madVR - high quality video renderer (GPU assisted)

[Motenai Yoda]

Does this also happen with the latest madVR v0.84.3 build? Are you using the MPC-HC refresh rate switcher? Try using the madVR built in display mode switcher instead.

ok
without MPC-HC internal switcher no switch.
with only MPC-HC internal switcher seems work.
with MPC-HC internal switcher and madVR doesn't work.

if i leave empty the modes textbox, work (with MPC-HC switcher)
with even 1024x768p60 (default for my screen) blank screen.

I've tryed even older build of madVR and MPC-HC.

my specs are w7 64bit, i7 920, hd4650, catalyst 12.6

[madshi]

ok
without MPC-HC internal switcher no switch.
with only MPC-HC internal switcher seems work.
with MPC-HC internal switcher and madVR doesn't work.

if i leave empty the modes textbox, work (with MPC-HC switcher)
with even 1024x768p60 (default for my screen) blank screen.

I've tryed even older build of madVR and MPC-HC.

my specs are w7 64bit, i7 920, hd4650, catalyst 12.6
Maybe there's a misunderstanding. You should not use 2 different display mode switchers at the same time! That won't work. In the worst case MPC-HC will switch to mode A, madVR's switcher will detect the change and switch to mode B. MPC-HC will switch back to mode A, madVR back to mode B. They might change modes all the time without pause. Please always only use one display mode switcher. Either MPC-HC or madVR, but not both switchers at the same time.

Mathias:
Having double checked, it would appear to me that having a final Gaussian blur step with a radius dependent on the enlargement ratio may be the ticket. Gremlins die before the blur is annoying. Gaussian blur with a reasonably small radius should be cheap (enough).
My guess is that the radius (or, equivalently, sigma) should obey a simple proportionality constant with the enlargement ratio. (Gaussian radius = half of enlargement ratio, I would guess, is about right. This means that the "effective area" over which the blur is noticeable is about the size of an original pixel, or a bit smaller.)
Thank you, that sounds like a good fix. Maybe I'll find a way to avoid the gremlins from appearing in the first place. If not, I'll try your blur idea. :)

[6233638]

Thanks... :) I'm still waiting for more test pictures from 6233638 and cyberbeing, though. Optimizing the AR algorithm for just one image is easy. Optimizing it to work well for all sorts of difficult situations is the hard thing.What sort of images are you looking for? Film captures?

The problem I have is that it's difficult to find video/film sources that don't already have a ton of ringing already encoded outside of Blu-ray. (which mostly doesn't need upscaling yet)

[aufkrawall]

I hate making the settings dialog more complicated than necessary. :( I'll think about it.

Thanks. My humble opinion: Checkboxes don't do any harm. :D
Of course I don't know if this is technically a little Sisyphean task-like for you to mess with the GUI, but from a user's viewpoint, a little checkbox, hidden relatively deep in the settings, won't make it more confusing.

Btw: I'd really appreciate if you could comment on this:
http://forum.doom9.org/showpost.php?p=1594957&postcount=14538

Thanks.

[NicolasRobidoux]

Mathias:
It indeed was an ImageMagick 7 bug. (IM7 is the alpha of the next major, API breaking, release, so I asked for it.)

[madshi]

What sort of images are you looking for? Film captures?
Well, you reported that the Jinc AR produces more (or uglier) ringing compared to e.g. Lanczos/Spline AR. I need more images which demonstrate that. Basically any images which make Jinc AR stumble somehow. And the more different the test images are the better. Thanks.

from a user's viewpoint, a little checkbox, hidden relatively deep in the settings, won't make it more confusing.
I have a very different opinion on this. A new madVR user will be tempted to look at all the settings and understand them. Adding billions of funny checkboxes which are hard to understand will bury the really important settings in a sea of noise. If I had added every funny option any madVR user ever asked for, the settings dialog would have 2 or 3 times as many options as it has now, many of which nobody would ever use or even understand. Don't get me wrong. I'm not saying that your wish is useless. I said I would think about it. I'm just saying that I generally do have to resist the tendency to overload the madVR settings dialog with funny options.

Btw: I'd really appreciate if you could comment on this:
http://forum.doom9.org/showpost.php?p=1594957&postcount=14538
DVD playback is not officially supported by madVR yet. I can't fix things that I don't support and which require hacking the OS for me to even be able to reproduce the problem.

[huhn]

i can confirm the dpc problems with my amd 6770 for me they are at 1300 and i didn't notice a problem at all at least nothing i'm aware off.
using a queue like 4/4 solved the "issue" but more interesting is the fact that my hd4000 in the same pc doesn't have any problems with queues like 24/24l the 6770 is still running but all displays are disconnected.
even with 4/4 the error comes back after some time
with evr cp there r no problem at all with dpc spikes.

bug report:

start a video in windows mode for example a 720 on a 1080p screen
press pause then stop and now press control+enter

the video will be at the top left in original size the rest is black
i can't test it in windows 7 overlay mode because my amd card can't handle it
in normal and fse mode it exists
i think someone else reported this before not sure

sry for the terrible English

[Blight]

madshi:
I addressed the over-complication of the settings dialog (and ZP has a lot of them) in three ways:
1. Split the options into a basic/advanced configuration.
2. Add a centralized documentation, in your case, maybe a wiki would be best.
3. Probably overkill in your situation, but I've also added a search function that scans through the dialogs and highlights the searched item.

I'm not really sure if you have the time for it, but it would simplify some of your interface design decision making.

[madshi]

bug report:

start a video in windows mode for example a 720 on a 1080p screen
press pause then stop and now press control+enter

the video will be at the top left in original size the rest is black
i can't test it in windows 7 overlay mode because my amd card can't handle it
in normal and fse mode it exists
i think someone else reported this before not sure
Which media player? Which OS? Aero enabled or disabled?

madshi:
I addressed the over-complication of the settings dialog (and ZP has a lot of them) in three ways:
1. Split the options into a basic/advanced configuration.
2. Add a centralized documentation, in your case, maybe a wiki would be best.
3. Probably overkill in your situation, but I've also added a search function that scans through the dialogs and highlights the searched item.

I'm not really sure if you have the time for it, but it would simplify some of your interface design decision making.
Yes, both a basic/advanced switch and a settings documention would be a good thing to have, but both would be a lot of work, and it's not the right time for that now, considering that madVR is still lacking a few key features.

[aufkrawall]

I have a very different opinion on this. A new madVR user will be tempted to look at all the settings and understand them. Adding billions of funny checkboxes which are hard to understand will bury the really important settings in a sea of noise. If I had added every funny option any madVR user ever asked for, the settings dialog would have 2 or 3 times as many options as it has now, many of which nobody would ever use or even understand. Don't get me wrong. I'm not saying that your wish is useless. I said I would think about it. I'm just saying that I generally do have to resist the tendency to overload the madVR settings dialog with funny options.

I agree that an overload of settings should be avoided, but regarding this issue I think it's different.
What is the filename mode detection for? I'd assume to fix wrong flags? That's a good idea, but imho not if it's the default because flags/file headers should be regarded more important than file names since latter can be messed up too easily.
It's like if an anti-virus would catch malware by file names. :D
It's a good thing though if it's optional.


DVD playback is not officially supported by madVR yet. I can't fix things that I don't support and which require hacking the OS for me to even be able to reproduce the problem.
I totally see your point.

DVDs would profit so much from madVR. :(

[NicolasRobidoux]

Mathias:
Suggestion:
Because you are effectively dealing with the most offensive common problem with a linear light toolchain, namely dark halos (your'e dealing with the light ones too, of course, but they are not the main issue with linear light resampling that uses negative lobe filters), and you are starting to do rather complicated things in the AR toolchain, including Gaussian blur, I suggest that you try your favorite AR toolchain through linear light (into linear light before the resampling filter, out of linear light after the final Gaussian blur) because this will restore, among other things, light <-> dark symmetry, and in particular would prevent the most annoying artifact of sRGB resampling (besides light halos), which is the "thinning" of light on dark features, which AR, I'm sure, can't do anything about.
I have the foolish hope that this may reduce the size of the worst gremlins, as well. (This being said, I still have not read your overall description of how AR works, so everything is based on "I bet it works more or less like this...")
Also, unless you have "fudge factors" in your AR algorithm, I also suggest that you see if you can see a difference when you use a LUT which has more points than what you use now.
-----
When you have time to waste, I suggest that you try EWA Robidoux. Not perfect. But good-cheap-solid.
P.S. "The Recommendations" are in need of an update. Will take a while. Don't take what's there for my current gospel.

[huhn]

Which media player? Which OS? Aero enabled or disabled?

os win xp 64 win 7 64 win 8 64
mpc hc atm 6086

aero tested with both on and off
double checked on win 7 64

[crotecun]

I agree that an overload of settings should be avoided, but regarding this issue I think it's different.
What is the filename mode detection for? I'd assume to fix wrong flags? That's a good idea, but imho not if it's the default because flags/file headers should be regarded more important than file names since latter can be messed up too easily.
It's like if an anti-virus would catch malware by file names. :D
It's a good thing though if it's optional.


From what I've read filename mode detection working that way is not a bug, it's a feature.

If you add such a tag to a file name, madVR thinks that you want to overwrite the frame rate to 60.000Hz. That's not a bug but intended behaviour. Change it to 1080p59 and the problem should go away.

You should tag it with the correct rate or perhaps leave it from the filename altogether to disable the automatic detection.

[NicolasRobidoux]

Also, unless you have "fudge factors" in your AR algorithm, I also suggest that you see if you can see a difference when you use a LUT which has more points than what you use now.Clarification: Use a larger LUT for the (EWA) filter itself. Or else, build something that takes into account that there is an epsilon error in the computed values fed to AR.
For example, at locations where there is some ambiguity RE: whether to "chop", blend instead.
AR looks overly eager to me. You can leave some halo. The first halo, in particular, can be argued to increase acutance, so it's not all bad.
Sometimes, "better" is better than "perfect".

[Pat357]

What is the filename mode detection for? I'd assume to fix wrong flags? That's a good idea, but imho not if it's the default because flags/file headers should be regarded more important than file names (
If the flags would have priority, then how would you ever be able to properly playback a file with bad flags ? Not possible !!
With the current system, you add the correct info to the filename and your set : perfect playback !

[aufkrawall]

If the flags would have priority, then how would you ever be able to properly playback a file with bad flags ?

Well, encoders/muxers usually flag correctly, or not?
In the end, filename detection is a somewhat "hacky" solution, especially if you have no influence on it (hidden magic that has confused you and me).


Not possible !!

With the current system, you add the correct info to the filename and your set : perfect playback !
That's why I like a checkbox.

[nevcairiel]

In the end, filename detection is a somewhat "hacky" solution, especially if you have no influence on it (hidden magic that has confused you and me).

You have all the influence on it by changing the filename. :)

[NicolasRobidoux]

To keep noobs out of trouble and the interface clean, have some options only accessible through config files. Emphasis on "only".
Then, your documentation can have a section called "Expert Configuration Options" with a blinking neon sign that says: "You are likely to break things if you mess with this stuff, which caters to the "needs" of the minority that eats bandwidth for breakfast, lunch and dinner on madVR forums. Not to forget the midnight snack."

[TheLion]

i can confirm the dpc problems with my amd 6770 for me they are at 1300 and i didn't notice a problem at all at least nothing i'm aware off.
using a queue like 4/4 solved the "issue" but more interesting is the fact that my hd4000 in the same pc doesn't have any problems with queues like 24/24l the 6770 is still running but all displays are disconnected.
even with 4/4 the error comes back after some time
with evr cp there r no problem at all with dpc spikes.



Yeah, there definitely seems to be a problem with DPC spikes with the combination of AMD/ATI 5xxx/6xxx series and madVR. After my last post I also ran into these issues again. With 8/4 queue sizes and 8 frames in advance it was quite stable but after playback runs a while these dpc spikes crept in again. This still leads to occasional audio drops (no matter how large I set the ASIO buffer). The only way to avoid this is using EVR - sadly.

[aufkrawall]

You have all the influence on it by changing the filename. :)
If you know that it works that way, but it's nowhere indicated.

[MSL_DK]

I am about to calibrate my TV/htpc. I know this is not quite right to do through mpc/madVR but I do not own a blu-ray player... yet.

When I calibrate through mpc/madVR would it be most optimal to choose 'this display is already calibrated' choose BT.709 select pure power curve 2.20 and 'disable GPU gamma ramps'?

I calibrate using X-Rite i1Display Pro, ColorHCFR and AVS HD 709 - Blu-ray & MP4 Calibration

[crotecun]

If you know that it works that way, but it's nowhere indicated.

That's more like a documentation problem than a configuration problem.

If madVR gets a checkbox to disable the file tokens it would be nigh useless since most users would not know about the filename tokens in the first place and they wouldn't check the option. But if it is checked by default, then madVR would lose a fair bit of functionality since the filename tokens would be disabled.

Basically by the time you know about the option to disable the filename tokens then you would also know that you could have done the same by just renaming the files yourself.

It's better to document this function thoroughly. If it hasn't been done already, perhaps a status in the on screen display whenever madVR follows a filename token to set the frame rate?

[6233638]

Well, you reported that the Jinc AR produces more (or uglier) ringing compared to e.g. Lanczos/Spline AR. I need more images which demonstrate that. Basically any images which make Jinc AR stumble somehow. And the more different the test images are the better. Thanks.I will try and find some more samples.

As I said, it's difficult outside of artificial tests, or recorded game footage. (most of the game footage I use has gone through an analogue video chain and is quite compressed, so it's not the same as an "artificial" screenshot) Most other content that needs scaled up already has a ton of ringing, and I find scalers like Jinc/Spline far too sharp for it.

I've had a quick look over some "problem discs" and so far it seems that with most real-world DVD content, Jinc is on-par or better than Spline as far as ringing is concerned with the anti-ringing filter enabled, but much better on aliasing. (as expected) I still find it far too sharp for SD content, and tend to go back to SoftCubic 80 though. Separate HD/SD scaling options would be nice, as I quite like Jinc 3 anti-ring for HD, and SoftCubic 80 for SD.
DVD playback is not officially supported by madVR yet. I can't fix things that I don't support and which require hacking the OS for me to even be able to reproduce the problem.The DVD problems don't require OS hacks to show up - as soon as I try to play a DVD in MPC-HC with madVR, I get Macrovision errors, whether it's ffdshow or LAV Filters doing the decoding, and this is with AnyDVD running.

The "OS hacks" are required to try and get DVD playback working in madVR, and now only require changing a registry key to point it to a different DVD Navigator instead of replacing system files.

If you get this package (http://www.filedropper.com/dvdnavigator), and use the executable to point it to the qdvd.dll contained within (Windows XP's DVD Navigator) DVD playback works, but it has a tendency to crash on seeking (video keeps playing, but MPC-HC is unresponsive) and menus don't always work correctly.

There was a brief period where you could get Windows 7's DVD Navigator working with specific builds of LAV Filters, but Nevcariel doesn't know what it was that enabled them to work, and he apparently has the Windows 7 DVD Navigator working correctly on his system in all current builds of LAV Filters. (but I can't see how - I only have madVR and LAV Filters running in my tests on a fresh install of Windows)

[nevcairiel]

The DVD problems don't require OS hacks to show up - as soon as I try to play a DVD in MPC-HC with madVR, I get Macrovision errors, whether it's ffdshow or LAV Filters doing the decoding, and this is with AnyDVD running.


He wasn't answering about Macrovision errors, but playback errors/player lockups when using the "hack", which someone inquired about.

[Pat357]

Well, encoders/muxers usually flag correctly, or not? .
No, the muxer/encoder often take whatever the user inputs : bullshit in -> bullshit out.

On playback the demuxers/splitters have (often) no choice then reading the flags from the file ; there is no way for the splitter to know whether these are correct for this file are not.
Some things like FPS can be "measured" for a few frames ; the splitter can use these as an "approximation" for the real FPS, without being sure of it;
This all makes that playback will look very bad : playing to fast/slow, shakey, combing, 3/2 jitter + not smooth, blurry, wrong colors, artifacts from de-interlacing progressive content , ....

[iSunrise]

The DVD problems don't require OS hacks to show up - as soon as I try to play a DVD in MPC-HC with madVR, I get Macrovision errors, whether it's ffdshow or LAV Filters doing the decoding, and this is with AnyDVD running.

The "OS hacks" are required to try and get DVD playback working in madVR, and now only require changing a registry key to point it to a different DVD Navigator instead of replacing system files.

If you get this package (http://www.filedropper.com/dvdnavigator), and use the executable to point it to the qdvd.dll contained within (Windows XP's DVD Navigator) DVD playback works, but it has a tendency to crash on seeking (video keeps playing, but MPC-HC is unresponsive) and menus don't always work correctly.

There was a brief period where you could get Windows 7's DVD Navigator working with specific builds of LAV Filters, but Nevcariel doesn't know what it was that enabled them to work, and he apparently has the Windows 7 DVD Navigator working correctly on his system in all current builds of LAV Filters. (but I can't see how - I only have madVR and LAV Filters running in my tests on a fresh install of Windows)
Did you already try nevcariel´s latest DVD betas? I also had problems with the standard DVD Navigator filter in Windows 7 with older LAV builds. DVD playback works perfectly fine here now, without having to change anything in the OS itself. Just leave AnyDVD running in the background. LAV even shows DVD menus now, which is awesome.

PS: I´m on Windows 7 x64 Ultimate. Playback chain is only LAV and madVR (using PotPlayer).

@madshi:
What I would really find useful is to have like 2 different upscaling setups, where either madVR can automatically choose between them (e.x. source: movie resolution) or/and the user can toggle (override) between them if needed. I think you already mentioned something like that in your last posts.

I usually prefer sharp algorithms, however, there is some content out there that just scales very badly to 1080p (or in the future even higher), where you need to think a bit differently. I love your work on Jinc, excited to see where you can take it.

[6233638]

Did you already try nevcariel´s latest DVD betas? I also had problems with the standard DVD Navigator filter in Windows 7 with older LAV builds. DVD playback works perfectly fine here now, without having to change anything in the OS itself. Just leave AnyDVD running in the background. LAV even shows DVD menus now, which is awesome.I have just tried installing the latest MPC-HC and LAV Filters builds from here (http://xhmikosr.1f0.de/) and it's still not working. (previously, I was using the builds with DVD playback enabled that were posted in the LAV Filters topic)

My filter list with DVDs is: ReClock Audio Renderer (I have tried changing audio renderer, it makes no difference)
madVR
LAV Audio Decoder
LAV Video Decoder
DVD Navigator
I get "Macrovision Fail" any time I try playing a DVD with the Windows 7 DVD Navigator. (Windows 7 Pro x64 SP1)

I just reinstalled Windows 7 a couple of weeks ago, after trying out Windows 8 (too many compatibility issues for me) so it's not like I have anything else installed on my system that should be causing it.

Are you running with an AMD or Nvidia card in your system? That's about the only other thing I can think it could be that is causing this to work fine on some systems, and not at all on others.

[strumf666]

You told me you were using "1080p23, 1080p24, 1080p50, 1080p60". Which was not true. You are actually using "1080p23, 1080p24, 1080p25, 1080p30, 1080p50, 1080p60". madVR consequently selects 1080p25 for 25Hz content, but unfortunately your GPU reports "bad mode" when trying to switch to 1080p25. You should only enter valid display modes into the madVR display switching field.
Apologies, I use projector and TV for output and I must have forgotten to delete the 25p from projector settings. Regarding the graphics card reporting it as bad, that's probably because my projector doesn't support it or do AMD cards have issues with it?
Thanks for your help :)

[dansrfe]

I have just tried installing the latest MPC-HC and LAV Filters builds from here (http://xhmikosr.1f0.de/) and it's still not working. (previously, I was using the builds with DVD playback enabled that were posted in the LAV Filters topic)

My filter list with DVDs is: ReClock Audio Renderer (I have tried changing audio renderer, it makes no difference)
madVR
LAV Audio Decoder
LAV Video Decoder
DVD Navigator
I get "Macrovision Fail" any time I try playing a DVD with the Windows 7 DVD Navigator. (Windows 7 Pro x64 SP1)

I just reinstalled Windows 7 a couple of weeks ago, after trying out Windows 8 (too many compatibility issues for me) so it's not like I have anything else installed on my system that should be causing it.

Are you running with an AMD or Nvidia card in your system? That's about the only other thing I can think it could be that is causing this to work fine on some systems, and not at all on others.

I have the DVD navigator working on both of my systems with ATI and nvidia. Did you apply the path hack that is in the dslibdvdnav thread?

[6233638]

While it's probably not what you were looking for, as it's still relatively artificial, here are a couple of samples that show ringing around the edge of subtitles that the current anti-ringing filter doesn't catch. (Munich (http://www.imdb.com/title/tt0408306/) DVD)

http://www.abload.de/thumb/screenshot179sap.png (http://www.abload.de/img/screenshot179sap.png) http://www.abload.de/thumb/screenshot280s67.png (http://www.abload.de/img/screenshot280s67.png)

I would say that Spline 3 AR exhibits slightly less ringing in these samples compared to Jinc 3 AR.

Mitchell-Netravali AR shows minimal ringing, and SoftCubic 80 shows negligible ringing with these images. Anything else is rings too much for my tastes. (but I'm pushing myself to use Jinc 3 to look for things like this)

Jinc 3 AR and SoftCubic 80 are the only resamplers that show good anti-aliasing properties on the diagonals. (but the source is very aliased)

As with most DVDs though, even if the ringing introduced by Jinc were eliminated, I think it is still too sharp to use due to the amount of ringing already encoded in the source.
I have the DVD navigator working on both of my systems with ATI and nvidia. Did you apply the path hack that is in the dslibdvdnav thread?I have that working fine, but other people here are reporting that the Windows 7 DVD Navigator is now working with the latest versions of LAV Filters and madVR, which I cannot get working at all. ("Macrovision Fail" error)

I can only get the XP DVD Navigator (qdvd.dll set using the tool provided with dslibdvdnav) or dslibdvdnav working with madVR, neither of which are great solutions. The XP DVD Navigator crashes on seeking 90% of the time (video plays, but the player becomes unresponsive) and dslibdvdnav doesn't work with a large number of DVDs that I own.

There was a brief period where an older version of LAV Filters worked with the Windows 7 DVD Navigator, and that played most of the discs I tried without issue, however that has now broken for me in updated versions of LAV Filters. (but there are new features/bugfixes that I need from them)


Apologies, I use projector and TV for output and I must have forgotten to delete the 25p from projector settings. Regarding the graphics card reporting it as bad, that's probably because my projector doesn't support it or do AMD cards have issues with it?
Thanks for your help :)I don't know of any display which accepts a 25Hz (25p) signal. With Nvidia, "25Hz" actually outputs 1080i50 if I remember correctly. (which is then deinterlaced to 1080p25)

[Pulstar]

Any idea why madVR switches to TV levels output with certain SD/AVC files, regardless of decoder? All monitors in madVR are set to PC levels.

[Peekstra]

Recently, I noticed some tearing with Madvr in the upper part of the image (mainly during horizontal movements):

http://imageshack.us/a/img443/5791/img20121015224837.th.jpg (http://imageshack.us/photo/my-images/443/img20121015224837.jpg/)

http://imageshack.us/a/img337/315/img20121015224749.th.jpg (http://imageshack.us/photo/my-images/337/img20121015224749.jpg/)

These (sorry) rather unclear photo's show it on my TV. I don't know if it happens with HD / no PAL material as well as I've been mainly watching SD (PAL) content for the last weeks which I don't do very often. As I don't very often watch PAL material its also possible that this problem already was present before the more recent Madvr builds although I'm inclined to think it has started happening with the more recent builds.

It's very difficult to reproduce because it can take hours before it happens. When I restart the player, the problem is gone.

The TV is set to the same refreshrate as the source material by Madvr and has a very low deviation (< 0.001%).

I'm using the latest Nvidia WHQL drivers and it happened with the previous one as well. The GPU is a GTX670 and the player is Zoomplayer 8.16

And I'm also using the HPET settings which were posted earlier in this thread.

Madvr is using all the default options for as far as I can tell.

[cyberbeing]

Crash with madVR v0.84.3 & MPC-HC ISR r6086
http://www.mediafire.com/?a9698iv9rucvr6r

Doesn't seem easily reproducible, so this could be a MPC-HC ISR bug and not a madVR bug. It crashed when the MPC-HC ISR was trying to display a line right after a chapter point. The line in question specified a missing font, so the MPC-HC ISR probably crashed when attempting to load or render the gdi fallback font.

[NicolasRobidoux]

While it's probably not what you were looking for, as it's still relatively artificial, here are a couple of samples that show ringing around the edge of subtitles that the current anti-ringing filter doesn't catch.IMHO, this is exactly the kind of image where EWA Quadratic B-spline smoothing shines. It's cheap (radius=1.5), it smooths out ringing, and it's not offensively blurry: http://web.cs.laurentian.ca/nrobidoux/misc/madVR/screenshot179sapEWAQuadratic.png
(http://web.cs.laurentian.ca/nrobidoux/misc/madVR/screenshot179sapEWAQuadratic.png)
On the other hand, I imagine that Mathias has something else to do than add such a "special purpose" filter: Unless the image is pixelated and/or full or ringing/halo almost to the point of being considered "defective", this filter is too soft, at least for luma.

[NicolasRobidoux]

I actually have programmed a relative of EWA quadratic B-spline smoothing, called VSQBS (Vertex Split with Quadratic B-Spline Smoothing) which is an unbelievably cheap scheme (my guess is would be screamingly fast on a GPU), in a different library than ImageMagick, namely VIPS (Virtual Image Processing System) and its GUI NIP2 (New Image Processor 2). It has slightly better jaggy reduction, at the cost of a touch more blur. http://web.cs.laurentian.ca/nrobidoux/misc/madVR/screenshot179sapVSQBS.png
(http://web.cs.laurentian.ca/nrobidoux/misc/madVR/screenshot179sapVSQBS.png)
VSQBS is not for downsampling. To get a similar scheme for downsampling, you should switch to EWA or tensor quadratic B-spline smoothing.

When enlarging, VSQBS uses the 8 input pixel values closest to the output pixel location to produce its value, and it's really easy to determine these 8.

(Note: Neither with ImageMagick nor VIPS did I do careful import and export of the png. This is why there is a slight difference in colouring in some browsers.)

P.S. Despite the reduced jaggies, I actually like EWA quadratic B-spline smoothing better than VSQBS.

[6233638]

IMHO, this is exactly the kind of image where EWA Quadratic B-spline smoothing shines. It's cheap (radius=1.5), it smooths out ringing, and it's not offensively blurry: http://web.cs.laurentian.ca/nrobidoux/misc/madVR/screenshot179sapEWAQuadratic.png
(http://web.cs.laurentian.ca/nrobidoux/misc/madVR/screenshot179sapEWAQuadratic.png)
On the other hand, I imagine that Mathias has something else to do than add such a "special purpose" filter: Unless the image is pixelated and/or full or ringing/halo almost to the point of being considered "defective", this filter is too soft, at least for luma.Soft filters are not necessarily bad for Luma. Unlike image processing where you might be upsampling a high quality photograph, the quality of commercially released DVDs is generally terrible. They have had all the high-frequency detail filtered out of them, and sharpened with very basic techniques that introduce a lot of ringing. Older discs also exhibit a lot of "mosquito" noise.

What I like about softer filters, is that they often blur out this offensive ringing/noise (note: I have no problem with film grain) while retaining most of the actual useful picture information.

This is perhaps going too far in the opposite direction, but SoftCubic 80 compared with Jinc 3 anti-ring:
http://www.abload.de/thumb/softcubici8rz3.jpg (http://www.abload.de/img/softcubici8rz3.jpg) http://www.abload.de/thumb/jinc55odk.jpg (http://www.abload.de/img/jinc55odk.jpg)

Source (http://www.abload.de/img/screenshotonpfh.png)

I actually have programmed a relative of EWA quadratic B-spline smoothing, called VSQBS (Vertex Split with Quadratic B-Spline Smoothing) which is an unbelievably cheap scheme (my guess is would be screamingly fast on a GPU), in a different library than ImageMagick, namely VIPS (Virtual Image Processing System) and its GUI NIP2 (New Image Processor 2). It has slightly better jaggy reduction, at the cost of a touch more blur.This looks pretty good. (along with your other example)

Recently, I noticed some tearing with Madvr in the upper part of the image (mainly during horizontal movements)
...

These (sorry) rather unclear photo's show it on my TV. I don't know if it happens with HD / no PAL material as well as I've been mainly watching SD (PAL) content for the last weeks which I don't do very often. As I don't very often watch PAL material its also possible that this problem already was present before the more recent Madvr builds although I'm inclined to think it has started happening with the more recent builds.
Do you have interpolation enabled on your TV? (motion smoothing) A number of displays can introduce their own screen tearing when using it. (I think just about all Sony TVs do with MotionFlow enabled for example) You should be able to take a screenshot that shows tearing, if it is not the display.
You are using an Nvidia graphics card, do you have "Adaptive V-Sync" enabled? This allows the GPU to tear.
You say that you don't often play PAL material, are you outputting it at 50Hz? (Or 24p with ReClock's PAL SpeedDown?)

[NicolasRobidoux]

On the Danny Glover shot, EWA quadratic B-spline smoothing has nothing over (tensor, I imagine) SoftCubic 80.
(Being sloppy, I can't quite align things using an ImageMagick command, so this is with the not-quite-aligned eyeball metric.)
P.S. I managed to get alignment. They're so close one could almost say they're clones. EWA Quadratic is a minuscule amount sharper and jaggier. I think SoftCubic wins by a whisker.
No "normal viewer" would see the difference, though. Even zooming in with a high quality image viewer (I use NIP2, of course: it's really fast, and I'm a dev).
P.S.2 Scrutinizing some more, SoftCubic 80 wins: less jaggies, without significant additional blur, and with very little additional halo.
P.S.3 But basically, you have to zoom in to break the tie.

[Peekstra]

You are using an Nvidia graphics card, do you have "Adaptive V-Sync" enabled? This allows the GPU to tear.
You say that you don't often play PAL material, are you outputting it at 50Hz? (Or 24p with ReClock's PAL SpeedDown?)[/list]

The TV is perfectly in sync at 50hz so that's not the problem.

While the TV does some smoothing but I don't think that's the culprit here. But....

It's probably the adaptive vsync option, it's enabled. I'll turn it off during playing a video and try if it happens again.

:thanks:

[yok833]

I have an random issue with all the version of Madvr... I have an ATI 6850 with a X4 925, 8 GB Ram and W7 64. The problem is after a certain period the screen is divided in 2 with a lot of artifacts.
Do you have an idea of the reason and how to fix??? At the moment the only solution is to reboot the system (close the session is not enough and the image is still splited in 2)

[Moony349]

My best guess is your video card is overheating.

[iSunrise]

Are you running with an AMD or Nvidia card in your system? That's about the only other thing I can think it could be that is causing this to work fine on some systems, and not at all on others.
I have a GTX580, currently still running on the 301.40 beta drivers. If it´s not because of the beta drivers, I must say that I am out of ideas, your setup seems to be very close to mine. If it´s not the graphics drivers, could this actually be caused by the soundcard´s drivers? I have a Creative X-Fi Titanium PCIe.

Soft filters are not necessarily bad for Luma. Unlike image processing where you might be upsampling a high quality photograph, the quality of commercially released DVDs is generally terrible. They have had all the high-frequency detail filtered out of them, and sharpened with very basic techniques that introduce a lot of ringing. Older discs also exhibit a lot of "mosquito" noise.

What I like about softer filters, is that they often blur out this offensive ringing/noise (note: I have no problem with film grain) while retaining most of the actual useful picture information.

This is perhaps going too far in the opposite direction, but SoftCubic 80 compared with Jinc 3 anti-ring:
http://www.abload.de/thumb/softcubici8rz3.jpg (http://www.abload.de/img/softcubici8rz3.jpg) http://www.abload.de/thumb/jinc55odk.jpg (http://www.abload.de/img/jinc55odk.jpg)
There are indeed some very bad commercial DVDs out there. I´m not sure why (compressibility? lack of time? botch job? all of them?), but a lot of NTSC DVDs tend to have a lot of ringing in them. PAL DVDs not so much. I own a big shelf of DVDs that I am going through, because I couldn´t get them to work with madVR until now.

The example you provided indeed looks horrible. What I would find interesting to play around with is some fine-tuning value for the anti-ringing filter, so for such cases where we actually want to remove the ringing some more, we could do that. We would still have the added benefits of the Jinc algorithm and the image (luma) would still be reasonable sharp. If we´re dealing with high quality content, tiny amounts of ringing are not bothering me at all (AR could be turned off altogether). It all also depends on how far someone is sitting away from the screen, because at a distance of 2-3 meters away, it´s extremely hard to impossible to make out ringing at all. In this case I usually prefer a sharper image.

[6233638]

I have a GTX580, currently still running on the 301.40 beta drivers. If it´s not because of the beta drivers, I must say that I am out of ideas, your setup seems to be very close to mine. If it´s not the graphics drivers, could this actually be caused by the soundcard´s drivers? I have a Creative X-Fi Titanium PCIe. Strange, I just use HDMI out from my GTX 570. You are definitely using the default Windows 7 DVD Navigator?

There are indeed some very bad commercial DVDs out there. I´m not sure why (compressibility? lack of time? botch job? all of them?), but a lot of NTSC DVDs tend to have a lot of ringing in them. PAL DVDs not so much. I own a big shelf of DVDs that I am going through, because I couldn´t get them to work with madVR until now. Actually, most of the DVDs I watch are PAL, and tend to suffer from a lot of ringing.

The example you provided indeed looks horrible. What I would find interesting to play around with is some fine-tuning value for the anti-ringing filter, so for such cases where we actually want to remove the ringing some more, we could do that. We would still have the added benefits of the Jinc algorithm and the image (luma) would still be reasonable sharp.To clarify, the ringing there is mostly encoded in the source, it is not Jinc at fault. I was just making the case that softer filters tend to be preferable with commercial DVD content. (Jinc seems well-suited to 720p content)

The anti-ringing filter does not remove ringing that is encoded in the source, only ringing that is caused by the scaling algorithm used.

[strumf666]

I don't know of any display which accepts a 25Hz (25p) signal. With Nvidia, "25Hz" actually outputs 1080i50 if I remember correctly. (which is then deinterlaced to 1080p25)
AMD outputs i50 as well when I select 25Hz refresh in catalyst and while I haven't tested it on current htpc, on my previous (E-350 using EVR) the result was unwatchable. Manually setting refresh to 50Hz did the trick though.
MadVR switches to 50Hz in case of 25p source now that I corrected the configuration so I am quite happy :)

[jmone]

Wow - been away for a few weeks so have just had a chance to play with the new Jinc / Anti Ring algorithm but it is too much for my 550ti in the desktop with heaps of dropped frames (96%+ GPU usage, spiking temp and fan ramps up on VC-1 1080(i) content) and is way to taxing for the iGPU on the i7-2600K in my HTPC with most material. While it looks great, what HTPC friendly GPU are people using?

[nevcairiel]

I'm running a 660 GTX now, the ASUS DirectCU II models have some good coolers which are really quiet. I'm a bit surprised that the 550Ti doesn't manage to do it, tbh.

[jmone]

So was I and it was only VC-1(i) content that caused the low render queue issue on the 550Ti as 264 1080(i) content was in the low 90% GPU level but was stable so it must be right on the max threshold for the 550Ti.

[Pat357]

Strange, I just use HDMI out from my GTX 570. You are definitely using the default Windows 7 DVD Navigator?
I think he is using PotPlayer and you're using MPC-HC, right ?
Well, I 've just discovered that LAV-video indeed works with MadVR for DVD-playback with *PotPlayer* !
I 've retested again with MPC-HC but no avail : "macrovision" error, even if the DVD has absolutely no copy-protection (made DVD myself).
To make things more complicated : I can watch DVD's using MPC-HC & MadVR if I use this freeware Open-source decoder : GPL MPEG1-2 decoder (http://sourceforge.net/projects/gplmpgdec/). It works perfect with the Win7 DVD navigator.
This makes me think that there must be something in the decoder that makes it work with the 7 DVDnav or not. But What is it ??

PS : can you try the open source GPL MPEG1-2 decoder (http://sourceforge.net/projects/gplmpgdec/) with MPC-HC & MadVR to play a DVD ? I want to be sure it's not a coincidence that it works for me.
If you don't like to run the installer, just open it using 7z or a .MSI unpacker : there is just the .ax and a .txt file IIRC.
Register .ax file : "regsvr32 filename.ax"

[Xaurus]

I'm running a 660 GTX now, the ASUS DirectCU II models have some good coolers which are really quiet. I'm a bit surprised that the 550Ti doesn't manage to do it, tbh.
That's the same gfx card I am running in my htpc, it's a great card and the Asus coolers are really quiet. I went from a passive gfx solution to that card and to be honest I can't hear any difference.

And it tackles anything you throw at it (madvr jinc etc).

[iSunrise]

Strange, I just use HDMI out from my GTX 570. You are definitely using the default Windows 7 DVD Navigator?
Yes, I am definately using the default Windows 7 DVD Navigator. I´m not using MPC-HC though. In PotPlayer under Preferences -> Hardware -> DVD/Blu-ray in the Navigator selection I have selected "Microsoft DVD navigation (recommended)", which is also the default.

Here´s my filter chain (as configured in PotPlayer):

DVD Navigator
LAV Video Decoder
Madshi Video Renderer
LAV Audio Decoder
DirectSound Audio Renderer

(I also have LAV splitter source configured for MPEG2 PS and MPEG 2 TS.)

I think he is using PotPlayer and you're using MPC-HC, right ?
Well, I 've just discovered that LAV-video indeed works with MadVR for DVD-playback with *PotPlayer* !
I 've retested again with MPC-HC but no avail : "macrovision" error, even if the DVD has absolutely no copy-protection (made DVD myself).
To make things more complicated : I can watch DVD's using MPC-HC & MadVR if I use this freeware Open-source decoder : GPL MPEG1-2 decoder (http://sourceforge.net/projects/gplmpgdec/). It works perfect with the Win7 DVD navigator.
This makes me think that there must be something in the decoder that makes it work with the 7 DVDnav or not. But What is it ??
The strange thing is that I have basically the same filter chain as 6233638. Even the video decoder is the same (LAV). The only things that are different are the audio renderer, the graphics card drivers and the media player.

The anti-ringing filter does not remove ringing that is encoded in the source, only ringing that is caused by the scaling algorithm used.
I know and that´s also how it should be. What I´m wondering though (just out of interest) is if it would be possible to use that algorithm in such a way that you could upscale e.x. with Jinc and let the anti-ringing filter remove ringing more aggressively, so that we don´t have to use SoftCubic which basically softens the whole image (if used on luma). Because that´s one of the reasons why you (or even me in some cases) use SoftCubic in the first place, to hide source artefacts (which alters the image to some extend).

[NicolasRobidoux]

What you may want to do is use a filter that's strongly smoothing, and post-sharpen.