madVR - high quality video renderer (GPU assisted)

[Aktan]

Originally Posted by Stereodude

That might work for DVDs, but it doesn't work for broadcast 1080i60 HDTV where there are lots of logos crawls and other crap outside of the center that will mess it up if you don't limit it to the center.

Ah, well then damn if you do, damn if you don't? I thought for broadcast, it be pure interlace anyway, not film video. I guess you need to change when needed? I don't know..

Originally Posted by Stereodude

It seems the HW decoders ignore the soft pulldown flags, not honor them. Further, I'm not sure what you two are seeing, but disabling deinterlacing with software decoding (avcodec) with the video card outputting 24/1.001Hz looks terrible and the render times are higher to boot. Maybe it's engaging smooth motion. I haven't spent a lot of time with it.

Sorry, I forgot to mention I was using LAV Video to software decode. It being MPEG2 and DVD resolution, I didn't think there be a need for HW acceleration. When you said looks terrible, what you mean. Lack of post processing or jittery movement?

[CarlosCaco]

Originally Posted by Thunderbolt8

Im a bit confused by the terms 8-bit, 10-bit etc. in relation to the number of colours. my monitor has 16.7 million colours, but I dont know if that is 8-bit or 10-bit.

I had this doubt too, since is not register in edid, or in display tecnical specs or in any site,
i found a lot of posts asking about, in one madshi says to use the test paterns especialy the colours and try different bitdepths, and see what looks best, also i made the procediment to know if my display supports 10 bit (there is a thread about in the forum) i tested milion times 6bit, 7bit and 8bit now i clear see the diference in the patterns in my case bellow 8bit produce banding and above 8bit i can see more banding and seens that the images has no dithering
after all of this i calibrated my display with dispcalgui and on iccprofileinfo was showing 8 bitdepth

[ibius]

Originally Posted by Thunderbolt8

Im a bit confused by the terms 8-bit, 10-bit etc. in relation to the number of colours. my monitor has 16.7 million colours, but I dont know if that is 8-bit or 10-bit.

When you do the math, 16.78 million is 8-bit, displaying 256 red, 256 green and 256 blue shades. Look up specs of your monitor, depending on how they decided to phrase it you might have to dig deeper, better yet find a detailed review. Look for 'Colour Palette' of your panel, if it says '1.06 billion' but the 'Colour Depth' says '8-bit', then it's not a true 10-bit panel (like 99% out there), but is capable of producing 1.07 billion colours. That this is done with an 8-bit colour depth and an additional Frame Rate Control (FRC) stage (Temporal Dithering).
But I doubt you will see a difference outputting 10-bit on a '8-bit + FRC' panel vs properly dithered madVR output on 8-bit panel.

On a side note, did anyone with a '8-bit + FRC' panel make some tests displaying 10-bit with madVR with dithering disabled to compare panel's dithering to that of madVR?
And, is madVR's Ordered dithering enough to smooth out the gradients of a dithered 10-bit on a 8-bit panel or is Error Diffusion required?

[CarlosCaco]

Originally Posted by ibius

When you do the math, 16.78 million is 8-bit, displaying 256 red, 256 green and 256 blue shades. Look up specs of your monitor, depending on how they decided to phrase it you might have to dig deeper, better yet find a detailed review. Look for 'Colour Palette' of your panel, if it says '1.06 billion' but the 'Colour Depth' says '8-bit', then it's not a true 10-bit panel (like 99% out there), but is capable of producing 1.07 billion colours. That this is done with an 8-bit colour depth and an additional Frame Rate Control (FRC) stage (Temporal Dithering).
But I doubt you will see a difference outputting 10-bit on a '8-bit + FRC' panel vs properly dithered madVR output on 8-bit panel.

On a side note, did anyone with a '8-bit + FRC' panel make some tests displaying 10-bit with madVR with dithering disabled to compare panel's dithering to that of madVR?
And, is madVR's Ordered dithering enough to smooth out the gradients of a dithered 10-bit on a 8-bit panel or is Error Diffusion required?

so i not loosing information outputing 10bit on 8bit display?

i don´t find any specs about my panel searched everywere,i don´t know if has FRC
i know that if i go d113d exclusive 10 bit with any dither algo the image is similar to 8bit with dithering disabled on madvr

[Knight77]

Originally Posted by Warner306

Try SuperRes(3) + crispen edges(1.0) for 2x scaling factors. I prefer this to SuperRes + thin edges.

Do you suggest this also for 1.5x scaling or in that case you would stick with SuperRes + thin edges?

[michkrol]

Originally Posted by CarlosCaco

i thought that madvr converts or dither down to 8 bit, but in my osd shows: h264, 10bit, 4:2:0 -> P010, 10 bit, 4:2:0

This line informs you what format the video is and what format madvr gets from the decoder, not display depth.
In your case the file is h264 10bit and madvr gets 10bit color format from the decoder. This is good because it means madvr does the dithering to 8bit before sending it to your display. This gets you higher quality than decoder's dithering.

Originally Posted by CarlosCaco

so i not loosing information outputing 10bit on 8bit display?

Technically you might be losing some information, but most people wouldn't notice the difference.
Another question is whether the studio used 10bit while mastering the source, which usually isn't the case.

Originally Posted by CarlosCaco

i know that if i go d113d exclusive 10 bit with any dither algo the image is similar to 8bit with dithering disabled on madvr

It's actually the other way round: 8bit + dither is really close to 10bit without dither.
That said, you should have dither always enabled.

To sum this up: when in doubt leave the settings at defaults to get highest quality. 10bit videos on 8bit displays should look so close to 10bit display that you can't notice a difference.
Your settings are alright, so don't over-think it, just enjoy the movies

[madshi]

Originally Posted by dansrfe

Is it possible for madVR to adjust display backlight over DDC/DI based on profile rules for HDR/SDR media? That way, like calibration curves, backlight can be adjusted automatically as well.

I've no idea if that's possible. And if it worked I've no idea if that would screw up things like calibration and stuff. Might be worth trying at some point, although I don't know if the OS/GPU driver even gives me access to DDC/DI. But this is not going to happen any time soon.

Originally Posted by rack04

I have exhibited random crashes in the past couple of weeks. I have updated madVR, LAV Filters, MPC-BE and graphic drivers to the latest versions but I can't narrow down the culprit. I'm hoping this log provide some kind of clarity. It is a large file so please forgive me.

Unfortunately for crashes debug logs don't really help at all in most cases. I'd need a crash report. When those crashes occur, do you get a window which looks somewhat similar to this?

http://madshi.net/exc-ss3.gif

If so, click on "show bug report", then press Ctrl+C. After that you have the crash report in your clipboard. Upload it somewhere for me to look at.

If the crash box looks different, it's probably a crash in the media player (or decoder or splitter or ...) and not in madVR.

Originally Posted by Warner306

Are there any guidelines for setting peak luminance? Does this even matter with non-HDR content? If so, this option could use better labelling.

It only matters for HDR content at this point. This is just the very first HDR version, so there are no real guidelines yet. Will need some testing and feedback from users.

Originally Posted by SweetLow

madshi in his infinite wisdom make possible to run any executables on (de)activating profile

Well, the option is there, but I think I've not actually implemented that feature yet? Not sure right now...

Originally Posted by mogli

FSE, but I don't think we should spent to much time on this problem if FRC isn't the right way to go for this case and ReClock solves it

Ok.

Originally Posted by Ver Greeneyes

By the way, I have my monitor calibrated to 120 nits (or 120 lux? I always get them confused), well below the minimum of 400 that madVR allows. Is 400 simply the point below which normal (non-HDR) luminance values would start getting cut off? In other words, what is the HDR luminance value of 'normal white'?

You know, I've asked this exact question to my UHD insider, and he didn't know, either. He said he was going to ask his contacts, but I'm not sure we'll get a useful answer. The HDR specs don't specify the luminance value of "diffuse white". Currently I'm assuming diffuse white to be 100 nits. But that's just a guess because that's what DVDs and Blu-Rays are usually mastered for, I think.

Originally Posted by BetA13

Oh an dteh ENhance Detail option, (sorry, just found out about it) i do like. how does it work exactly? at first glanze it look just like an sharpener to me, but its different, i cant tell what it is, but call me interessted.

It increases the contrast between darker and brighter pixels in non-edge areas, which makes fine image detail easier to see, but it may also increase noise, grain and compression artifacts.

The other sharpeners all concentrate on edges, only, and intentionally leave non-edge areas untouched. So the "enhance detail" algorithm basically works on different image regions compared to all other sharpening algorithms.

Originally Posted by BetA13

BUT, is there a way to go lower them 400nits?

Its still a bit too dark for my taste, maybe 200?

Originally Posted by omarank

@madshi: I played a couple of HDR sample videos with the new build, at normal 100 nits peak brightness level. I too feel that I would prefer a brighter output than it is at 400 nits setting of “display peak luminance”. Can you please offer 100 nits option in the dropdown for “display peak luminance”?

Originally Posted by 6233638

So my display is currently calibrated to 100 nits, since that is the intended reference level for SDR.
I can get that up to 300-350 or so if I disable backlight scanning, trading motion clarity for brightness.
But with HDR content not being very common right now, I'd be more inclined to just compress HDR sources to 100 nits than mess with creating a new preset on the TV that I have to switch to. Do you plan on extending the range of options for this?

Ok, the next build will have options for 265, 180 and 120 nits.

Originally Posted by BetA13

After downloading the new madTestPatternSource samples, i wanted to try them, but found out that my MPCHC64bit didnt want to. cant handle the file.
I did install teh new madTestPatternSource stuff and the bat and all..

It did work with MPCHC32bit.
So i guess they only work with 32bit player?, or am i missing something?

Yes, the test patterns only work with 32bit players atm.

Originally Posted by 6233638

Something that I was wondering though: I believe that one of the videos I have is HDR, but lacking any of the metadata, as it has that very desaturated, washed-out look that HDR videos did prior to this update.
Is there any way to tag a file as HDR, or is it basically beyond repair if that metadata is missing?

Seems like it doesn't even have the proper flags for the SMPTE 2084 transfer function. I guess it might be possible to make it work ok by providing the missing information via file name tags, but at this point I don't think it's worth adding those tags, especially because you'd probably have to guess the exact values, anyway.

Originally Posted by 6233638

So having support in madVR would be a much better solution.
I don't expect it to be a high priority item, but I would like to see BFI support at some point.

This is unlikely to happen before we get displays which can do really high FPS. Maybe in 2-3 years when we get fast OLED displays, we can revisit BFI. At the moment most displays top out at 60Hz, which makes BFI impossible without major flickering.

Originally Posted by GREG1292

Madishi on a 9500LC crt projector what is the nit capabilities of the projector? Have all the latest direct chain mods and Moomee card and I can see changes all the way up to 10000nit.

I've no idea. Just try which value looks best to your eyes.

Originally Posted by 6233638

I can't think of many consumer displays, CRT or not, which would exceed 400 nits brightness.

True. However, things are more complicated than simply measuring the brightness. E.g. if you're watching a movie in a "bat cave" with black walls/floor/ceiling and a good projector, your eyes will adjust to the much lower overall brightness level compared to an LCD flat panel. So the same e.g. 400 nits setting might work well for such front projection, too, due to the magic of our eyes.

Basically this means that we have to take ambient light levels and that kind of stuff into account, too.

[QUOTE=Stereodude;1750030]I've come across a few scenes in some NTSC DVDs where the madVR IVTC algorithm freaks out and drops the 3:2 pattern and goes unknown before recovering when the shot changes./QUOTE]

Originally Posted by Stereodude

Except the scenes are not fades. They just cause misdetections. I will cut one of them from the DVD and upload a sample.

Edit: Here's a sample. Enable deinterlacing and watch.

Thanks for the sample. The more samples you can provide, the better. So if you stumble over this problem again, please create more samples.

Originally Posted by Warner306

Can someone tell if there is any difference at all between choosing "disable calibration controls for this display" and "this display is already calibrated?"

"disable calibration control" completely disables any gamma and gamut adjustments. If you set madVR to "this display is already calibrated", then depending on the video properties, madVR might do some adjustments or not.

Blu-Rays are usually encoded to BT.709 primaries. NTSC DVDs are usually encoded to SMPTE-C primaries. PAL DVDs are usually encoded to EBU/PAL primaries. In theory, in order to watch all these 3 different sources with correct colors, you'd have to calibrate your display 3 times with 3 different presets, one for each of those gamuts. I don't think anybody is doing that, not even professional calibrators.

Most displays these days try to get near to BT.709. Most calibrators aim to achieve BT.709, too. If your display is calibrated to give you correct colors for BT.709, then Blu-Rays will look correct in your setup, regardless of whether you set madVR to "disable calibration control" or "this display is already calibrated". However, if in this situation you play e.g. an NTSC DVD, there will be a difference. With "disable calibration control" disabled, the content will be sent to the display as is, and the colors will be slightly incorrect. When using "this display is already calibrated to BT.709", madVR will convert the NTSC colors to look correct on a BT.709 calibrated display.

So basically the recommendation is the following: If your display comes close to BT.709 (either by factory, or by manual calibration), then you should use "this display is already calibrated". If your display is not calibrated at all, then the colors you'll see are likely to be incorrect in any case. In that situation you could just as well use "disable calibration control" because there's no way to achieve correct colors, in any case. The difference between a properly calibrated display and a non-calibrated display is probably much bigger than the difference between BT.709 and SMPTE-C.

Originally Posted by nanpan

I use Windows 10 Pro x64(10586.29)+AMD Fury X.
When I update VGA driver to new driver【radeon-crimson-15.12-win10-64bit】

The NNEDI3 in madVR(chroma upscaling or image doubling)all could not work.
Maybe the madVR NNEDI3 OpenCL kernel something wrong in new VGA driver.

I always use SVP4 Pro+madVR together.
SmoothVideo Project SVP(3.17a and 4 pro)has the same error when using GPU-acceleration(OpneCL).

Any idea to fix it ?

If there is no way to fix it, I just only go back to old VGA driver.
Because old VGA driver(not crimson)all works fine.

I'm not sure why it fails. Probably a GPU driver problem. Nobody else reported this specific problem with Crimson, though. So it might be worth a try simply reinstalling the very same driver. Maybe the OpenCL driver didn't get installed properly or something. If all else fails, you may have to go back to an older driver.

Originally Posted by djsolidsnake86

0.89.18 playback is stopped (only audio work) with a cloned 2nd monitor
reverting to .17 is fine

Need more details. Does the madVR debug OSD (Ctrl+J) still work in this situation? Do the media player buttons and menus still react to your mouse in this situation? Does it make any difference if you disable/enable D3D11 presentation and/or FSE mode?

Originally Posted by CarlosCaco

if i play a 10 bit content on my 8 bit display, i'm loosing some color information?

No. madVR uses dithering to automatically convert higher bitdepths to lower bitdepths. Dithering works very well, the only disadvantage is a raised noise floor.

Originally Posted by Thunderbolt8

Im a bit confused by the terms 8-bit, 10-bit etc. in relation to the number of colours. my monitor has 16.7 million colours, but I dont know if that is 8-bit or 10-bit.

8bit = 2^8 = 256 steps per channel.
10bit = 2^10 = 1024 steps per channel.

If each RGB channel (R, G and B) has 8bit, that is 256 * 256 * 256 = 16.777216 million colors.

[madshi]

Originally Posted by hubblec4

I have posted in the LAV thread an issue.

With the new madVR0.89.17 play the DVD Aliens(test dvd without the big-vobs) with 4:3 format (menu and video), but it should be 16:9. With EVR it plays fine with 16:9 format.

It seems to play fine as 16:9 on my PC. But I always get to the "Deutschland / Österreich" copyright screen, not any further, due to missing VOBs. Which media player are you using? And which aspect ratio settings in the media player?

Originally Posted by Nullack

When using the current madVR on a 4K UHD display with high DPI say 300%, the madVR CTRL J renderer stats are too small to read.

OSD font size will scale with the rendering window size in the next build.

Originally Posted by Aktan

I think it's more like the madVR isn't requesting the frames from the capture source. Here is the screenshot of Ctrl+J in Potplayer using the capture source: http://i.imgur.com/8KwExdw.jpg and here is the screenshot of Ctrl+J in Potplayer using a captured AVI from the same capture source: http://i.imgur.com/7anTyFl.jpg.

According to the screenshots the decoder isn't delivering the frames to madVR fast enough. madVR is not "requesting" frames. That's not how it works. The decoder should be sending frames as fast as it can decode them. The decoder has the active role here. It's madVR which has the passive role. And if madVR doesn't receive decoded frames fast enough, obviously all queues stay empty.

Have you tried different decoders?

Originally Posted by retrue

Hi
I think I have found a bug. I am using the last development version of MPC-HC with the last version of XYSubFilter and the last version of MadVR
The bug happens in MadVR plus XYSubFilter when displaying upscaled VobSub subtitles.
It is an anime episode, 848x480 internal resolution, MPEG4 (H264) video, softcoded VobSub subtitles.
When displaying the vid at 1x resolution the subtitles are correctly displayed at the bottom of the screen.
When the vid is at 50% resolution (in window mode) it doesn't show subtitles at all.
If the vid is upscaled, resolution 2x (window mode) or it is upscaled fullscreen at 1920x1080 the subtitles appear on the left and in the middle of the screen and the subtitles themselves are not upscaled (they appear with the same size than in the resolution 848x480).
MadVR with other subtitles filters (VSFilter and Internal MPCHC filter) display correctly the subtitles upscaled and the bottom of the screen.
The MPCHC Personalized Enhanced Video Renderer with XYSubFilter shows the subtitles correctly too.
You can see screenshots of the vid here with the subtitles. In all these screenshots I am using MPCHC, MadVR and XYSubFilter.
If you need more info, please, tell me.
http://www72.zippyshare.com/v/8bgIjJfZ/file.html

Hmmmm... I suppose you're using DXVA upscaling, is that correct? Can you confirm that this problem goes away if you use e.g. Catmull-Rom upscaling? If the problem still occurs with Catmull-Rom upscaling, can you create a small sample for me, please, so that I can reproduce the problem on my PC? If the problem is fixed by using Catmull-Rom, then I don't need a sample.

Originally Posted by KoD

- movie file here (35Mb)
- settings files here; when everything is fine, jinc is used for resizing; when it's not fine, DXVA is used for resizing
- screenshots here, for how it looks when DXVA resizing is used, in windowed and fullscreen modes, because of deinterlacing performed by madVR

Thanks, I can reproduce the problem. Will try to fix it for the next build.

Originally Posted by KoD

- VIDEO_TS folder here (67Mb)

This seems to work just fine on my PC. The intro is shown as 4:3, the menu switches to 16:9. Which media player are you using? And which aspect ratio settings?

[6233638]

Originally Posted by madshi

This is unlikely to happen before we get displays which can do really high FPS. Maybe in 2-3 years when we get fast OLED displays, we can revisit BFI. At the moment most displays top out at 60Hz, which makes BFI impossible without major flickering.

There are a lot of monitors out there which support 72Hz inputs. Though it would flicker, most Plasma TVs ran 24p inputs at 48Hz, which would be 72Hz with 2 video frames for every 1 black frame. The question is not "will it flicker?" but rather "do I prefer motion clarity over flicker?" Growing up with 50Hz CRTs, I don't mind flicker one bit.

And higher refresh rates are becoming increasingly common. There's an ever-increasing number of TVs which support 120Hz inputs now, and I don't know anyone that's bought a new monitor in the last few years which wasn't 120-144Hz - and the latest models now offer 165-200Hz. (and that 200Hz monitor is a 21:9 "ultrawide", which seems to be gaining popularity)

It's less common with higher-end home theater equipment like projectors - though there are some 120Hz DLPs, but there should be 4K120 displays on sale next year. I just hope that LG or Samsung OLEDs are a part of that.

[madshi]

madVR is more meant for big TVs and projectors. I'm not going to invest time into BFI if it mostly only works on computer monitors. IMHO we need at least 96Hz before madVR BFI can start making sense, ideally 120Hz. The number of users with such TVs is probably very low atm. Furthermore there's no way for me to test this properly without having such a high-refresh-rate TV or monitor myself, which is currently not the case.

[e-t172]

Originally Posted by madshi

I've no idea if that's possible. And if it worked I've no idea if that would screw up things like calibration and stuff. Might be worth trying at some point, although I don't know if the OS/GPU driver even gives me access to DDC/DI. But this is not going to happen any time soon.

Do typical consumer HDTVs and projectors actually support DDC/CI, anyway? Even computer monitors rarely support it, except maybe professional ones.

[chros]

Originally Posted by ibius

But I doubt you will see a difference outputting 10-bit on a '8-bit + FRC' panel vs properly dithered madVR output on 8-bit panel.

On a side note, did anyone with a '8-bit + FRC' panel make some tests displaying 10-bit with madVR with dithering disabled to compare panel's dithering to that of madVR?
And, is madVR's Ordered dithering enough to smooth out the gradients of a dithered 10-bit on a 8-bit panel or is Error Diffusion required?

I agree with you. As it turned out I have one, 8bit+FRC, if it's true this is the one http://www.panelook.com/LC550WUD-SCA...iew_11833.html and it passes the 10bit gradient test ( (I had to enable FullRGBHDMI hack for the Intel driver, thanks madshi for this!)).
I'll take a good look and try to see the difference, but I didn't see at the first time.

Originally Posted by CarlosCaco

an relative old laptop with an disgusting optmus system,
after milions of tests i can assume that my display output 8 bit,

Honestly, I'd be surprised if it was 8 bit at all I vote for 6bi+FRC at most!!!
What's your laptop? And what string of the display can you see in madvr under Devices? (I need the full string)
Mine is: LG LP156WF1-TLF3, and panellook reports it: http://www.panelook.cn/LP156WF1-TLF3...iew_12274.html

[Stereodude]

Originally Posted by Aktan

I thought for broadcast, it be pure interlace anyway, not film video.

Lots of broadcast 1080i60 HTDV in the US is film based material that has gone through a telecine process. Things like the local news, or talk shows, no. But the primetime scripted shows are usually IVTC-able to 1080p24.

When you said looks terrible, what you mean. Lack of post processing or jittery movement?

Well, smooth motion was driving up the render times to higher than 1/23.976... Turning it off helped solved that issue, but I still can't quite put my finger on what it's doing. I think motion looks slightly different, but it might just be in my head. I guess I'd need to remove the soft pulldown and compare the playback of the file. My HTPC has an AMD R9 380 and a old Intel E5300 CPU. So no CUVID or Quicksync HW decoding options. Both the Radeon HW (copyback) and SW decoding leave a 3:2 pattern for madVR to find, which I don't understand if it's pure progressive output. However, I will agree that stepping through the video in slow motion doesn't show any interlaced or duplicated frames.

When I play it on another PC and use Quicksync as the HW decoder madVR only finds a 2:2 pattern which AFAIK means it is pure progressive. HW decoding with CUVID also leaves a 3:2 pattern for madVR to find, like SW and the Radeon HW unit.

[CarlosCaco]

Originally Posted by madshi

"disable calibration control" completely disables any gamma and gamut adjustments. If you set madVR to "this display is already calibrated", then depending on the video properties, madVR might do some adjustments or not.

Blu-Rays are usually encoded to BT.709 primaries. NTSC DVDs are usually encoded to SMPTE-C primaries. PAL DVDs are usually encoded to EBU/PAL primaries. In theory, in order to watch all these 3 different sources with correct colors, you'd have to calibrate your display 3 times with 3 different presets, one for each of those gamuts. I don't think anybody is doing that, not even professional calibrators.

Most displays these days try to get near to BT.709. Most calibrators aim to achieve BT.709, too. If your display is calibrated to give you correct colors for BT.709, then Blu-Rays will look correct in your setup, regardless of whether you set madVR to "disable calibration control" or "this display is already calibrated". However, if in this situation you play e.g. an NTSC DVD, there will be a difference. With "disable calibration control" disabled, the content will be sent to the display as is, and the colors will be slightly incorrect. When using "this display is already calibrated to BT.709", madVR will convert the NTSC colors to look correct on a BT.709 calibrated display.

So basically the recommendation is the following: If your display comes close to BT.709 (either by factory, or by manual calibration), then you should use "this display is already calibrated". If your display is not calibrated at all, then the colors you'll see are likely to be incorrect in any case. In that situation you could just as well use "disable calibration control" because there's no way to achieve correct colors, in any case. The difference between a properly calibrated display and a non-calibrated display is probably much bigger than the difference between BT.709 and SMPTE-C.

Testing the HDR new feature i can see cleary the diference, if i check disable calibration controls, madVR not do the conversion to the correct colour space of HDR content, the result are bad compared to this display already calibrated to bt 709

Do my calibration settings in dispcalgui i thought the correct was use the already calibrated to bt 709 with pure power curve transfer function

but testing the hdr content i only achieve goods resuls if i set the bt 709/601 curve

could someone explain the difference between the the two transfer functions?

There was a shader in sweetfx that simulated the appearence of HDR, was terrible because even in the minimal seetings the black were so darker e the withe so brighter that vanishes a lot of elements in the darkker/brighter areas

tweaking madVR i achieve the same results setting the transfer function calibration to bt 709/601 and setting gamma processing to pure power curves(terrible results, i was only messing the gama to see what happens), so i wondering what the difference of the transfer functions and in what cases should we use then?
iits possible to a tv or monitor were calibrated to bt 709 with pure power curve? what should correct?
in dispicalgui there´s option to use bt 709 tone curve or gama 2:20, what should the correct transfer function to work properly on madVR color conversions?

[CarlosCaco]

Originally Posted by chros

Honestly, I'd be surprised if it was 8 bit at all I vote for 6bi+FRC at most!!!
What's your laptop? And what string of the display can you see in madvr under Devices? (I need the full string)
Mine is: LG LP156WF1-TLF3, and panellook reports it: http://www.panelook.cn/LP156WF1-TLF3...iew_12274.html

Hey man, my laptop is a Dell Xps 15 l502x, and the display is a AU Optronics KYYVK€B156HW1 full hd
I serchead so many time for some information/specs about that, but didn not find anything relevant (i had hard times to do a correct calibration profile for use with madVR, i fid a lot of people in foruns complaining about the colours of this display, some people sayng that is wide gamut, so i searched a lot, but with no results, the colors are a bit saturated but i do know nothing about, found that are different versions of this panel that came with lenovo, acer and hp laptops, but in my case there is no documetation about, so i don´t if is v1, v2, v3,v4,v5

http://www.panelook.cn/rfqs_cn.php?p...2%82%ACB156HW1


tried to search AU Optronics KYYVK-B156HW1 too, because dispcalgui reports that way
just not informations at all, if you help me to findi will be very glad

edited

searched again and found but that are 8 versions on the site, thanks a lot! but i still don´t know what version is my...

http://www.panelook.cn/modelsearch.p...umber=B156HW01

[huhn]

Originally Posted by CarlosCaco

Testing the HDR new feature i can see cleary the diference, if i check disable calibration controls, madVR not do the conversion to the correct colour space of HDR content, the result are bad compared to this display already calibrated to bt 709

Do my calibration settings in dispcalgui i thought the correct was use the already calibrated to bt 709 with pure power curve transfer function

but testing the hdr content i only achieve goods resuls if i set the bt 709/601 curve

could someone explain the difference between the the two transfer functions?

There was a shader in sweetfx that simulated the appearence of HDR, was terrible because even in the minimal seetings the black were so darker e the withe so brighter that vanishes a lot of elements in the darkker/brighter areas

tweaking madVR i achieve the same results setting the transfer function calibration to bt 709/601 and setting gamma processing to pure power curves(terrible results, i was only messing the gama to see what happens), so i wondering what the difference of the transfer functions and in what cases should we use then?
iits possible to a tv or monitor were calibrated to bt 709 with pure power curve? what should correct?
in dispicalgui there´s option to use bt 709 tone curve or gama 2:20, what should the correct transfer function to work properly on madVR color conversions?

bt 709 has a pure gamma of 2.22 so bt 709/ bt 601 should be the same but the result is more like sRGB which is BT 709 with a different gamma function. this doesn't have an effect without "color or gamma -> enable gamma processing" at least for bt 709 content.

if you use this option for HDR than the results are different.
it's like madVR wants to take these settings and change it back to normal gamma 2.2.

that's why changing to BT 709/bt 601 curve is crushing black or at least rising the gamma and a higher gamma is lowering the gamma.
but "color & gamma -> enable gamma processing" does nothing with HDR content.

"correct" would be a pure gamma curve of 2.2.
if you are using DISPCAL GUI just think about an 3D LUT.

[Aktan]

Originally Posted by madshi

According to the screenshots the decoder isn't delivering the frames to madVR fast enough. madVR is not "requesting" frames. That's not how it works. The decoder should be sending frames as fast as it can decode them. The decoder has the active role here. It's madVR which has the passive role. And if madVR doesn't receive decoded frames fast enough, obviously all queues stay empty.

Have you tried different decoders?

This is raw YUY2 from an analog capture source (device). I don't think it can deliver frames faster than realtime as it captures in realtime. I thought a setting like "delay playback start until render queue is full" would fix the problem, but there is no delay. Did you look at the 2nd screenshot that shows decoding an AVI works fine? That has decoder delay since it is losslessly compressed, but it still works fine. Another note is if I pipe the raw YUY2 from FFMPEG encapsulated in an ASF container to MPC-HC, this problem does not happen. But then there is an obvious input delay which wouldn't work for playing games.

Edit: Now thinking about it, I can see why "delay playback start until render queue is full" would not fix the problem since it is a decoder queue problem. Would it be possible to add the option of something like "delay playback start under decoder queue is full"?

[seiyafan]

I have a computer which I use for amature photo editing at home and it's calibrated with i1 Display pro and basICColor (it generates an icm file which is loaded by Windows at startup) Just today I found out about dispcalGUI and Argyll CMS. Since I also watch movies on this computer, what's the advantage of dispcalGUI over my existing calibration besides the 3D LUT? Is the current movie playback with MadVR using the .icm of the existing calibration?

[Aktan]

Originally Posted by Stereodude

When I play it on another PC and use Quicksync as the HW decoder madVR only finds a 2:2 pattern which AFAIK means it is pure progressive. HW decoding with CUVID also leaves a 3:2 pattern for madVR to find, like SW and the Radeon HW unit.

I'm surprise it finds a 2:2 pattern at all as I thought that is usually used on PAL DVDs. But it is interesting that 1/3 of the time when I force FILM deinterlacing from madVR, I do get a 2:2 pattern as said by the OSD.

[retrue]

Originally Posted by madshi

Quote:
Originally Posted by retrue View Post
Hi
I think I have found a bug. I am using the last development version of MPC-HC with the last version of XYSubFilter and the last version of MadVR
The bug happens in MadVR plus XYSubFilter when displaying upscaled VobSub subtitles.
It is an anime episode, 848x480 internal resolution, MPEG4 (H264) video, softcoded VobSub subtitles.
When displaying the vid at 1x resolution the subtitles are correctly displayed at the bottom of the screen.
When the vid is at 50% resolution (in window mode) it doesn't show subtitles at all.
If the vid is upscaled, resolution 2x (window mode) or it is upscaled fullscreen at 1920x1080 the subtitles appear on the left and in the middle of the screen and the subtitles themselves are not upscaled (they appear with the same size than in the resolution 848x480).
MadVR with other subtitles filters (VSFilter and Internal MPCHC filter) display correctly the subtitles upscaled and the bottom of the screen.
The MPCHC Personalized Enhanced Video Renderer with XYSubFilter shows the subtitles correctly too.
You can see screenshots of the vid here with the subtitles. In all these screenshots I am using MPCHC, MadVR and XYSubFilter.
If you need more info, please, tell me.

Hmmmm... I suppose you're using DXVA upscaling, is that correct? Can you confirm that this problem goes away if you use e.g. Catmull-Rom upscaling? If the problem still occurs with Catmull-Rom upscaling, can you create a small sample for me, please, so that I can reproduce the problem on my PC? If the problem is fixed by using Catmull-Rom, then I don't need a sample.

At Madshi

Yes, I am using DXVA upscaling, but when I use any other upscaling with this video (using last versions of MPC-HC, XYSubFilter and MadVR) I don't get any image at all. The sound plays fine but the screen is black. With some of the upscaling methods I can see the info when typing Ctr+J, but in most cases I see nothing at all. All the upscaling methods work fine with other videos that I have tried.

How I can create the sample?
I took a sample with Avidemux but it ignored the subtitles. Then I extracte d the subtitles from the original video and I added them to the sample. But something is wrong. Selecting DXVA upscaling it was possible to reproduce the problem: subtitles not upscaling and appearing in a wrong position, but when I tried Catmul-Rom on the sample the computer crashed.