madVR - high quality video renderer (GPU assisted)

[Stereodude]

How do you know that?
I looked at the benchmark results posted by Sunset1982 and made a little interpolation.

http://forum.doom9.org/showthread.php?p=1774940#post1774940

The RX 480 is on par with the GTX 1060 in madVR and neither is much faster than a GTX 970 with madVR. The RX 480 has ~50% more FLOPS than the GTX 1060, but those don't apparently show up with madVR. The high FLOPS of the various R9 cards showed up in madVR.

AFAIK, the R9 380 was just slightly more powerful than the GTX 970 in madVR. The 390 scaled linearly from the 380 with madVR pretty much in direct relationship to the FLOPS. The 390 about 47% faster than the 380. So, having the RX 480 about 23% faster than the GTX 970 leaves it well short of the R9 390 despite having more FLOPS.

[ionutm80]

But in my case we are not talking about interlaced video as the flag showed by MediaInfo clearly states 25fps progressive and I can confirm that is was ripped from a PAL DVD Disney movie.
Did LAV ignored the flag and forced a deinterlacing w/o being the case?

Case closed :)
I have put no software deinterlacing in LAV and now there are no more repetead frames and movie info in madVR shows 25fps instead of 50 fps.

[huhn]

all mepg2 dvds are interlanced i can only guess that this DVD doesn't need an deinterlancer and works fine without. but that a guess.
media info is not 100% reliable.

the repeated frames are still a bug "somewhere"

if it is in madVR or lavfilter is sending work frame times or if the file is "broken" i have no clue.

[Stereodude]

Is there a flag for 3D that can be used to pick profiles? I didn't see anything in the 3rd post (http://forum.doom9.org/showpost.php?p=1271417&postcount=3). I see HDR, but no 3D. :(

Or is there some other way to identify when madVR switches to 3D (frame packed) output?

Edit: 3D is a valid boolean value for profiles. The profile guide in post 3 is outdated.

[AntonP]

You would have no visible drops with 23.971 with reclock either. I've been using it this way for years. You may have to clean out the timings in reclock config so it can learn the 23.971 refresh rate. As far as I know for 3d, you get 23.971 with Nvidia. I tried making a custom refresh rate with nvidia and the 3d got amazingly messed up. I gave up on the Nvidia at this point as I'd like to be able to try DTSX and Atmos and they need to bitstream with 23.976.

Reclock cant adapt 23.976 to 23.971, it hasnt such an option (usually its 24 or other STANDART refresh rate).
And 24fps movie + 23.971 refresh rate = drops

[Asmodian]

Hopefully not to bring begin a chroma discussion yet again here.. but I'd totally recommend you evaluate super-xbr 100 and 125 AR in comparison to Recon soft + SR1.

You are quite right, super-xBR + SR1 is very nice, I think I will switch to super-xBR 100 + SR1. I had tested super-xBR for chroma when it was first added but with the current SR1 I like the look of super-xBR for chroma upscaling a lot more. Thanks. :)

Reclock cant adapt 23.976 to 23.971, it hasnt such an option (usually its 24 or other STANDART refresh rate).
And 24fps movie + 23.971 refresh rate = drops

Reclock adapts the audio to your actual measured refresh rate. This is why it has its video clock database that you can clear, it is constantly measuring and updating the audio speed adjustment to keep everything in perfect sync. Watching 23.976 video on a 23.971 Hz display, without frame drops or audio skips, is what Reclock is for.

As far as I know there are no 24fps movies, I have never seen one anyway. Almost everything is 24/1.001 fps.

[stefanelli73]

With classic blu ray you must activate the HDR functions madVR or not?

[ryrynz]

You are quite right, super-xBR + SR1 is very nice, I think I will switch to super-xBR 100 + SR1. I had tested super-xBR for chroma when it was first added but with the current SR1 I like the look of super-xBR for chroma upscaling a lot more. Thanks.

NP. Hope you ticked AR. I still think SR1 is a tad over the top with chroma and prefer to just notch up to 125 myself.

[chros]

that's not how it works.
film in lavfilter is just halve frame rate deintelerancing which is usually not recommand at all. because they are not restoring the orginal frame rate.

if the source is interlanced than it has 60/50 fields worth of 60/50 different frames so film mode in lavfilter is "just" throughing halve of them away.

than we have interlanced content that is PsF and don't need a deinterlancer.
and we have telecine which is 24 frames on 60 fields (and other stuff) the film mode in lavfilter is not restoring these the film mode in madVR is.

if the source is telecine a IVTC (film mode) algorithm is clearly better.

50i usually needs a deinterlancer or can be treated progressive.
Yes, thanks for the correction.

[nevcairiel]

With classic blu ray you must activate the HDR functions madVR or not?

Classic Blu-rays are not HDR, so you do not need to do anything there.

[ibius]

As far as I know there are no 24fps movies, I have never seen one anyway. Almost everything is 24/1.001 fps.

Of the top of my head, Yip Man (2008) is 24.000 fps, and I recently watched 5 to 7 (2014), also 24fps, there were others, can't remember now.

[Sparktank]

As far as I know there are no 24fps movies, I have never seen one anyway. Almost everything is 24/1.001 fps.

There are.
Some regional blurays outside of North America and UK will sometimes transfers movies in 24fps.

[DragonQ]

There are.
Some regional blurays outside of North America and UK will sometimes transfers movies in 24fps.

Yep. Don't forget that 23.976 fps only really exists because of legacy reasons. Films are usually shot at 24 fps and some are transferred to BD like that.

[Sunspark]

NP. Hope you ticked AR. I still think SR1 is a tad over the top with chroma and prefer to just notch up to 125 myself.

Honestly, I feel chroma Bicubic 60 AR paired with image Jinc AR is a good combination. The image doesn't get oversharpened, and the colour has a smoother almost creamier tonality to it.

A video test I believe that helps prove this is the first few seconds of the opening of the movie the Revenant. The entire movie is filmed in natural light. The bit where the boy is backlit by the flames, with bicubic 60 AR you see a soft red glow around him on the ground. With some of the other chroma algos this glow seems to be faded or less saturated.

[Sparktank]

Yep. Don't forget that 23.976 fps only really exists because of legacy reasons. Films are usually shot at 24 fps and some are transferred to BD like that.

I'm fully aware of that. I have been for a long time but it's good to let others know.

[Georgel]

On 720P anime which is already reasonably detailed these settings come across to me to be quite excessive and would likely be harming image quality considerably, debanding at maximum for 720P?
How bad are these videos?

The first thing you should be doing is viewing high bitrate videos, I keep mine on low, set a shortcut to change it if the source requires.

On 720P I usually just upscale with NNEDI3 (which doesn't make that much of a difference on a 1080 screen TBH) downscale with SSIM 1D without AB maybe do some reasonable SE and CE sharpening and that's it. Line thinning was kinda useful on some soft ~640x480 anime I had but it even at lower settings in some areas didn't produce good results which is why I asked madshi about adding warpsharpen to madvr which is considerably better for anime.



Link me a sample file that you feel your settings look good on and upload your settings.dat and I'll do a screenshotcomparison vs what I would use.
I might be able to make you reconsider what is considered good image quality.



Madshi, would it be possible to add to madVR the ability for it to remember settings for certain videos?

For example if I watch a video that has poor color and lots of banding and I hit a couple of hotkeys to change these defaults, could another hot key be pressed to remember those changes for any future videos played within that directory? That way if watching a series the user doesn't need to perform these adjustments again and again.


I actually like the smoothing effect that debanding on high has.

One observation though, for some anime* reducing ringing and reducing halos thingies can degrade seriously the gradients and the overall quality of the image.

[Georgel]

Debanding as set there has nothing to do with the banding due to lack of dithering, in fact on high/high it would make it worse by blurring details that help disguise the banding caused by rounding down (instead of dithering) madVR's internal 16-bit data to 8-bit.

Are you watching close to a large low resolution display?



Debanding high/high? Thin edges 8.0? Sharpen edges 4.0? Dithering off?

I watch mostly anime and I would never use anything like that and strongly recommend no one use them unless testing for themselves. Definitely settings that should not be recommended generally.

Those settings give off a sensation of higher quality / resolution without adding color noise, or creating what I call false texturing.

The idea for anime is that some color portions are supposed to be smooth, not have a certain color noise texture. I want to erase that texture and be left with very smooth gradients on full color areas.

Dithering is a cause of [false texture]. It adds a texture that is not supposed to be there. And because madVR has a high bit depth, I am able to use no dithering, otherwise it would look really awful. Testing through many settings, for Dithering to work nicely on off, one has to select 10 bit in Devices -> Properties.

Then again, some people might expect color texture instead of smooth color gradient areas, and as such, might not fully like these settings.

Observation: Not all anime look their best on these settings, since some anime intentionally have a texture, that if erased, the entire thing looks awkward, this type of setting works best with most anime, not all.

[Georgel]

Happy Happy Joy Joy then, what matters is that you like what you see coz at the end of the day the only thing that matters is your own satisfactshun.

All this said I have no idea how anyone can bear having any of the two dithering sub-options enabled so maybe you could try vanilla ED2 with both of these pesky things disabled as one makes the noise dance around on top of the picture and the other creates very faint color fringing :angry:

I will keep testing, It's a pretty nice hobby :D

One thing though, when testing all dithering features, I still can't escape the new texture created thingy (?)

I do wonder why we would need dithering at 10 or even 8 bit.

It's easily visible on what dithering brings to the table on less than 8 bit processing, but it's unclear on what the advantages are on more than 8 bit processing. If someone wants to post comparation images that can show a visible advantage, I'm very curious. When doing this, also have activated debanding on high, since that alleviates all other problems that ditehring was supposed to.

[ryrynz]

One observation though, for some anime* reducing ringing and reducing halos thingies can degrade seriously the gradients and the overall quality of the image.

If you're aware of limitations on certain content with reducing ringing please upload before and after screenshots or a sample, thanks. The halo reduction as stated in the options isn't recommended for anime.

[huhn]

it is not hard to show why dithering is needed.

dithering (https://abload.de/img/dithering56q1o.png)
no dithering (https://abload.de/img/noditheringg0qil.png)

it is needed for correct color anyway.

[ryrynz]

it is needed for correct color anyway.

He's aware that he's destroying his image and he's okay with it ;)

[Georgel]

He's aware that he's destroying his image and he's okay with it ;)

it is not hard to show why dithering is needed.

dithering (https://abload.de/img/dithering56q1o.png)
no dithering (https://abload.de/img/noditheringg0qil.png)

it is needed for correct color anyway.

If you're aware of limitations on certain content with reducing ringing please upload before and after screenshots or a sample, thanks. The halo reduction as stated in the options isn't recommended for anime.

How do you get that image ?

But even so...

There is something going on on my side.

Just let me try one thing.

This is processed through all of my filters. Adding dithering will just add noise. but it's an exact frame, and can;t find the episode or the frame right now.

https://www.dropbox.com/s/zvj84qdbj8u00o9/Illya-120.pngdl=0

https://www.dropbox.com/s/zvj84qdbj8u00o9/Illya-120.png?dl=0

Click for full res I think?

The point is, it looks better than original, better than dithered, and has no visible problems, it's more high resolution and looks smooth.

EDIT ::: What I don't actually get is... It does erase some things... like smoothens things out... like, this would look kinda too smooth, or way too smooth for move materials, unnatural. but for anime, it's just so much cleaner and higher resolution looking... With anime, all that it erases are actually bits of noise? I don't know, when it erases something, what gets erased looks like it was not overly relevant.

[huhn]

the dithered image is "perfect" make sure it isn't scaled.

[Georgel]

the dithered image is "perfect" make sure it isn't scaled.

Oh.

Looking over the reasons why adding dither might have noise, my panel is 6 bit. Now this is messed up, but it's the most sold laptop high end display, and it has just 6 bits.

Maybe this is a reason why on my setup I see un-dithered images better? I mean, it would make sense for it to be otherwise, as far as common sense goes.

http://www.panelook.com/LP173WF4-SPF1_LG%20Display_17.3_LCM_overview_21056.html

Something seems wrong on my side honestly.

Well, many many people have this display, so the instructions are not entirely wrong, but curiosity is burning me on what is going on.

It just looks cleaner to me when deactivating dithering. On the display. Man, I can't wait to get my 100% adobeRGB AUO to escape this nonsense.

The more scary thing is that I too think that it should be the other way around, I do believe that dithering should make it better. It just does not work on my setup, adding a layer of visible noise :angry:

[Ver Greeneyes]

The more scary thing is that I too think that it should be the other way around, I do believe that dithering should make it better. It just does not work on my setup, adding a layer of visible noise :angry:
Dithering does add noise, it spreads the error out so on average pixels will have the right color. This is best seen on a video of a still image with "change dither for every frame" checked (and playback not paused). The result will be slightly more noisy, but the colors should be much more correct on average.

Since your panel is 6-bit (it may have its own high frequency dithering up to 8-bit, usually indicated with the acronym "FRC"), have you tried telling madVR this by settings its "native display bitdepth" to 6 bit in the devices section? 8 bit dithering should be almost imperceptible, so I expect that 6 bit dithering will be more noisy, not less - but 6 bit does match your display.

[Georgel]

Dithering does add noise, it spreads the error out so on average pixels will have the right color. This is best seen on a video of a still image with "change dither for every frame" checked (and playback not paused). The result will be slightly more noisy, but the colors should be much more correct on average.

Since your panel is 6-bit (it may have its own high frequency dithering up to 8-bit, usually indicated with the acronym "FRC"), have you tried telling madVR this by settings its "native display bitdepth" to 6 bit in the devices section? 8 bit dithering should be almost imperceptible, so I expect that 6 bit dithering will be more noisy, not less - but 6 bit does match your display.

I've been using 10 bit in display setup, since setting that to 6 bit seemed to affect the image in a negative way, seemed to induce more noise. I think that you're right and the display already applies a layer of dithering over the image. It's theoretically capable of a bit larger color space coverage than an average LCD despite being 6bit, so I think it has some software doing this.

Maybe this is how it works for this display only... Too bad it almost the universal display for 17" FHD high end gaming laptops. It's even Gsync capable, but not on my Acer...

I guess that I'll probably note a change on the settings I recommend after getting a 99% adobeRGB display to work on :rolleyes:

[Patrik G]

question:
why is motion resolution so low with madrv and the FPD motion test?
this 500M has over 800 lines of motion resolution but with madvr and this test it judders like crazy.
even down to 100 lines.

i have tested with different settings and there is no dropped frames.
checking the same motion resolution test with windows build in movie viewer and its smooth as it should.

what trix needs to be disabled in madvr to get back the default motion resolution?

[huhn]

what settings are you using and is your TV using interpolation at what refreshrate is it running?

can you amke a screen of the OSD and a screen with athe windows movie player?

[ryrynz]

https://www.dropbox.com/s/zvj84qdbj8u00o9/Illya-120.pngdl=0


That does't look like it has thin edges 8 applied to it at all. I'm not seeing any edge issues there..Really need the original frame, but if I'm judging this picture on it's appearance, I can't fault it.
In saying that, other content may well look worse with those settings.

[pose]

ryrynz, here is comparison from his own site. To me the after picture looks unnatural and with lots of details missing.
https://3.bp.blogspot.com/-wv05mfBisf4/Vrpgn1275jI/AAAAAAAAALs/3tMaRr_TRKc/s1600/Illya%2B-%2Ball%2Benhancements%2Bturned%2Boff.png
https://1.bp.blogspot.com/-S6k4Cru_JJs/VrphhvEy78I/AAAAAAAAAL4/ZgY7p58vGBA/s1600/Illya%2B-%2Bedge%2Bthinning%2B%252B%2Bedge%2Bsharpening%2Bx4.png

[ryrynz]

ryrynz, here is comparison from his own site. To me the after picture looks unnatural and with lots of details missing.
https://3.bp.blogspot.com/-wv05mfBisf4/Vrpgn1275jI/AAAAAAAAALs/3tMaRr_TRKc/s1600/Illya%2B-%2Ball%2Benhancements%2Bturned%2Boff.png
https://1.bp.blogspot.com/-S6k4Cru_JJs/VrphhvEy78I/AAAAAAAAAL4/ZgY7p58vGBA/s1600/Illya%2B-%2Bedge%2Bthinning%2B%252B%2Bedge%2Bsharpening%2Bx4.png

Thanks. It's certainly further than I'd go, the edges aren't something you want to thin down considerably.. better to sharpen them and just thin them just a little.
The overall detail is still there there's actually little change to most of the image except for the lines, I don't understand the reason why people want to thin them so much, it ruin's the appearance of the characters.

Sometimes I found even thin edges at 1.0 went a bit overboard so I stopped using it and just rely on a Awarpsharpen, and a touch of line darkening through Avisynth along with some sharpening through madVR depending on the source.
I can make IMO 480 anime look similar to 720 content. It's all about having a firm yet gentle hand, it's easy to over do it.

[Georgel]

Thanks. It's certainly further than I'd go, the edges aren't something you want to thin down considerably.. better to sharpen them and just thin them just a little.
The overall detail is still there there's actually little change to most of the image except for the lines, I don't understand the reason why people want to thin them so much, it ruin's the appearance of the characters.

Sometimes I found even thin edges at 1.0 went a bit overboard so I stopped using it and just rely on a Awarpsharpen, and a touch of line darkening through Avisynth along with some sharpening through madVR depending on the source.
I can make IMO 480 anime look similar to 720 content. It's all about having a firm yet gentle hand, it's easy to over do it.

Mostly because my screen in FHD and I want all anime to look FHD :D

Some algorithms are less needed in some anime that were natively produced in FHD, so I'm playing around with my settings a lot, but that image represents exactly why I love edge thinning.

It;s mostly about getting rid of blurry and too thick edges that make things look fuzzy. Not that much about thinning really, as much as about making edges look tangible, opposed to the gaussian blurred filtered look they have sometimes.

Well, I agree that sometimes it might be overkill and the image might degrade, so I think I might have to also step down a bit on the details.

I also think that for my images examples I'm using upscaling refinement -> edge thinning 8.0 while for image enhancements I'm stepping it down a lot, as it does break eyes of anime characters.

For real media content, I think that the sweet spot is (for me?) around 2.8 for edge thinning on image enhancements, whilst I'm still testing for upscaling refinement.

[Georgel]

ryrynz, here is comparison from his own site. To me the after picture looks unnatural and with lots of details missing.
https://3.bp.blogspot.com/-wv05mfBisf4/Vrpgn1275jI/AAAAAAAAALs/3tMaRr_TRKc/s1600/Illya%2B-%2Ball%2Benhancements%2Bturned%2Boff.png
https://1.bp.blogspot.com/-S6k4Cru_JJs/VrphhvEy78I/AAAAAAAAAL4/ZgY7p58vGBA/s1600/Illya%2B-%2Bedge%2Bthinning%2B%252B%2Bedge%2Bsharpening%2Bx4.png


I don't think that details are missing. The backgrounds look less blurry really, even closer to a clear image :eek:

[Patrik G]

what settings are you using and is your TV using interpolation at what refreshrate is it running?

can you amke a screen of the OSD and a screen with athe windows movie player?

settings doesnt matter as i have changed all and it doesnt affect the motion test.
i use 60Hz at 1080P
run the test for yourself and see what you come up with.
it runs at 60fps
http://privat.bahnhof.se/wb192876/FPD_Motion%20Resolution.mp4

but madvr and its smooth motion must be doing something good because motion on 24p content is way better (more fluid) than on the build in windows media player which this test runs smooth on.
so i think this motion test doesnt say it all about motion performance.
not for 24p content anyway.

but just for the papers

[huhn]

this test file is kind of broken.

it is interlanced and madVr isn't deinterlancing it because it is saying that is has 60 FPS before deinterlancing.

when lavfilter is deinterlancing than madVr things this file is 120 hz.

[a8213711]

I just discovered the results of Image Enhancements don't appear in the images saved with MPC-HC; are there many other functions which don't appear when saving images?

[huhn]

if i'm not mistaken nearly everything is not applied there.

[Asmodian]

if i'm not mistaken nearly everything is not applied there.

Doesn't that function actually get the video before madVR has even done scaling? I think it grabs the frame after decoding but before madVR.

[huhn]

you still need to do a YCbCr->RGB conversation and madVR has to do that.

i have no clue if the selected chroma scaler is used for that or not.
if i remember correctly it is dithering even with disabled dithering(don't do this).

[Georgel]

I just discovered the results of Image Enhancements don't appear in the images saved with MPC-HC; are there many other functions which don't appear when saving images?

But the enhancements appear when you press print screen, if you want to notice differences, as far as I know.

[vivan]

if i remember correctly it is dithering even with disabled dithering(don't do this).AFAIK it dithers on CPU (using error diffusion).

[ibius]

Doesn't that function actually get the video before madVR has even done scaling? I think it grabs the frame after decoding but before madVR.

'Crop black bars' gets applied to saved images, so madVR is active.

@madshi: speaking of cropping black bars, could we get file name tags for that, please?
Your last reply to leeperry on the subject didn't say whether you would implement it or not.

[ryrynz]

I don't think that details are missing. The backgrounds look less blurry really, even closer to a clear image :eek:
You quoted me for something I didn't say. I ran both through a picture comparison tool and there's little change in detail nothing really noticeable.. It's just the lines which we already know are thinned too aggressively.

Sent from my Nexus 5 using Tapatalk

[v0lt]

@madshi
I have a mk3d file with the following parameters:
MultiView_Count : 2
MultiView_Layout : Side by Side (left eye first)
Is there a way to transform this video to "row interleaved" format? With madVR of course.

[Warner306]

@madshi
I have a mk3d file with the following parameters:
MultiView_Count : 2
MultiView_Layout : Side by Side (left eye first)
Is there a way to transform this video to "row interleaved" format? With madVR of course.

madVR will only accept frame-packed 3D. So, no, it can't be converted.

[Stereodude]

madVR will only accept frame-packed 3D. So, no, it can't be converted.
SBS or TAB are just 2D, so of course madVR can accept them. It can output row interleaved too. However, I don't know if you can tell madVR that a 2D input video is a certain type of 3D and have it handle it accordingly.

[huhn]

that's what warner is saying.

in it's current state madVR only detects frame packed 3D as 3D and it can only convert these. more 3D features where planned AFAIK like the feature Volt is talking about reading some flags in mkv and convert the input to the selected output.

[Sunspark]

I can make IMO 480 anime look similar to 720 content. It's all about having a firm yet gentle hand, it's easy to over do it.

The after image linked was too heavy a hand. If you compare the red and yellow bows they wear, especially the one in the center, the yellow portions have actually shrunk and are now smaller than they were before by several mm.

[fhoech]

how does your 3d lut know the brightness of the file?

It doesn't. All that it knows is the HDR peak brightness (of the display), as set during 3D LUT creation, then applies the BT.2390 rolloff + saturation adjustment, then maps into the display colorspace (with imo great results btw, my implementation has come some way in the past three weeks and will be available in the next version of DisplayCAL). Of course, a 3D LUT cannot react to dynamic metadata, which could be seen as a drawback. So far, I don't think it matters much, though.

I'm afriad it's not in the current versions from v0.90.22:
Sure, I know that. I'm currently working around it by having two versions of the files: One with HDR metadata (handled by madVR's HDR implementation) and one without (HDR handled by the 3D LUT). The only thing that's a little bothersome when I do comparisons is that I have to remove the HDR 3D LUT when using madVR's HDR, otherwise the HDR processing is done twice (once from madVR itself and once through the 3D LUT).

You can try v0.90.20: if it's working there you can ask madshi to revert/modify that change.
When I downgrade I won't be able to easily do comparisons. Having a way to disable madVR's HDR processing but not disable the 3D LUT would be useful. Notifying @madshi.

[AngelGraves13]

For some reason I keep going back to Jinc for all my scaling needs. It seems to look more natural and has almost no artifacts unlike the other scalers.

NNEDI3 128 does look quite good. 256 looks a tiny bit better, but my 1080 GTX tends to get hot with that setting and the render times are at about 39 ms, which is too close to the frame time of 24p content. Should I lower the setting if the max sometimes hits 45 ms, but the average is around 30 ms? I'd ideally want the max under the 41.7 ms of 24p content?