for. I'll wait for x264.nl to chuck a build online before I start transcoding. :)

[kemuri-_9]

Of course. The option for "impatient". :-)
As we have seen already 971 available

yeah psy-rd is included for 970 and hardamard added to psy-rd in 971

looking at the raw diffs, seems that psy-rd will be on by default and psy-trellis is off by default,
which is actually better for me since i've seen psy-trellis be hit and miss on my anime encodes.

-Edit-
meh, i'll go ahead and get a build in and manually put in the version definitions

-Edit2-
Git seems to be working again

[LoRd_MuldeR]

looking at the raw diffs, seems that psy-rd will be on by default and psy-trellis is off by default.

How do you enable Psy Trellis now? Simply with "--trellis 1" respectively "--trellis 2" or is there a new switch?

Git seems to be working again

Confirmed :)

[kemuri-_9]

How do you enable Psy Trellis now? Simply with "--trellis 1" respectively "--trellis 2" or is there a new switch?
Confirmed :)

it's the same switch, just do --psy-rd 1.0 or 1.0:1.0 to enable both


--psy-rd Strength of psychovisual optimization ["1.0:0.0"]
#1: RDO (requires subme>=6)
#2: Trellis (requires trellis, experimental)

[LoRd_MuldeR]

it's the same switch, just do --psy-rd 1.0 or 1.0:1.0 to enable both

Oh, I see :o

[elguaxo]

no more patches? http://img89.imageshack.us/img89/7123/mspacman2thwb6.gif

[kemuri-_9]

no more patches? http://img89.imageshack.us/img89/7123/mspacman2thwb6.gif

the hrd pulldown is the only left as the mostly used ones go

and the devs already stated they won't commit it until it has some factual basis for working (iirc)

[stanjr]

no more patches? http://img89.imageshack.us/img89/7123/mspacman2thwb6.gifMaybe this one will still be useful for those of us encoding on Linux:Second Pass ETA Patch:
Current: http://www.benswebs.com/public/x264/patches/x264_fp-eta.01.r680.diff
Creator/Maintainer: morph166955
Description: Forces x264 to use the frame count from the stats file on a second pass if the frame count can't be calculated for some reason (such as the use of a fifo pipe).

[kemuri-_9]

Maybe this one will still be useful for those of us encoding on Linux:

that'll be useful only in the cases where piping occurs, which is mostly the current standard practice for the linux side
(since they for some reason don't want to save .y4m/.yuv files and encode from them :p)

but it also makes me wonder how avisynth 3.0 has been designed to handle passing off to linux based encoders (since it is platformless)....

[stanjr]

How in the heck is that patch supposed to be applied? I get this when trying to apply it:stanjr@stanjr-desktop:~/Programs/Source/mplayer-x264/x264$ patch -p1 -i ../x264diffs/x264_fp-eta.01.r680.diff
patch: **** Only garbage was found in the patch input.

[Ranguvar]

You probably need to convert it to the Unix line-end convention. Try dos2unix.

[kemuri-_9]

patch -p 0 -i ../x264patches/x264_fp-eta.01.r680.diff
patching file encoder/ratecontrol.c
Hunk #1 succeeded at 409 (offset 57 lines).
patching file x264.c
Hunk #1 succeeded at 799 (offset 36 lines).
patching file x264.h
Hunk #1 succeeded at 156 with fuzz 2 (offset 9 lines).


my text editor already recognizes it as LF only (unix style)

[skystrife]

x264.973.modified.exe (http://www.mediafire.com/?dmrqpxqhmag) - Alternate Download (http://skystrife.com/x264/x264.973.modified.exe)

Patches used:

x264_hrd_pulldown.09_interlace.diff

gcc 3.4.5 fprofiled build with -march=pentium2.

psyrd and b-adapt 2 patches have been committed!

[gav1577]

x264.973.modified.exe (http://www.mediafire.com/?dmrqpxqhmag) - Alternate Download (http://skystrife.com/x264/x264.973.modified.exe)

Patches used:

x264_hrd_pulldown.09_interlace.diff

gcc 3.4.5 fprofiled build with -march=pentium2.

psyrd and b-adapt 2 patches have been committed!

Nice 1 thanks

[Romario]

Please, skystrife, can you compile 64-bit build of your patched versions of x264 ?

It's about time, it's shame not to have 64-bit build in year 2008.

[LoRd_MuldeR]

Please, skystrife, can you compile 64-bit build of your patched versions of x264 ?

x264 doesn't support x64 Windows. I think it could be built for x64 Windows by disabling all the Assembly, but that wouldn't make sense, speed-wise...

[kemuri-_9]

x264 doesn't support x64 Windows. I think it could be built for x64 Windows by disabling all the Assembly, but that wouldn't make sense, speed-wise...

yes i have made x64 builds but the assembly does not work with it (seg faults right at the start), and when disabling the assembly it's slower than the x86 with assembly.

[LoRd_MuldeR]

yes i have made x64 builds but the assembly does not work with it (seg faults right at the start), and when disabling the assembly it's slower than the x86 with assembly.

That's because Win-x64 uses a different calling convention than Linux-x64 and hence the x264 assembly won't work on Win-x64 (but on Linux-x64 it does).
It's a known problem, but currently none of the developers is willing to fix it...

[akupenguin]

None of the developers has win64 nor is willing to install it. But you can help by donating a shell account on a win64/cygwin box.

[LoRd_MuldeR]

None of the developers has win64 nor is willing to install it. But you can help by donating a shell account on a win64/cygwin box.

I'm running WinXP x64 Edition. If you tell me what to do, I'll do what I can...

[Ranguvar]

Same as above.

[akupenguin]

I don't actually know how, since cygwin isn't designed to be a whole multiuser environment. But there is a sshd, maybe that's enough.

[kemuri-_9]

None of the developers has win64 nor is willing to install it. But you can help by donating a shell account on a win64/cygwin box.

win64, msys + mingw 64, and RDP?
does that sound sufficient, or would you still rather prefer sshd?

[akupenguin]

win64, msys + mingw 64, and RDP?
does that sound sufficient, or would you still rather prefer sshd?
Sufficient if it works. The only times I've used RDP was RDP-over-X-over-compressed-ssh :)

[Manao]

sshd works well with cygwin, and is surprisingly easy to set up.

[Shinigami-Sama]

Sufficient if it works. The only times I've used RDP was RDP-over-X-over-compressed-ssh :)

rdesktop works wonderfully well on solaris
so it should work on any other *nixy OS

[Ranguvar]

Check the Git log! Awesome new changes!

Home (http://sites.google.com/site/ranguvar13/x264-builds)
Direct download (http://sites.google.com/site/ranguvar13/x264-builds/rang_x264_r0977.7z?attredirects=0), Mirrors (http://www.rapidspread.com/file.jsp?id=abz7nxcc8z)

x264 r977 from Git (patched, fprofiled).
Thanks to the x264 devs, including those who made the patches I use.
Compiled by Ranguvar on September 16th, 2008, with GCC 4.3.2.
DON'T think that because I used march=athlon it restricts the CPUs you can use.
It seems to improve performance (VERY slightly) for all CPUs.

Open this archive with the free, multi-platform tools 7-Zip or p7zip. Compressed with LZMA.
The src folder contains the patched source code.
The bin folder contains a binary executable, and a DLL for those apps that use it
(May not work in AviDemux. Get those DLLs from LoRd_MuldeR).

Git: git://git.videolan.org/x264.git
Info, and source tarballs: http://www.videolan.org/developers/x264.html
Changelog: http://git.videolan.org/gitweb.cgi?p=x264.git
Vanilla builds: http://x264.nl/
Discussion: http://forum.doom9.org/forumdisplay.php?f=77
http://forum.doom9.org/showthread.php?t=130364


Applied patches, in the order applied (included, unchanged, in the patches folder):

patch -p1 -i ../x264diffs/rang_x264_version.diff
patch -p1 -i ../x264diffs/x264_dll_alignment_fix.01.diff
patch -p1 -i ../x264diffs/x264_hrd_pulldown.09_interlace.diff
patch -p0 -i ../x264diffs/x264_fp-eta.01.r680.diff


CLI used for build: ./configure --enable-shared --extra-cflags="-march=athlon -pipe"
make fprofiled VIDS="../enctests/deadline_cif.y4m"

Platform: X86
System: MINGW
asm: yes
avis input: yes
mp4 output: yes
pthread: yes
gtk: no
debug: no
gprof: no
PIC: no
shared: yes
visualize: no

[Avenger007]

Check the Git log! Awesome new changes!
:D lol
I guess you're referring to
commit a92bbc2325942d2ceab2ab043727f5943a33bd64 r974
Author: Jason Garrett-Glaser <darkshikari@gmail.com>
Date: Tue Sep 16 00:31:26 2008 -0700

Cache motion vectors in lowres lookahead
This vastly speeds up b-adapt 2, especially at large bframes values.
This changes output because now MV prediction in lookahead only uses L0/L1 MVs, not bidir. This isn't a problem, since the bidir prediction wasn't really correct to begin with, so the change in output is neither positive nor negative.
This also allowed the removal of some unnecessary memsets, which should also give a small speed boost.
Finally, this allows the use of the lowres motion vectors for predictors in some future patch.
Btw, thanks for the Athlon builds. :)

[Inventive Software]

Yeah, plus the VBV ones. ;)

[Ranguvar]

Yep :)
And I use march=athlon not for the Athlon CPU, but because it seems to improve performance on all CPUs (especially the Athlons), including my Core 2 Quad :)

[skystrife]

x264.977.modified.exe (http://www.mediafire.com/?a22z8pdp6aw) - Alternate Download (http://skystrife.com/x264/x264.977.modified.exe)

Patches used:

x264_hrd_pulldown.09_interlace.diff

gcc 3.4.5 fprofiled build with -march=pentium2.

Excited about the b-adapt 2 speedboost!

[kemuri-_9]

bah *stabs the planet*
I'm having problems accessing the git repository again... guess i'll try waiting it out.... :(

[Ranguvar]

here ya go, unpatched, straight git clone :)

http://www.rapidspread.com/file.jsp?id=dwephtordq

[kemuri-_9]

thanks, but it worked for me this time!
the repository needs to stop being so finicky!

[skystrife]

thanks, but it worked for me this time!
the repository needs to stop being so finicky!

I blame the new VLC release.