Hey everyone!
I just start some test with "Dither_convert_8_to_16" filter, but I got problem when i used Dither_out filter
here is my code:
the video resolution become 2560x720 .Code:ffvideoSource("E:\encoding\test_new\TEST[1080p].mkv") Dither_convert_8_to_16() Dither_resize16nr(1280,720, csp = "YV24") Dither_out()
any help plz
thanx.
+ Reply to Thread
Results 1 to 22 of 22
-
Last edited by alkoon; 30th Dec 2017 at 11:10.
-
Can't this be a problem?
BernixResizes a 16-bit clip.
The function can resize interlaced content, but only if presented as separated, interleaved fields. It uses the Avisynth parity information to detect fields and maintain the correct chroma and luma positions. If this behaviour is not desired, add an AssumeFrameBased() before the resizing call. -
This is completely normal. AviSynth (not AviSynth+) cannot natively handle >8 bit so if you view the script it will look like this with double width (basically it converts the 16 bits into 8+8 bits - there is also stacked variant). You just have to tell your encoder that input is actually 16 bit.
E.g.:
avs2pipemod -rawvideo "script.avs" | x264 - --demuxer raw --input-depth 16 --input-res 1280x720 --fps 24000/1001 -o "output.264"
You can read about it in the Dither documentation. It's very good. (But not developed anymore AFAIK. People have switched to AviSynth+ or VapourSynth.) -
Actually I have AviSynth+ (r1576) in my laptop, and in my server I have AviSynth 2.6
in both I got same result , even with >>
but everything is okay with I used DitherPost(), but Dither_out() is the right filter to useavs2pipemod -rawvideo "script.avs" | x264 - --demuxer raw --input-depth 16 --input-res 1280x720 --fps 24000/1001 -o "output.264"
so I think isn't okay to use DitherPost instead of Dither_out !DitherPost = Adds the dither, turning a 16-bit clip generated by Dither1Pre, Dither2Pre or a mod16 denoiser into a regular 8-bit clip.
Dither_out() =This function allows Avisynth to output 16-bit YUV pixel components
thank you. -
Are you encoding 10bit444 ? You have to use a 10bit x264 binary and specify --input-csp i444 --output-csp i444
(If not 444 then why "YV24" in your script ?) -
Two possible problems:
A. As poisondeathray saw I missed the csp = "YV24". Then the command is a bit different:
avs2pipemod -rawvideo "script.avs" | x264 - --demuxer raw --input-depth 16 --output-depth 10 --input-csp i444 --output-csp i444 --input-res 1280x720 --fps 24000/1001 -o "output.264"
B. You are using AviSynth+ with ffms2. If ffms2 is recent and your source is not 8 bit it may be in non-8-bit-AviSynth+ format. But Dither package doesn't understand those and treats input like 8 bit. Then you get problems.
To properly handle either:
or simply with native AviSynth+Code:ffvideoSource("E:\encoding\test_new\TEST[1080p].mkv") ConvertBits(16) ConvertToStacked() Dither_resize16nr(1280,720, csp = "YV24") Dither_out()
For the last the command is a bit simpler:Code:ffvideoSource("E:\encoding\test_new\TEST[1080p].mkv") ConvertBits(16) ConvertToYUV444() Spline16Resize(1280, 720)
avs2pipemod -y4mp "script.avs" | x264 - --demuxer y4m --output-depth 10 --output-csp i444 -o "output.264"
Note that for 10 bit output you may now need to specify --output-depth 10. Latest x264 now supports multi-depth builds (like x265). -
her is my bat script
Maybe i should use the Avisynth+ r2xx ver & up!!avs2yuv test.avs - | x264.2851kMod.10bit.x86_64 --preset veryslow --crf 16 --input-depth 10 --input-csp i444 --output-csp i444 --deblock 1:-1 --aq-mode 2 --subme 10 --no-fast-pskip --chroma-qp-offset -3 --qcomp 0.65 --aq-strength 0.7 --fps 24000/1001 --output "testPC4.mkv" - --demuxer y4m
gonna download Avisynth+ r2580 ver and i will try.
Thanx everyone =). -
1. Use avs2pipemod, not avs2yuv.
2. Use the commands I used, don't change them however you like.
If it still does not work show what script and command you are using now. There have been several examples and now I don't know what you are using right now. -
where can i download the latest x264 ?
I download from here http://www.x264.nl/x264_main.php & http://komisar.gin.by/
both of them gives me an error, unknown --output-depth 10
isn't --input-depth 10 enough ?!
So I just tried this (without --output-depth 10 )
script.avs :
>> Video resolution become 2560x720ffvideoSource("TEST[1080p].mkv")
ConvertBits(16)
ConvertToStacked()
Dither_resize16nr(1280,720, csp = "YV24")
Dither_out()
trim(50,100)
bat:
avs2pipemod.exe -y4mp "script.avs" | x26410b.exe - --demuxer y4m --input-depth 16 --input-res 1280x720 --output-csp i444 --fps 24000/1001 -o "output.264"
the encoder couldn't encode the input file only 1 frameLast edited by alkoon; 30th Dec 2017 at 14:27.
-
Latest official binary is r2893. It's very recent, komisar doesn't have it yet. (x264.nl has been abandoned for years)
-
ok
I used this one here :
https://download.videolan.org/x264/binaries/win32 (x264-10b-r2851-ba24899.exe - 26-Jun-2017 20:04 )
same I got an error :
unknown --output-depth 10 -
ok
bat:
result: attachmentavs2pipemod.exe -y4mp "script.avs" | x264-r2893-b00bcaf.exe - --demuxer y4m --input-res 1280x720 --output-depth 10 --output-csp i444 -o "output.264" -
Don't mix and match commandlines. Only raw requires specifying parameters, y4m has header information and should convey info like resolution, frame rate
It works for me with komisar's old 10bit x264 . I tested x86 . Mine is renamed to "x264k32_10.exe"
input.mp4 was a "regular" 8bit AVC 1920x1080 24fps (not 23.976 like yours, so you would use --fps 24000/1001) . ffms2 would have worked tooCode:avs2pipemod -rawvideo input.avs | x264k32_10.exe --demuxer raw --input-depth 16 --input-res 1280x720 --fps 24 --input-csp i444 --output-csp i444 -o output.mp4 -
Code:LSmashVideoSource("F:\input.mp4") Dither_convert_8_to_16() Dither_resize16nr(1280,720, csp = "YV24") Dither_out() -
Thank you guys, but the problem still exist with avs script
this command help now to encode 1280x720 video, but the major problem still a same before & after encoding
avs2pipemod.exe -rawvideo "script.avs" | x264-r2893-b00bcaf.exe - --demuxer raw --input-res 1280x720 --input-depth 10 --output-depth 10 --output-csp i444 -o "output.264"-
before
after
Last edited by alkoon; 30th Dec 2017 at 23:01.
-
For what you do Vapoursynth instead of Avisynth might be a better choice, 64bit Vapoursynth, starting to forget about 32bit apps,
I noticed that guys working with anime videos are jumping that direction. -
How are you decoding the output file ?
Cut a piece of your mkv with mkvtoolnix and upload the sample here -
here it's :
https://files.fm/f/pngrnbgn
Actually I used Vapoursynth but still not familiar with it, but i think I'm going use it again.For what you do Vapoursynth instead of Avisynth might be a better choice, 64bit Vapoursynth, starting to forget about 32bit apps,
I noticed that guys working with anime videos are jumping that direction.
thank you guys for help =). -
-
Similar Threads
-
Best Resize Filter For...
By ben45 in forum Video ConversionReplies: 9Last Post: 6th Aug 2016, 15:44 -
IR filter in videos.
By ressurex in forum EditingReplies: 5Last Post: 29th Sep 2015, 08:29 -
ND or CP filter for GoPro
By jucas85 in forum Camcorders (DV/HDV/AVCHD/HD)Replies: 7Last Post: 22nd Dec 2014, 11:50 -
Detect used filter?
By xicudiz in forum Newbie / General discussionsReplies: 6Last Post: 27th Mar 2014, 20:22 -
HQDering filter
By brusno in forum Video ConversionReplies: 5Last Post: 23rd Mar 2013, 06:19


Quote
