Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Capturing and Editing Video > Avisynth Development

Reply
 
Thread Tools Search this Thread Display Modes
Old 28th November 2016, 19:48   #2321  |  Link
leoenc
Registered User
 
Join Date: Mar 2007
Posts: 192
Quote:
Originally Posted by Groucho2004 View Post
In that case use DSS or AVIsource, no indexing required.
DSS doesn't support 10-bit with AVIsynth+. ffms does.
leoenc is offline   Reply With Quote
Old 29th November 2016, 13:34   #2322  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 6,753
Well, if ya ain't got no time for a keyframe index, ya ain't got time for optimal quality either.
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 30th November 2016, 09:37   #2323  |  Link
leoenc
Registered User
 
Join Date: Mar 2007
Posts: 192
Quote:
Originally Posted by LigH View Post
Well, if ya ain't got no time for a keyframe index, ya ain't got time for optimal quality either.
You have a point
You can't rush art...
leoenc is offline   Reply With Quote
Old 29th December 2016, 13:40   #2324  |  Link
Myrsloik
Professional Code Monkey
 
Myrsloik's Avatar
 
Join Date: Jun 2003
Location: Kinnarps Chair
Posts: 2,548
Here's the RC for FFMS 2.23.1. The only change is that it's been compiled with a recent FFmpeg which should fix some issues. For example conversion to planar RGB output now works properly.

Report success/failure. I'll make the release official if nothing is broken.
__________________
VapourSynth - proving that scripting languages and video processing isn't dead yet
Myrsloik is offline   Reply With Quote
Old 29th December 2016, 16:49   #2325  |  Link
Atak_Snajpera
RipBot264 author
 
Atak_Snajpera's Avatar
 
Join Date: May 2006
Location: Poland
Posts: 7,806
Quick question. Is FFM2 v2.23.1 with Interlaced AVC footage still reports double number of frames?
Atak_Snajpera is offline   Reply With Quote
Old 29th December 2016, 18:43   #2326  |  Link
dipje
Registered User
 
Join Date: Oct 2014
Posts: 268
@Myrsloik It's been a while since I tried FFMS2 builds. I tested 2.23.1 x64 through Vapoursynth. My usual tests are to see if it opens variants of Cineform, Prores and DNXHD OK or not.

I tried Cineform YUV422 P10, Cineform RGB, Prores HQ (422 P10), Prores 4444 without alpha (444 P10) and Prores 4444XQ with and without alpha (444 P10, recently added to ffmpeg so unknown vtag in previous ffms2 builds), and DNXHD 422 P10 and DNXHD 444 P10.

All open OK, but I have issues with Cineform RGB. A recent ffmpeg build (ffprobe) reports it as gbrap12le(10 bpc, bt709, progressive). I see that as planar RGB with alpha (which sounds correct). I don't get why the 'p12' and then the '10 bpc'. Cineform is often named as '12bit RGBA' while internally everything works as 10bpc as far as I know.

ANYWAY (I digress), the issue I have is that if I just try to open it without any special parameters, Vapoursynth (or ffms2, don't really know) gives an error 'Resize error 1026: RGB color family cannot be YUV'. The pixel format of the clip is reported as vs.RGB48. I only get the error while trying to preview / display the clip in Vapoursynth Editor. If I just 'run' the script VapoursynthEditor reports a valid script with RGB48 output. It seems as if it actually gives YUV data but reports it as RGB or the other way around or something? If I open the .vpy script in VDFilterMod (with RGB48) I just get the correct video showing. It seems as if the video output is OK, but wrongly tagged that trips up the preview-code or the resizers in Vapoursynth / VapoursynthEditor.

(I know realize this might also be VapoursynthEditor related which of course is not an official part of Vapoursynth).

If I add the 'format = vs.RGB48' or 'format = vs.RGB30' options the script still won't display in VapoursynthEditor, but if I add 'format = vs.YUV444P16' or 'format = vs.YUV444P10' the script opens and displays fine n VapoursynthEditor. But I don't know if I'm loosing some data or precision now? (Inside of FFMS2 the clip is apparently RGB '12bit' which it puts in RGB48, so if I add the format parameter it gets converted 'inside FFMS2' before being passed to Vapoursynth I guess, so the only thing I'm loosing in precision is RGB to YUV444 conversion I guess.

edit: Tried with L-Smash r921 (latest in the dropbox folder as of now) and it crashes on the Cineform RGB (RGBA) and the prores 4444 XQ Alpha file, so l-smash seems to have an issue with alpha channels so I can't compare. FFMS2 wins for now

Last edited by dipje; 29th December 2016 at 18:47.
dipje is offline   Reply With Quote
Old 29th December 2016, 18:50   #2327  |  Link
Myrsloik
Professional Code Monkey
 
Myrsloik's Avatar
 
Join Date: Jun 2003
Location: Kinnarps Chair
Posts: 2,548
Quote:
Originally Posted by dipje View Post
@Myrsloik It's been a while since I tried FFMS2 builds. I tested 2.23.1 x64 through Vapoursynth. My usual tests are to see if it opens variants of Cineform, Prores and DNXHD OK or not.
...
Are you really using the latest versions of everything? That Vapoursynth Editor issue sounds familiar
__________________
VapourSynth - proving that scripting languages and video processing isn't dead yet
Myrsloik is offline   Reply With Quote
Old 29th December 2016, 18:54   #2328  |  Link
Mystery Keeper
Beyond Kawaii
 
Mystery Keeper's Avatar
 
Join Date: Feb 2008
Location: Russia
Posts: 724
L-Smash doesn't set correct matrix property for the clip after conversion. It confuses VS internal resize plugin which is used in VSEditor for preview.
__________________
...desu!
Mystery Keeper is offline   Reply With Quote
Old 29th December 2016, 20:30   #2329  |  Link
dipje
Registered User
 
Join Date: Oct 2014
Posts: 268
I might have an outdated VapoursynthEditor on this machine, hold on let me update (since it only happened with rgb-a with ffms2 didn't think of anything else).

edit: Nope, that's not it. Latest r13 vs editor: Error on frame 0 request: Resize error 1026: RGB color family cannot be YUV

Last edited by dipje; 29th December 2016 at 20:32.
dipje is offline   Reply With Quote
Old 2nd January 2017, 19:22   #2330  |  Link
VS_Fan
Registered User
 
Join Date: Jan 2016
Posts: 98
Quote:
Originally Posted by Myrsloik View Post
Here's the RC for FFMS 2.23.1. The only change is that it's been compiled with a recent FFmpeg which should fix some issues. For example conversion to planar RGB output now works properly.

Report success/failure. I'll make the release official if nothing is broken.
Good: It's working fine reading FFV1 videos (Versions 1 and 3). Previous release of FFMS2 (2.23) crashed immediately.

Thanks
VS_Fan is offline   Reply With Quote
Old 2nd January 2017, 22:20   #2331  |  Link
Myrsloik
Professional Code Monkey
 
Myrsloik's Avatar
 
Join Date: Jun 2003
Location: Kinnarps Chair
Posts: 2,548
Now it's officially released as 2.23.1. Everyone should update. Should actually be usable as a high bitdepth image source now. Kinda.
__________________
VapourSynth - proving that scripting languages and video processing isn't dead yet
Myrsloik is offline   Reply With Quote
Old 3rd January 2017, 09:54   #2332  |  Link
dipje
Registered User
 
Join Date: Oct 2014
Posts: 268
So my yuv/rgb issue is classified as a problem with VapoursynthEditor ?

Since it happens only with ffms2 and only with Cineform RGB I thought it was something funky ffms2 was signaling. (Something like 'hey Cineform, that is yuv p10, but then not expecting that recent ffmpeg/libavcodec versions are returning RGB p10 or RGB p12 or something like that )
dipje is offline   Reply With Quote
Old 4th January 2017, 07:06   #2333  |  Link
Mystery Keeper
Beyond Kawaii
 
Mystery Keeper's Avatar
 
Join Date: Feb 2008
Location: Russia
Posts: 724
Quote:
Originally Posted by dipje View Post
So my yuv/rgb issue is classified as a problem with VapoursynthEditor ?

Since it happens only with ffms2 and only with Cineform RGB I thought it was something funky ffms2 was signaling. (Something like 'hey Cineform, that is yuv p10, but then not expecting that recent ffmpeg/libavcodec versions are returning RGB p10 or RGB p12 or something like that )
There's no problem with VapourSynth Editor. There's problem with L-SMASH.
Instead of
Code:
res = core.lsmas.LWLibavSource(a,threads = 1,format = "RGB48")
Use
Code:
src_file = r"D:\vstests\遊魂 2 -you're the only one- 遊戲開場動畫(中文字幕版).mp4"
src = core.lsmas.LWLibavSource(src_file, threads = 1)
src = core.resize.Spline16(src, format=vs.RGB48)
__________________
...desu!
Mystery Keeper is offline   Reply With Quote
Old 4th January 2017, 11:06   #2334  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 6,753
Quote:
Originally Posted by TheFluff View Post
Quote:
Originally Posted by LigH View Post
Is FFIndex (either in ffms2.dll or via ffmsindex.exe) able to support a list of VOB segments as source (like DGMPGDec does), if not even an IFO (ideally with a PGC/angle definition, or at least a PGC/angle VOB set stripped by the ripper in IFO/movie mode or PGCDemux), and an MPLS Blu-ray playlist?

If not yet ... where is the recommendable place to add such a feature request?
__

P.S.: Exists already...
It's been on the todo list for ages, issue #33 is an even older feature request for the same thing.
Phew, 4 years ago already. And since converters are moving towards 64 bit, the existence of source plugins supporting DGMPGDec indexes (*.d2v) could be useful.
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid
LigH is offline   Reply With Quote
Old 4th January 2017, 11:21   #2335  |  Link
dipje
Registered User
 
Join Date: Oct 2014
Posts: 268
Quote:
Originally Posted by Mystery Keeper View Post
There's no problem with VapourSynth Editor. There's problem with L-SMASH.

.. but we're talking about ffms2 here, not l-smash.
dipje is offline   Reply With Quote
Old 4th January 2017, 13:52   #2336  |  Link
Jamaika
Registered User
 
Join Date: Jul 2015
Posts: 697
Quote:
Originally Posted by dipje View Post
@Myrsloik It's been a while since I tried FFMS2 builds. I tested 2.23.1 x64 through Vapoursynth. My usual tests are to see if it opens variants of Cineform, Prores and DNXHD OK or not.

I tried Cineform YUV422 P10, Cineform RGB, Prores HQ (422 P10), Prores 4444 without alpha (444 P10) and Prores 4444XQ with and without alpha (444 P10, recently added to ffmpeg so unknown vtag in previous ffms2 builds), and DNXHD 422 P10 and DNXHD 444 P10.

All open OK, but I have issues with Cineform RGB. A recent ffmpeg build (ffprobe) reports it as gbrap12le(10 bpc, bt709, progressive). I see that as planar RGB with alpha (which sounds correct). I don't get why the 'p12' and then the '10 bpc'. Cineform is often named as '12bit RGBA' while internally everything works as 10bpc as far as I know.

If I add the 'format = vs.RGB48' or 'format = vs.RGB30' options the script still won't display in VapoursynthEditor, but if I add 'format = vs.YUV444P16' or 'format = vs.YUV444P10' the script opens and displays fine n VapoursynthEditor. But I don't know if I'm loosing some data or precision now? (Inside of FFMS2 the clip is apparently RGB '12bit' which it puts in RGB48, so if I add the format parameter it gets converted 'inside FFMS2' before being passed to Vapoursynth I guess, so the only thing I'm loosing in precision is RGB to YUV444 conversion I guess.

edit: Tried with L-Smash r921 (latest in the dropbox folder as of now) and it crashes on the Cineform RGB (RGBA) and the prores 4444 XQ Alpha file, so l-smash seems to have an issue with alpha channels so I can't compare. FFMS2 wins for now
Never converts RGB48 to RGBxx CineForm, ProRes or DNxHD. This is absurd.
There are limitations with the colormatrix. ProRes is always in RGB BT601 and CineForm BT709.
The sony vegas 14 is illegal and incorrect operation. Even using the V210 or R210 then I would use yuv422p10le.
In my opinion this is not a problem editors only muxing encoders.
Jamaika is offline   Reply With Quote
Old 4th January 2017, 14:39   #2337  |  Link
Myrsloik
Professional Code Monkey
 
Myrsloik's Avatar
 
Join Date: Jun 2003
Location: Kinnarps Chair
Posts: 2,548
Quote:
Originally Posted by LigH View Post
Phew, 4 years ago already. And since converters are moving towards 64 bit, the existence of source plugins supporting DGMPGDec indexes (*.d2v) could be useful.
It's stuck in "patches welcome" land. I never need it and now we have d2vsource as well...
__________________
VapourSynth - proving that scripting languages and video processing isn't dead yet
Myrsloik is offline   Reply With Quote
Old 4th January 2017, 14:49   #2338  |  Link
LigH
German doom9/Gleitz SuMo
 
LigH's Avatar
 
Join Date: Oct 2001
Location: Germany, rural Altmark
Posts: 6,753
... for VapourSynth. And what do users of AviSynth+ 64-bit use?

Oh, there is one.
__________________

New German Gleitz board
MediaFire: x264 | x265 | VPx | AOM | Xvid

Last edited by LigH; 4th January 2017 at 14:51.
LigH is offline   Reply With Quote
Old 4th January 2017, 15:02   #2339  |  Link
Reel.Deel
Registered User
 
Join Date: Mar 2012
Location: Texas
Posts: 1,664
Quote:
Originally Posted by LigH View Post
... for VapourSynth. And what do users of AviSynth+ 64-bit use?

Oh, there is one.
MPEG2DecPlus: https://github.com/chikuzen/MPEG2DecPlus
Binary here; this version is for AVS+ only, support for avs 2.6 was added but you'll have to compile it yourself.
Reel.Deel is offline   Reply With Quote
Old 4th January 2017, 17:26   #2340  |  Link
dipje
Registered User
 
Join Date: Oct 2014
Posts: 268
Quote:
Originally Posted by Jamaika View Post
Never converts RGB48 to RGBxx CineForm, ProRes or DNxHD. This is absurd.
There are limitations with the colormatrix. ProRes is always in RGB BT601 and CineForm BT709.
The sony vegas 14 is illegal and incorrect operation. Even using the V210 or R210 then I would use yuv422p10le.
In my opinion this is not a problem editors only muxing encoders.
Not sure what you're saying here. I'm trying to load Cineform files.
A simple core.ffms2.Source() seems to return RGB48 for the Cineform files (reported as 12bpp RGB, so I guess it makes it 16bpp RGB because no support for 'RGB36'), but somehow still sets the format-tag to something YUV based which trips up the resizers in Vapoursynth which means I can't convert / resize it to anything.

edit:
Btw: (not to try to prove you wrong, but more to explain to me how it could be )
I'm in After Effects (which doesn't understand the concept of YUV, it's completely RGB compositer), I make a little comp, export it yo ProRes 4444.
With ffms2 I load the file into Vapoursynth which returns it as '4444'. I convert it with fmtconv to RGB _with specifying the bt709_ color matrix. The end result in RGB is _exactly_ the same as my After Effects start. So the file seems to be bt709.
Ffmpeg had a time (or still does?) where it writes incorrect bt601 matrix-tags in the output file for ProRes, but those are ignored (in FCPX and the Adobe suite at least) anyway.

As another example, I can load a .MTS AVCHD file from one my cameras, which opens as YUV420P8. I use fmtconv to 'convert' it to YUV422P10. That YUV422P10 I pass to ffmpeg to encode to prores.
When I then load the original .MTS file and my ffmpeg-generated .mov file in After Effects and toggle the layers on and off, I see no differences. Which to me also seems to suggest the Prores data is bt709 and being interpreted as bt709. (The AVCHD .mts file is bt709 for sure).

Last edited by dipje; 4th January 2017 at 17:54.
dipje is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 09:51.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.