x265 HEVC Encoder

[vadlerg]

Thank You for sharing the build.

--aq-auto 10 works. It is about 3% faster than --aq-mode 4 and may preserve a bit more detail.

When using --rd 4 --dynamic-rd 3 options I've got the warning

x265 [WARN]: Dynamic-rd disabled, requires RD <= 4, VBV and aq-mode enabled

That seems to be not rational. There must be some typo in the source, the build itself is OK.

[Boulder]

Originally Posted by vadlerg

Thank You for sharing the build.

--aq-auto 10 works. It is about 3% faster than --aq-mode 4 and may preserve a bit more detail.

When using --rd 4 --dynamic-rd 3 options I've got the warning


That seems to be not rational. There must be some typo in the source, the build itself is OK.

The error message is a bit bad English to me. Do you have VBV enabled as well? I believe the AQ-mode checks ok if AQ auto is enabled.

if (m_param->dynamicRd && (!bIsVbv || !p->rc.aqMode || p->rdLevel > 4))
    {
        p->dynamicRd = 0;
        x265_log(p, X265_LOG_WARNING, "Dynamic-rd disabled, requires RD <= 4, VBV and aq-mode enabled\n");
    }

[vadlerg]

VBV is set automatically because of the parameters

--crf 18 --preset slower --profile main10 --level-idc 52

.
aqMode should be 2 by default, but --aq-auto 10 sets it to

AQ: mode / str(edg) / qg-size / cu-tree : auto-hyst-aq5 / 1.0(1.0) / 32 / 1

rdLevel is set to 4 and encoder still disabling dynamic-rd, even if I set --rd to 3.
In the INFO section I see

x265 [INFO]: tools: rect amp limit-modes rd=4 psy-rd=2.00 rdoq=2 psy-rdoq=1.00
x265 [INFO]: tools: rskip mode=2 rskip-edge-threshold=0.03 limit-tu=4 signhide
x265 [INFO]: tools: tmvp b-intra strong-intra-smoothing deblock sao
x265 [INFO]: tools: selective-sao

[Boulder]

Originally Posted by vadlerg

aqMode should be 2 by default

Then it must be because of --aq-auto. I'll take a look at it when I'm back home from work and adjust that check in the code accordingly to make it work again.

[LeXXuz]

Originally Posted by vadlerg

aqMode should be 2 by default, but --aq-auto 10 sets it to

Isn't that what it is supposed to do?
https://forum.doom9.org/showthread.p...00#post1981600

[vadlerg]

Of course it is what it supposed to be but this should not disable dynamicRd.

[Boulder]

Originally Posted by vadlerg

Of course it is what it supposed to be but this should not disable dynamicRd.

Yes, it's probably just that validation which has not been updated accordingly. I don't remember I've ever seen the dynamic RD option used by anyone so it's definitely understandable.

[Boulder]

I found the cause of the issue, it's not auto-aq related. The validation is almost ok, it just doesn't understand that --level-idc enables VBV. If you add the VBV parameters to the command line, it will work.

EDIT: Here's a build where the validation is fixed (if VBV is not set and --level-idc is 0, it will disable dynamic-rd): https://drive.google.com/file/d/1U9c...usp=share_link

[benwaggoner]

Originally Posted by Boulder

Yes, it's probably just that validation which has not been updated accordingly. I don't remember I've ever seen the dynamic RD option used by anyone so it's definitely understandable.

I've used it before, with high resolution CRF with VBV encodes. It lets the encoder run faster when it can achieve CRF, and then slows down to improve quality when VBV limitations keep CRF from being hit.

It can improve throughput across many clips, but perf improvements for any given encode are unpredictable, varying with how often the content would hit the VBV.

Of course, this will increase ABR versus just using full complexity the whole time, so it's quite scenario specific. There is also the risk of the different modes causing some "stylistic" discontinuities.

[vadlerg]

Thank You. That's it.

[jpsdr]

I'm stuck with an issue i've no idea for now how to solve, and i may think that's there is just no solution at all (unless a loooooot of rework of the code).
After adding just the command line asked by Boulder, i've hit the C1061 error issue with MS compiler...
Why have they done this...????
As it's specific to MS compiler, i still can make the clang version, but for now, it's the only build i can do.
To do the others builds, i have to see how i can build x265 with gcc, maybe i'll do that, but not the time for now...

[jpsdr]

Finaly i made some tests for gcc build, but when things want to get in your way, they are very efficients...

I've tested msys script build, but what i'm looking for is where i can find the parameter list you see in command like this:

cmake -G "MSYS Makefiles" ../../../source -DHIGH_BIT_DEPTH=ON -DEXPORT_C_API=OFF -DENABLE_SHARED=OFF -DENABLE_CLI=OFF -DMAIN12=ON

I want to make a static standalone .exe, not a one when you start it asked a lot of gcc dll...!
This is possible with cmake-gui (i have trouble to make it start also), but cmake-gui doesn't give all the options (you don't have 12bits for exemple).
Also, i would like to add "-mavx2" in the gcc command, no idea how to do that's kind of stuff (this is why i choose the Visual Studio solution, where i have access to all the compiler options in the ide).

Edit
Even with "static link CRT" checked in cmake-gui, the .exe from gcc build is still asking for dll... Not the same than my first attempt without, but still...

So, gcc builders, can you provide my your whole complete build script ?
As i said, when things want to get in your way, they are very efficients !!

[Boulder]

For GCC builds, I use Media Autobuild Suite to take care of the cumbersome scripting. You can use custom patches, for example with x265, create a file named x265_git_extra.sh and add the patch in the main 'build' folder and it will get applied during the compile operation. This is what I have in the file.

#!/bin/bash

_pre_ninja(){

patch -p 1 -i "$LOCALBUILDDIR/x265-jpsdr-aqauto-v5.patch"

}

[qyot27]

Originally Posted by jpsdr

I've tested msys script build, but what i'm looking for is where i can find the parameter list you see in command like this:

cmake -G "MSYS Makefiles" ../../../source -DHIGH_BIT_DEPTH=ON -DEXPORT_C_API=OFF -DENABLE_SHARED=OFF -DENABLE_CLI=OFF -DMAIN12=ON

cmake . -LAH will emit the full list of variables that get set with their values, although it isn't all that nicely formatted. In many cases it's just easier to open CMakeLists.txt and see what things are either listed as option() or being toggled as an environment variable with set() or some such.

I want to make a static standalone .exe, not a one when you start it asked a lot of gcc dll...!
This is possible with cmake-gui (i have trouble to make it start also), but cmake-gui doesn't give all the options (you don't have 12bits for exemple).
Also, i would like to add "-mavx2" in the gcc command, no idea how to do that's kind of stuff (this is why i choose the Visual Studio solution, where i have access to all the compiler options in the ide).

Edit
Even with "static link CRT" checked in cmake-gui, the .exe from gcc build is still asking for dll... Not the same than my first attempt without, but still...

So, gcc builders, can you provide my your whole complete build script ?
As i said, when things want to get in your way, they are very efficients !!

See example here (yes, x265 is in the guide, but for clarity's sake use a smaller example):
https://github.com/qyot27/mpv/blob/1...dious.txt#L475

Most GCC toolchains that are widely distributed are built assuming shared runtime by default, you have to either pass something like -static-libgcc -static-libstdc++ to LDFLAGS (in CMake, -DCMAKE_SHARED_LINKER_FLAGS for libraries and -DCMAKE_EXE_LINKER_FLAGS for executables) or you can just turn this off entirely by building GCC with static runtime only (which is what I do when building the MinGW/GCC toolchain, hence why the guide doesn't use those flags).

I need to do a good refresh on that guide because while I typically do a once-through every six months after moving to the current release of Ubuntu, I was so stressed out last year and now spend >95% of my time in Linux that I still haven't done so since last May or something like that.

[jpsdr]

Ok, translation issue from my part, my question is not properly made.

I was asking, from people using gcc, their whole complete build script for building static x265 (not building gcc...).

[LigH]

Mine (depends on the media-autobuild suite directory tree and its MSYS2/MinGW64 shell).

[jpsdr]

@LigH
Is in your script the following line:

export CXXFLAGS=

allow to add compiler options (like -mavx2) that will be used afterward ?

[LigH]

No, just basic adaptions of the Multicoreware templates.

[jpsdr]

No luck, i've tried several options, but i always have an error message telling me that libstdc++6.dll is missing when i try to run the x265.exe builded...
Don't know what option i have to use to make a standalone version.

[LeXXuz]

Originally Posted by jpsdr

@LeXXuz
For the same command line effect, you have to replace --sbrc --sbrc-aq5 --sbrc-hyst with --aq-auto 10 assuming you don't want the new sbrc (wich has totaly nothing to do with the first patch) and just keep the auto AQ feature.
(And still check with the log file )

jpsdr, is it normal that Mediainfo shows all kinds of aq-mode when encoding with --aq-mode 5? Some files show aq-mode=2, others 3, 4 or 5. All have been encoded with --aq-mode 5 and --aq-auto 10.