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 > Hardware & Software > Software players

Reply
 
Thread Tools Search this Thread Display Modes
Old 21st October 2011, 18:05   #10321  |  Link
pirlouy
_
 
Join Date: May 2008
Location: France
Posts: 553
Quote:
Originally Posted by nevcairiel View Post
Maybe i feel like venturing into that land some day when i have the "major" blockers in LAV Filters taken care of.
I like a challenge, and to learn something new, and this would fit the bill perfectly.
Start today please !
__________________
Windows 10 x64; 1080p 60" TV, GTX 750; madVR, LAV Filters, MPC-HC
pirlouy is offline   Reply With Quote
Old 21st October 2011, 18:09   #10322  |  Link
nand chan
( ≖‿≖)
 
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
I think SoftCubic is just too blurry to use properly. When you watch a lot of Anime, the sharpness matters 100x as much. The blurring makes anime basically unwatchable.
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management
nand chan is offline   Reply With Quote
Old 21st October 2011, 18:20   #10323  |  Link
Budtz
Registered User
 
Join Date: Apr 2011
Posts: 130
I have copied settings from hdtvtest. These are calibrated settings. Witch options should i choose in madvr when i have a calibrated set?
Budtz is offline   Reply With Quote
Old 21st October 2011, 18:25   #10324  |  Link
markanini
Registered User
 
Join Date: Apr 2006
Posts: 293
'Scale in linear light' looks great on the right sources, bringing out faint details. Oversharpened sources might need a softer kernel though.
markanini is offline   Reply With Quote
Old 21st October 2011, 18:52   #10325  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 8,757
Quote:
Originally Posted by nand chan View Post
Ps. madshi why aren't you on IRC?
Why would I be?

Quote:
Originally Posted by TheShadowRunner View Post
Unfortunately, no easy way to recreate the issue.
That's too bad, it makes it very hard for me to fix the problem.

Quote:
Originally Posted by TheShadowRunner View Post
Btw, do you still plan to enhance madVR's internal frequency switcher taking media height into account?
Remind me, what was this about?

Quote:
Originally Posted by dansrfe View Post
Just wanted to make sure saving frames to image format (png/jpg) is still on your list of things to do. It's a pretty handy feature... Thanks.
Sure, it's somewhere on my list.

Quote:
Originally Posted by Budtz View Post
Seems for a long time there has been a small bug where madvr ether crashes or exclusivemode fails. This happens if I switch to the next file in the folder in exclusivemode, but only if I have the mouse at the bottom of the screen so the search bar is showing. If there is no search bar then it works just fine switching video files. I have a setup where I press page-up or down to switch to the next video. It is when I press these buttons, it happens.
Which media player are we talking about? Can anybody else reproduce this?

Quote:
Originally Posted by Boltron View Post
The latest madVR has additional info with ^J. Among other things I see a "limited range (says upstream/bitstream)". Is this referring to RGB output levels 16-235 v.s. 0-255?
Quote:
Originally Posted by nand chan View Post
No, it's referring to the input levels. The output levels are always what you specified for your display in the settings.
Correct.

Quote:
Originally Posted by nand chan View Post
about the “upstream” vs “bitstream” issue I'm guessing upstream means the filter guessed it, and bitstream means it's written into the file headers itself. Either that, or it's the other way around.
Oh, I should have explained that. There are 3 new items in the Ctrl+J OSD:

(1) "matrix" - this is the source's decoding matrix
(2) "primaries" - this is the source's gamut/primaries
(3) "full/limited range" - this is the source's range

madVR tries to autodetect the correct values. The OSD shows in brackets how madVR autodetected the specific value. There are various different ways that madVR can autodetect these things (sorted by priority):

(a) "says ffdshow" - madVR has detected that ffdshow is sending RGB and madVR has found out which range/levels ffdshow is sending
(b) "says upstream" - the upstream filter has officially informed madVR
(c) "says bitstream" - madVR has read the info from the compressed video bitstream (h264, MPEG2 or VC-1)
(d) "best guess" - madVR has guessed, based on resolution (matrix/primaries) or color space (range)

Quote:
Originally Posted by jmone View Post
Once it happens it tends to stay till the player is restarted. Is it possible to enable logging between plays and would this give you what you need. This is one that a few have commented on (incl nevcairiel).
I don't really know if I can fix this without being able to reliably reproduce it.

Quote:
Originally Posted by nevcairiel View Post
The way the subtitle interface works right now is simple. You get the surface with the video image from the renderer, and you paint your subtitles on there. The renderer doesn't have to do anything.

libass is a rendering library, it just provides you with the info where to paint which pixel in which RGBA color (including alpha).
I'm wondering right now, if you did create a new external subtitle renderer, how much of your development time would go into:

(1) creating and maintaining the DShow framework
(2) finding the best way to forward the subtitles to the renderers
(3) actual work on implementing libass etc

I'm thinking right now that you'd probably spend 90%+ of your time on (1) and (2), would you agree with that? If so, I'm wondering whether it really makes sense. If I would implement libass directly in madVR (without support for other renderers, of course), I could probably get along with 20% of the time you'd need to invest. Or do you think my estimates are off?

Quote:
Originally Posted by 6233638 View Post
I took the time to measure the results I get with madVR and the built-in yCMS calibration, when giving it accurate data. (no hand-tweaking to try and control the results)

On the left are the automatically calculated madVR/yCMS results, on the right are hand-tweaked values (outside of madVR) to show what the display is capable of.
Ouch. I would say that yesgrey has some work to do to get this right. I'm convinced that this is a problem with yCMS, though, and not a problem with the 3dlut technique.

Have you tried nand chan's tools? Doesn't he have a "competeting" solution to yCMS? Maybe his tools create better 3dlut files?

Quote:
Originally Posted by 6233638 View Post
I'd go as far as saying that the option probably shouldn't even be available with anything other than the higher levels of SoftCubic when upscaling, and disabled on Lanczos/Spline when downscaling.
Yeah, I have my doubts about the use of linear light scaling, too. It's just that in another thread a user claimed that I would be taking shortcuts because I'm not doing linear light scaling. So I added it now, but I don't think it's all that useful, except maybe for downscaling. That said, maybe it will be more useful when I implement some better scaling algorithms.

Quote:
Originally Posted by 6233638 View Post
And as a side-note, it would be really nice if selecting madVR from the filters list brought up the preferences right away as it used to do, rather than it being a two-step process. It really slows things down when trying to do a comparison like this.
That is technically not possible, because the new madVR settings dialog is not created by madVR.ax, anymore, but by the external madHcCtrl.exe process. That said, I'm usually changing settings through the madVR tray icon menu. That works faster for me.
madshi is offline   Reply With Quote
Old 21st October 2011, 18:57   #10326  |  Link
Budtz
Registered User
 
Join Date: Apr 2011
Posts: 130
The Mediaplayer i use is MPC-HC.
Budtz is offline   Reply With Quote
Old 21st October 2011, 19:01   #10327  |  Link
nevcairiel
Registered Developer
 
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 8,746
Quote:
Originally Posted by madshi View Post
I'm wondering right now, if you did create a new external subtitle renderer, how much of your development time would go into:

(1) creating and maintaining the DShow framework
(2) finding the best way to forward the subtitles to the renderers
(3) actual work on implementing libass etc

I'm thinking right now that you'd probably spend 90%+ of your time on (1) and (2), would you agree with that? If so, I'm wondering whether it really makes sense. If I would implement libass directly in madVR (without support for other renderers, of course), I could probably get along with 20% of the time you'd need to invest. Or do you think my estimates are off?
1) will certainly require some work, but i don't think its that much, since i already have some good basis and experience to start with from my other decoders.
2) Is something actually wrong with the subtitle interface you use for MPC-HC? I thought all the issues we are having were due to bugs in the sub renderer, not in the interface.
3) Will certainly be fun!

Like mentioned before, one blocking factor for me is the time spend trying to figure out Direct3D, as i haven't done much with it.
I would probably add it to 3), but i would take longer with it then you would, for example, as you have the D3D experience.

I just think that the benefit of being able to use it with other renderers (and possibly without renderer support as a last resort, painting on the image, or for transcoding) outweights the extra efforts in this case, because the only alternative is VSFilter or ffdshow (which really are based on the same code), and those projects are really long dead, and unlikely to receive much love.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders

Last edited by nevcairiel; 21st October 2011 at 19:07. Reason: typos
nevcairiel is offline   Reply With Quote
Old 21st October 2011, 19:13   #10328  |  Link
nevcairiel
Registered Developer
 
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 8,746
Quote:
Originally Posted by madshi View Post
I don't really know if I can fix this without being able to reliably reproduce it.
This is Re: "Exclusive Mode Failed"

I can't "reliably" reproduce it, but i can reproduce it.
What i usually do:

Watch a file, let it run to the end (or seek close to the end). MPC-HC goes to Window mode.
I click my mouse thumb button, which goes to the next file in the folder, playback resumes.
Go back to FSE mode.

Now repeat that procedure above until it fails. *Maybe* there are some other circumstances, i'll try some things and see if i can make it break more reliably.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders
nevcairiel is offline   Reply With Quote
Old 21st October 2011, 19:58   #10329  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 8,757
Quote:
Originally Posted by nevcairiel View Post
2) Is something actually wrong with the subtitle interface you use for MPC-HC? I thought all the issues we are having were due to bugs in the sub renderer, not in the interface.
Well, the interface is working reasonably well, but there are a couple of things I don't like:

(1) the D3D work is done by the subtitle renderer, I'd prefer to get simple bitmaps with an alpha channel and do the drawing myself; e.g. I would run into big trouble if I wanted to use a full D3D11 rendering path, that simply wouldn't work with the current subtitle interface
(2) alpha blending on 16bit textures doesn't work with older GPUs, making it hard for me to have the subtitles run through 3dluts
(3) I'm not sure the current interface will work for 3D content

I think if you create an external filter, it would be better to create a new interface, which would just forward the subtitles in bitmap form with an alpha channel. That way you also wouldn't have to do any D3D stuff. At least not for madVR.

Quote:
Originally Posted by nevcairiel View Post
*Maybe* there are some other circumstances, i'll try some things and see if i can make it break more reliably.
That would be great.
madshi is offline   Reply With Quote
Old 21st October 2011, 20:00   #10330  |  Link
nevcairiel
Registered Developer
 
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 8,746
Quote:
Originally Posted by madshi View Post
I think if you create an external filter, it would be better to create a new interface, which would just forward the subtitles in bitmap form with an alpha channel. That way you also wouldn't have to do any D3D stuff. At least not for madVR.
I suppose your reasoning makes sense. I would just create a RGBA image in subtitle rendering resolution, and let you deal with scaling and alpha blending.
I'll think some about it, but i don't have any immediate plans eitherway.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders

Last edited by nevcairiel; 21st October 2011 at 20:18.
nevcairiel is offline   Reply With Quote
Old 21st October 2011, 20:45   #10331  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 8,757
Quote:
Originally Posted by nevcairiel View Post
I would just create a RGBA image in subtitle rendering resolution, and let you deal with scaling and alpha blending.
Alpha blending: Yes. Scaling? Not sure. If you render in SD and ask me to scale to HD then the subtitles will look very blurry and low-res. It'd be better if the subtitles were rendered at target resolution. Then scaling should not be necessary.
madshi is offline   Reply With Quote
Old 21st October 2011, 20:53   #10332  |  Link
nevcairiel
Registered Developer
 
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 8,746
For performance reasons people might want to render in lower resolutions, or maybe a bitmap subtitle just is lower resolution. I can scale it as well, but it would be equally blurry.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders
nevcairiel is offline   Reply With Quote
Old 21st October 2011, 21:05   #10333  |  Link
HeadlessCow
Registered User
 
Join Date: Nov 2002
Posts: 131
Quote:
Originally Posted by madshi View Post
Ouch. How can a compiler treat arr[0,0] as arr[0]? That doesn't make *any* sense to me whatsoever.
It's used for putting two statements into a spot that can only syntactically hold one.

For loops are the only place I ever see it used.

for(int i = 0, j = 10; i < 10; i++, j--) {
...
}
HeadlessCow is offline   Reply With Quote
Old 21st October 2011, 21:12   #10334  |  Link
pirlouy
_
 
Join Date: May 2008
Location: France
Posts: 553
Why would you want to scale subtitles ?
People choose something like "Arial, 20", then you display the subtitle always in size "20", you put it in an image for madVR, and then no blur.

But I guess there's some stuff I don't understand at all, then sorry.
__________________
Windows 10 x64; 1080p 60" TV, GTX 750; madVR, LAV Filters, MPC-HC
pirlouy is offline   Reply With Quote
Old 21st October 2011, 21:19   #10335  |  Link
nand chan
( ≖‿≖)
 
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
Quote:
Originally Posted by madshi View Post
Alpha blending: Yes. Scaling? Not sure. If you render in SD and ask me to scale to HD then the subtitles will look very blurry and low-res. It'd be better if the subtitles were rendered at target resolution. Then scaling should not be necessary.
Three factors here:

1. Performance as mentioned
2. 3D transformations and such will simply look wrong if not rendered at the correct resolution, or so I've been told
3. In some cases, blurry subtitles might be desired because they would fit into the scenery better (eg. softsubbed signs), whereas a noticeably too-HD sign would stand out too much.

Quote:
Originally Posted by madshi View Post
Why would I be?
It makes communication like this a great deal faster when all involved parties are present.

Quote:
Have you tried nand chan's tools? Doesn't he have a "competeting" solution to yCMS? Maybe his tools create better 3dlut files?
Unfortunately my tools are not capable of calibration, not by themselves nor do I have any plans to include such functionality. I rely on existing tools such as LittleCMS, yCMS, ArgyllCMS and ICC profiles to generate a 3dlut. It's just a “unification” engine, so to say - it unifies the color pipeline into a well-defined 64-bit space and allows you to mix any sorts of different stuff (.cal files, .3dluts at every bit depth imaginable, .icc profiles, custom logic) into the same pipeline then output that.

It is not intended for nor capable of any form of measurement-based display calibration.
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management

Last edited by nand chan; 21st October 2011 at 21:23.
nand chan is offline   Reply With Quote
Old 21st October 2011, 21:27   #10336  |  Link
HeadlessCow
Registered User
 
Join Date: Nov 2002
Posts: 131
Quote:
Originally Posted by pirlouy View Post
Why would you want to scale subtitles ?
People choose something like "Arial, 20", then you display the subtitle always in size "20", you put it in an image for madVR, and then no blur.

But I guess there's some stuff I don't understand at all, then sorry.
Arial, 20 on a windowed 640x360 image might look fine, but once you upscale it (like going to full screen) you now have blurry subtitles that are 2 inches tall.
HeadlessCow is offline   Reply With Quote
Old 21st October 2011, 21:34   #10337  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 8,757
Quote:
Originally Posted by nevcairiel View Post
For performance reasons people might want to render in lower resolutions, or maybe a bitmap subtitle just is lower resolution. I can scale it as well, but it would be equally blurry.
True.

Quote:
Originally Posted by nand chan View Post
Unfortunately my tools are not capable of calibration, not by themselves nor do I have any plans to include such functionality. I rely on existing tools such as LittleCMS, yCMS, ArgyllCMS and ICC profiles to generate a 3dlut. It's just a “unification” engine, so to say - it unifies the color pipeline into a well-defined 64-bit space and allows you to mix any sorts of different stuff (.cal files, .3dluts at every bit depth imaginable, .icc profiles, custom logic) into the same pipeline then output that.

It is not intended for nor capable of any form of measurement-based display calibration.
I thought your tools could be used to convert ICC profiles to 3dlut files? And I thought you implemented that with LittleCMS and not with yCMS. Guess I was wrong about that? So your something -> 3dlut conversions always end up using yCMS?
madshi is offline   Reply With Quote
Old 21st October 2011, 22:38   #10338  |  Link
pirlouy
_
 
Join Date: May 2008
Location: France
Posts: 553
Quote:
Originally Posted by HeadlessCow View Post
Arial, 20 on a windowed 640x360 image might look fine, but once you upscale it (like going to full screen) you now have blurry subtitles that are 2 inches tall.
Why upscale it ?
Let it at 20 in window mode, and 20 in Full Screen.

And if it's possible for the subtitle renderer to know if we're in Full Screen, then we can choose a size for window mode, and a size for Full Screen; like this, no upscale/downscale.

Any upscale/downscale will be uglier than a defined size.
__________________
Windows 10 x64; 1080p 60" TV, GTX 750; madVR, LAV Filters, MPC-HC
pirlouy is offline   Reply With Quote
Old 21st October 2011, 22:51   #10339  |  Link
nand chan
( ≖‿≖)
 
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
Quote:
Originally Posted by madshi View Post
I thought your tools could be used to convert ICC profiles to 3dlut files? And I thought you implemented that with LittleCMS and not with yCMS. Guess I was wrong about that? So your something -> 3dlut conversions always end up using yCMS?
You're misunderstanding. The ICC profile in question here already exists, thus it's technically not a device calibration done by any part of my tools. You are correct, I use LittleCMS and not yCMS (yCMS is only used for generating .3dluts from .ti3 files since I haven't added Argyll support yet, but if anybody would be interested in continued development of the project that would be near the top of my todo list).

They can be used to convert ICC profiles to .3dlut files, but in this case, he doesn't even have an ICC profile - he's looking for software that will generate an ICC profile (either device or link) to his requirements. My software is not capable of generating ICC profiles whatsoever, just using them.
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management
nand chan is offline   Reply With Quote
Old 22nd October 2011, 00:41   #10340  |  Link
TheShadowRunner
Registered User
 
TheShadowRunner's Avatar
 
Join Date: Feb 2004
Posts: 396
Quote:
Originally Posted by madshi View Post
That's too bad, it makes it very hard for me to fix the problem.
Yes sorry, wish I could make it easier for you ^^;
So the address reported when the crash occurs doesn't help at all I guess.
I think for now let's wait for Blight's view on the issue, from my testing the bug is due to ZP displaying its own OSD/Control bar while a frequency switch occurs. If I force ZP to never show OSD/Control bar, the bug never shows up..
What led me to believe it was a madVR issue is that it doesn't happen with the other renderers. Tough one!

Quote:
Originally Posted by madshi View Post
Remind me, what was this about?
Simply to make madVR's internal frequency switcher as flexible as ReClock's so it's finally possible to ditch ReClock's RunEvent completely (no more need to switch freq. behind madVR's back!)
Currently madVR's switcher only takes framerate into account. The request is about having madVR's switcher also take media height into account. You seemed to be OK with it. ^^;

A related question: madVR doesn't currently "know" which frequency is used, does it?
Because I noticed a strange thing. If in the "list all display modes madVR may switch to" I input only "720p50" (omitting purposefully 720p60), and play a 29.97fps media while on 720p60, madVR switches to 720p50!?
It seemed strange to me. Of course it doesn't happen if I input 720p60 in the list too.
__________________
XP SP3 / Geforce 8500 / Zoom Player
TheShadowRunner is offline   Reply With Quote
Reply

Tags
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, upsampling

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 14:31.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2017, vBulletin Solutions, Inc.