[HEVC] x265.EXE: mingw builds

[vhelp]

..i was still unable to play it back with lentoud decoder, that .12 new version, mind you.

i am using this version, hevcdecfltr.dll v2.0.1.6 but i dont know how to tell what levels it uses. i only know it handles the .hm10 i encode. i don't think the current version of x265 goes beyond hm10 anyway. i could be wrong.

Are you 100% sure that we can now use wpp? On my laptop, using the last non-leak build ( sorry, don't remember the name ), i was still unable to play it back with lentoud decoder, that .12 new version, mind you.
Also, just for the notes, any1 interested in simplistic gui avi/mp4/mkv->mp4/mkv with hevc video?

well, i have my own gui i'm working on. i added a playback tab for the .hm10 encoded files. i added avs script load, but will add a preview for it soon so i can preview what i'm scripting, then i will include a real-time preview for the encoding phase so i don't have to look at a progressbar only.

[Marchand]

We look forward to seeing your gui running and posted to test .... good job.

[El Heggunte]

NFO: source = http://forum.doom9.org/showthread.php?p=1641677#post1641677

Chikuzen released a new build of RawSource.

Result: HM 11 decodes with the same error. So it may be the proof for an encoder issue.

— frame 44

Raw HEVC file (~10 MB), x265 r3512, current command line options:

-q 20 -b 3 --weightp --weightbp --no-early-skip --no-fast-cbf --rdpenalty 1

[vhelp]

i don't understand. rawsource (an avisynth plugin) was updated to decode "raw" hevc (hm10) video ?

edit: nevermind. it is not the case.

[El Heggunte]

No, it was updated to support AviSynth 2.60 MT beta 4.

And LigH wanted to use it for viewing the YUV output from the HM reference decoder.

EDIT: too late for saying "nevermind"

[DarrellS]

I'm about to throw this computer out the window. I've been trying all weekend to install my boot drive backup onto a 160GB drive, erasing the drive with killdisk and formatting it to a 150GB NTFS partition only to have Acronis turn it into a drive with 74.5 GB of usable space and 74.5 GB of unallocated space. I've never had this problem before and don't know what to do to prevent it from happening. There are no settings to tell it not to do it and no options to tell it to do it so I have no clue why it only wants to give me 74.5 GB of usable space. The original boot drive was an 80 GB drive.


EDIT: I guess I have to convert to dynamic disk to extend the partition or use some other partition software and hope it doesn't screw it up more?


EDIT: Installed a program called EaseUS Partition Master, converted the unallocated space to a partition and merged C:\ and D:\ together to create a 149 GB C:\. Come to find out, my shortcut to Windows Explorer had gotten corrupted and creating a new shortcut on the desktop fixed my problem so all the troubled I went through this weekend was for nothing. I should've known since I had had this same problem a few years back but Acronis is known for causing the same type of problem with Windows always trying to search through the compressed TIB files and I have had problems with the IDE cable a few times in the past. Guess I can hook my IDE DVD burner back up although I rarely use it anymore. Maybe I'll just keep it around in case the SATA Burner wears out. At least I have a clean copy of XP with just the essentials I can use for testing. My case takes a SATA drive on the top that I can push in when needed and pulled out when I don't so it won't be too hard to switch back and forth. One way of dual booting I guess. Just have to tweak the BIOS when I switch boot drives.

[LigH.de]

Regarding the posts in the doom9 forum which El Heggunte reported:

I will run this test again with the most current x265.exe I can find here. Revision 3512 is already a little old.
__

Ran with a shorter and possibly simpler command line parameter set (-q 20 -b 1 --weightp --weightbp --rdpenalty 1); but x265.exe 0.3+504 crashed at frame 42 (MS APPCRASH dialog).
__

x265.exe 0.3+504 also crashes at frame 42 with: -q 20 -b 3 --weightp --weightbp --rdpenalty 1

Now with a very brief parameter set...
__

Only "-q 20": crashes at frame 46.

Again with version 0.3+355 (rev 3512) by MasterNobody: Passes frame 100 and continues...

[Vid01]

@LigH.de

Your crash is almost certainly a result of the known memory leak which has been found to exist in x265 after revision 0.3+381-6d14b538caa6 was commited on August 16th. You can confirm that by opening Task Manager and looking at the memory usage graph while the encoder is running. You should be able to see the distinct "stairway to crash" pattern and the encoder should normally halt when all available system memory has been consumed.

A bug report has been already filed at the x265 devs mailing list and hopefully the matter will be sorted soon. MasterNobody's build does not suffer from the leak, as it was generated before the 16/08 date.

[El Heggunte]

@ LigH: for chiming in, and welcome to Videohelp

@ Vid01: for reporting the memory leak issue to the x265 mailing list

TBH, sooner or later I myself was going to do that, because some users are "too busy" to send an e-mail to Steve Borho, and also it's too bad the guy from MCW doesn't watch this thread as often as he should

[El Heggunte]

[Not-so-OT]

To whom this may interest , Jan Ekström ( a.k.a. JEEB ) has just released his builds of HM 12.0,

http://x264.fushizen.eu/builds/hevc-hm/hm_12.0_r3541_release.7z

[El Heggunte]

For the record:

I tried to compile commit da8281e {"dpb: pass free list by reference (fixes memory leaks)"} ,
and this was the result:

[ 72%] Building CXX object common/CMakeFiles/common.dir/reference.cpp.obj
g:/0000/Mercurial263/TESTER/x265/source/common/reference.cpp: In member function 'void x265::MotionReference::generateReferencePlane(int)':
g:/0000/Mercurial263/TESTER/x265/source/common/reference.cpp:272:17: error: declaration of 'offset' shadows a member of 'this' [-Werror=shadow]
cc1plus.exe: all warnings being treated as errors
make[2]: *** [common/CMakeFiles/common.dir/reference.cpp.obj] Error 1
make[1]: *** [common/CMakeFiles/common.dir/all] Error 2
make: *** [all] Error 2

Now = 18:00 GMT, and the most recent commit is 407d8c6 {"dpb: add default case for switch, remove trailing white-space"} ---

--- but no, I won't mess around with GCC again until tomorrow, at least.

[x265]

@ LigH: for chiming in, and welcome to Videohelp

@ Vid01: for reporting the memory leak issue to the x265 mailing list

BTW, sooner or later I myself was going to do that, because some users are "too busy" to send an e-mail to Steve Borho, and also it's too bad the guy from MCW doesn't watch this thread as often as he should

Yes, thanks for reporting the bug through the proper channel -> the x265 development mailing list.

We use a mailing list to communicate and review patches:

• email: x265-devel@videolan.org
• signup: http://mailman.videolan.org/listinfo/x265-devel
• patchworks: http://patches.videolan.org/project/x265-devel/list/

Please do not email the lead developer directly with bug reports or feature requests. This will slow our efforts. Feature requests can come to me (privately or publicly), but you should appreciate that the companies funding this effort are at the head of the line when it comes to features or platform support.

We appreciate everyone who is testing x265, but please understand that we are not at version 1.0 yet. In other words, all features have not yet been implemented, and if you are creating a build from the development tip, you are likely to experience issues that come from features that are not fully implemented / optimized yet, like weighted prediction.

The guy from MCW subscribes to this thread, and he reads every post. I am in regular communication with our whole development team. Part of my job is to "fly air cover" for this team - to let them focus on the highest priorities. This team is extremely busy, and we have some big milestones ahead that we need to reach. Thanks for understanding if we don't respond to your bug report or feature request right away.

Tom
MulticoreWare

[Koti]

Thanks for the explanation Tom.
Wishing you and the developers the best on making deadlines/milestones and the best hevc encoder.

[x265]

By the way, we are in the midst of implementing significant changes to improve parallelism. The current build from the development tip crashes, and our developers are aware of the issue (caused by changes made last night).

[x265]

By the way, we are in the midst of implementing significant changes to improve parallelism. The current build from the development tip crashes, and our developers are aware of the issue (caused by changes made last night).

... aaaaand it's fixed.

[Edit]
This no longer crashes, but MD5 hashes of recon pictures are no longer matching decoder pictures following 63e0736c1f87, so more fixes remain to be done.

[deadrats]

Please do not email the lead developer directly with bug reports or feature requests. This will slow our efforts. Feature requests can come to me (privately or publicly), but you should appreciate that the companies funding this effort are at the head of the line when it comes to features or platform support.

i want to thank you for your honesty. i also want you to remember this statement when the day comes that you realize that x265 has failed to become the hevc encoder of choice for most consumer level users you hope it will be and are wondering why.

i certainly appreciate the desire and need to earn a buck, i owned my own business for years, and i understand the commitment to the bottom line. but you guys, as i've said before screwed up royally before you even got off the ground. you wanted to copy the x264LLC dual licensing scheme that they employ but i don't believe that is in a business' best interest.

development of your product should be guided by the desire to create the best quality product you possibly can not by what features some corporation funding the project wants to see included. they have their own agenda, which more often than not will clash with what the general public wants.

the result will be long development times, bloated code base and feature set and a product that fails to gain traction with the general public.

as for feature request, i will once again offer the opinion that you should choose your most talented programmer and assign him the task of porting the reference hevc encoder and decoder code to CUDA. don't worry about optimizing it for maximum performance as you port, just concentrate on getting to run on a discrete gpu and try to make it as feature complete, compared to the x86 version, as you possibly can.

one of the classic mistakes developers make is that they try and optimize the code as they go along then they find they have a hard time adding new features. optimization should be the last thing that is done, after all the features have been included and tested to work correctly.

[El Heggunte]

deadrats does have a point: it seems The X265 Project is being royally mismanaged.

[El Heggunte]

[ deleted Deleted DELETED ]

[Jacobr]

Just for anyone interested in checking out, my very basic and experimental gui, will upload soon.
Be sure to read notes, it was meant to just be a test script at the beginning, so you might run into unexpected crashes or similar things.
Once you get past 5 seconds preview should be fine

don't ask why 'create avs' is blinking or why some controls hide while others disable ..... just don't

also, you can obviously replace the x265 and/or ffmpeg etc.

[Marchand]

Jacobr

Gui basic functions normally
Video: OK
Mux audio: can not mix any audio with video

http://files.videohelp.com/u/223740/Videos.zip

Thanks

[El Heggunte]

Whatever.

[Jacobr]

Gui basic functions normally
Video: OK
Mux audio: can not mix any audio with video

http://files.videohelp.com/u/223740/Videos.zip

What was the problem exactly ?
You know, intermediate.avi is just the temp file, i can get mkvs and mp4s with audio on both laptop and pc automatically generated. They are written to the same dir as output, with .mkv or .mp4 extensions respectively and (2) added to name to avoid any possible overwrite if you are writing to the same dir as source. Muxing subs or chapters works flawlessly too.


Output dir should have .hm10 file, optionally with the sound file and possibly .ffindex, they aren't deleted just in case anyone needs them later on.

Whatever.

um.. ok

[vhelp]

test..something is screwing with the forum board.<br>

[vhelp]

corrupt reponse message.

[vhelp]

i have mixed success in latest x265 ea85b67 build.

encoding specs: 
--frames 122 --q 20 --keyint 24 --rect --max-merge 1 --hash 1    --wpp --no-rdo --no-rdoq --tu-intra-depth 1 --tu-inter-depth 1 --no-tskip --input "_video.y4m" -o "video.hm10"
------------------------------------------------------- 
computer #1, desktop pc, amd dual core x2 3600+, 2g ram winxp home sp2, results: random crashes, incomplete encodes
x265 [info]: detected SIMD architectures SSE2 SSE3 
x265 [info]: thread pool with 2 threads, WPP enabled (8 streams)
------------------------------------------------------- 
computer #2, netbook, atom processor, 500mb ram, win xp home sp3, results: success 
x265 [info]: detected SIMD architectures SSE2 SSE3 SSSE3 
x265 [info]: thread pool with 2 threads, WPP enabled (8 streams)
------------------------------------------------------- 
computer #3, dell cpu I3 4g ram windows xp pro sp3, results: success
x265 [info]: detected SIMD architectures SSE2 SSE3 SSSE3 SSE4.1 SSE4.2 
x265 [info]: thread pool with 4 threads, WPP enabled (8 streams)

@ all

i would like to make the suggestion that we use multiple cpu's if available because so far, i am getting success and failure on three. i ran the same test source three times and they all succeeded except when on the amd cpu. maybe it is a cpu problem. test your rig and post short specs if possible. don't need 200 pages to show this.

[Jacobr]

5000 frames:

x265.exe --q 23 --keyint 40 --max-merge 2 --me 3 --rect --hash 0 --no-rdo --no-rdoq --tu-intra-depth 1 --tu-inter-depth 2 --no-tskip --no-tskip-fast  --no-wpp  --amp  --input "C:\...intermediate.y4m"  --output "sthg.hm10"

5000 frames wpp and b frames:

C..x265.exe --q 23 --keyint 40 --max-merge 2 --me 3 -b 4 --rect --hash 0 --no-rdo --no-rdoq --tu-intra-depth 1 --tu-inter-depth 2 --no-tskip --no-tskip-fast  --amp  --input "C:\...intermediate.y4m"  --output "C....hm10"

1)
simd up to avx:
Intel(R)_Core(TM)_i5-3210M_CPU_@_2.50GHz
4gb ram
windows 8
/-------------------/
Both cmdlines encoding 100% correctly

[ozok]

I updated my GUI to build 92, you can download it from here https://bitbucket.org/ozok/x265gui/downloads. If you selected to disable audio muxing could fail. Now it should work without a problem.

[Marchand]

I updated my GUI to build 92, you can download it from here https://bitbucket.org/ozok/x265gui/downloads. If you selected to disable audio muxing could fail. Now it should work without a problem.

Ozok

Great Job. Audio and Video - Test OK

General
Complete name : C:\Users\Marchand\Videos\00200.mp4
Format : MPEG-4
Format profile : Base Media
Codec ID : iso4
File size : 348 KiB
Duration : 7s 465ms
Overall bit rate mode : Variable
Overall bit rate : 382 Kbps
Encoded date : UTC 2013-08-29 00:09:29
Tagged date : UTC 2013-08-29 00:09:29

Video
ID : 1
Format : HEVC
Format/Info : High Efficiency Video Coding
Format profile : Main@L9.0
Codec ID : hvc1
Codec ID/Info : High Efficiency Video Coding
Duration : 7s 200ms
Bit rate : 258 Kbps
Maximum bit rate : 330 Kbps
Width : 720 pixels
Height : 576 pixels
Display aspect ratio : 5:4
Frame rate mode : Constant
Frame rate : 25.000 fps
Standard : PAL
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Bits/(Pixel*Frame) : 0.025
Stream size : 226 KiB (65%)
Title : h265:FMT=HEVC@GPAC0.5.1-DEV-rev4707
Encoded date : UTC 2013-08-29 00:09:29
Tagged date : UTC 2013-08-29 00:09:29

Audio
ID : 2
Format : AAC
Format/Info : Advanced Audio Codec
Format profile : LC
Codec ID : 40
Duration : 7s 465ms
Source duration : 7s 488ms
Bit rate mode : Variable
Bit rate : 129 Kbps
Maximum bit rate : 141 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Compression mode : Lossy
Delay relative to video : -21ms
Stream size : 118 KiB (34%)
Source stream size : 118 KiB (34%)
Tagged date : UTC 2013-08-29 00:09:29

Jacobr

Jacobr, I've tried everything and could not audio. I switched mp4box.exe, graph2.exe, x265.exe and no audio.

I think it must be something in the configuration of my Haali Media Splitter, time to send your script to graphstudionext because I do not use ffdshow and klite mega codec pack.

Likewise, I think the problem is in my filters same configuration. I'll keep trying.

Thank you.

[DarrellS]

5000 frames:

x265.exe --q 23 --keyint 40 --max-merge 2 --me 3 --rect --hash 0 --no-rdo --no-rdoq --tu-intra-depth 1 --tu-inter-depth 2 --no-tskip --no-tskip-fast  --no-wpp  --amp  --input "C:\...intermediate.y4m"  --output "sthg.hm10"

5000 frames wpp and b frames:

C..x265.exe --q 23 --keyint 40 --max-merge 2 --me 3 -b 4 --rect --hash 0 --no-rdo --no-rdoq --tu-intra-depth 1 --tu-inter-depth 2 --no-tskip --no-tskip-fast  --amp  --input "C:\...intermediate.y4m"  --output "C....hm10"

1)
simd up to avx:
Intel(R)_Core(TM)_i5-3210M_CPU_@_2.50GHz
4gb ram
windows 8
/-------------------/
Both cmdlines encoding 100% correctly

Neither crash but both only encode at .70 fps. The first code only uses one thread. I'll try using --thread 4 and see if it helps or if it crashes.

The second code uses 4 threads and 17 streams but still only encodes at .70 fps using my Q6600 CPU.

Oh, and MP4Box won't accept .hm10


EDIT: I didn't check to see if the --wpp file would play either and that has been the issue all along. I get great speeds with x265's suggested argument (over 2 fps) but the files won't play. Don't know if it's an XP issue, my CPU or if the splitter or decoder.

[DarrellS]

Well, I just encoded a few files a while ago and now I can't get any of my bat files to work. They produce a y4m file and then nothing. I tried replacing the x264.exe file with a fresh copy but still no go. The only thing I can think of is that I downloaded that GUI up above that didn't work and it has control of the encoder some how.

I'll restart the computer and see what happens.

It was the frame size. 512x388 instead of 512x384. encoded at 3.60 fps. Did another 1920x1080, this time with black background and got 1.61 fps. I guess that's OK for 1920x1080. Better than .70 fps. All the wpp files played in MPC-HC except for the 1920x1080 files. They play in the Osmos4 player and will open in MPC-HC after I open them in the Osmos4 player but they won't play and the frame is corrupt.

Something isn't right though because the original is a 294 KB x264 mkv (man that's pretty small for 480 frames) and the x265 mp4 is 1,154 KB.