madVR - high quality video renderer (GPU assisted)

[ABDO]

I wonder if I should maybe remove the "quality" option and auto pick, based on the strength?


that would hurt my gtx 150 ti boost :D, i realy do not know.

[cork_OS]

Thanks for the feedback, it matches my own test results. I wonder if I should maybe remove the "quality" option and auto pick, based on the strength?
Please don't remove this option, RCA high is too slow for GTX 1060 and below.

[Werewolfy]

Thanks for the feedback, it matches my own test results. I wonder if I should maybe remove the "quality" option and auto pick, based on the strength?


Unfortunately, I don't think so because the GPU load is quite higher with high quality. Even on my Geforce GTX1080 I can see a difference so I supsect weaker GPU will struggle with this.

[Fabulist]

Some sharpeners bloat more than others. E.g. "crispen edges" doesn't bloat by design, but LumaSharpen does. Using anti-bloating removes the bloating, but that also reduces the overall sharpening effect, so you may have to increase the sharpening strength when using anti-bloating. Using 100% seems like a safe choice, but you can pick any value that looks good to your eyes.

Thank you for taking your time to reply. Does Sharpen Edges really do any bloating on 1080p+ sources? I mean objectively and in technical terms, by like a specific amount or %?

I run tests on various sets and sources and I am unable to see any kind of bloating while running Sharpen Edges at 4 (like other sharpeners do) on up to 65 inches. Is there something I am missing or am I not looking at what I should be looking at? Does it bloat in a different way other sharpeners do, one which I cannot identify, like a different kind of video distortion?

[Blackwalker]

Yell you could try but really it needs to be on and not off. If you have an AV receiver between the PC and TV you should also make sure that supports deep color or that the HDMI port your using supports it.

You could also try a different HDMI port of TV / AV receiver.

Hi, the htpc hdmi is connected directly to tv..!



My impression is that you have a tendency to ignore (or accidently miss) what people say. It has been suggested to you that you should use an older NVidia driver (e.g. 385.2x), but last time you reported your Nvidia driver version, it was still 387.xx or 388.xx. Also make sure you have the OS option "HDR and Advanced Color" turned off.



Thank you madshi,

Sorry.. i’m not ignoring, i miss it, again :(
I’ll try with 385.2x driver and check the “HDR and Advanced Color” tourned off.
Thx for your time and your work with this unique and fantastic toolb:thanks:

[Gopa]

Unfortunately, I don't think so because the GPU load is quite higher with high quality. Even on my Geforce GTX1080 I can see a difference so I supsect weaker GPU will struggle with this.

Agree. Usually, some "other" madVR enhancement compromises, with a GTX1070, using RCA high quality. Usually, fewer upscaling refinements &/or, a lower luma image upscaling setting, are required to use RCA high quality, with higher strength setting. RCA high quality, with average quality 720p and 1080p upscaled anime, is often, a very, very good madVR enhancement (sometimes, RCA medium quality, or reduce random noise, is just enough). Maybe, something between, medium & high quality RCA, might be nice, if possible? Lower GPU load? Madashi, says, I will love, the next madVR version, because of my preference for RCA high quality!

[yukinok25]

It looks like a crash in the Nvidia OpenCL driver. Try using NGU-Anti-Alias instead of NNEDI3. That should fix the problem. NGU AA looks better than NNEDI3 IMHO, anyway, and is faster at the same time.

Thank you madshi.

It's strange thus, I am actually using NGU-sharp at medium as a chroma, why the crash would report NNEDI3?

It's because I left the option "Let MadVR decide" on somewhere?

[Gopa]

newbie questions: I have a 65" 4k TV & so, upscaling 720p anime, results in a heavy GPU load - direct x4 luma upscaling to 2880p & then, downscaling to 2160p. Why is it not possible to just upscale 720p to 2160p (x3 instead of x4)? Is this just not possible, at all, or a limitation of madVR image upscaling? I think, that my Samsung TV, does this (direct x3 upscaling of 720p?), although the difference in upscaling quality, is significantly lower, compared to madVR upscaling + downscaling (or maybe, my TV upscales x4 & then downscales, like madVR?). I prefer 720p anime (original resolution), over upscaled 1080p, because of madVR 4x upscaling, so pre-upscaling anime to 1080p, is not the best option, unless I could "improve" the quality (fewer artifacts/sharper image, without overall quality loss). Any suggestions, other than upgrading my graphics card? And, yeah, I am simply striving for perfection: A GTX1070, is almost perfect, for what I am viewing. Very pleased with madVR - especially RCA high quality & luma image upscaling - limitation on using image refinements is a fairly minor issue (a little sharper image would be nice).

[thighhighs]

NGU soft\standard\sharp can upscale 2x and direct 4x, 3x not supported. Also no direct quadrupling for NGU AA, it work only in doubling twice mode.
Last i'm remember about NGU 3x variant, NGU 3x speed\quality ratio not good vs NGU 4x + downscaling. But i hope NGU 3x not dead.
These features is what exactly i would like to see in madvr:

- NGU direct 3x
- NGU AA direct 4x, 3x
- Also better SLI support: a) SLI-friendly madvr without negative scaling b) extra performance with second GPU (just dream :))

[Gopa]

NGU soft\standard\sharp can upscale 2x and direct 4x, 3x not supported. Also no direct quadrupling for NGU AA, it work only in doubling twice mode.
Last i'm remember about NGU 3x variant, NGU 3x speed\quality ratio not good vs NGU 4x + downscaling. But i hope NGU 3x not dead.
These features is what exactly i would like to see in madvr:

- NGU direct 3x
- NGU AA direct 4x, 3x
- Also better SLI support: a) SLI-friendly madvr without negative scaling b) extra performance with second GPU (just dream :))

GTX1070 – average quality 720p anime: RCA high quality, including strength 10 + chroma ngu sharp low/medium + luma ngu sharp direct quadruple high + this compromise: refine image only once, instead of after every 2x upscaling (thin edges, soften edges, luma sharpen, anti-bloating, anti-ringing) + a few low strength image enhancement sharpeners, if more sharpening is still needed. Refine image only once, is not quite as smooth as after every 2x. Otherwise, this seems to be a pretty good compromise, to use RCA high quality, along with ngu high x4 luma upscaling. Images look great (to me). Likely, no compromises, at all, with a GTX1070ti, or GTX1080? Better sources, would mean, no need for RCA high quality, of course & RCA medium quality, or reduce random noise, use a lot less rendertime. RCA high quality + ngu sharp 4x high luma, looks better than, ngu sharp very high luma, without RCA high quality (to me). Most likely, because of less than great source material. Upscaling 720p to 4k, without madVR, is not something, that I care to do, ever again! Thanks for the information, about past attempts with 3x upscaling. I once thought, that SLI, would be the answer: a more powerful graphics card, makes more sense to me, now, unless also using SLI, for gaming. Very happy with a GTX1070. Yeah, just a little bit more power, would still be nice! All of this is because I chose to get a 4k TV, instead of a 1080p TV (better overall features, with 4k TV’s). Not so sure, this was the right decision, though. A 65” 1440p monitor, might be the best, if any existed. I like sitting pretty close & this is not possible, with a 1080p monitor. Always open to suggestions for alternative settings.

[heiseikiseki]

As mentioned by heiseikiseki, Intel seems to plan to replace their internal GPUs with AMD soon, so maybe Intel NUCs might do the job then. But the GPU chip alone won't be enough, we also need fast VRAM, and ideally 4GB+ for 4K.


some of Kabylake-G use 4GB HBM2 memory for the GPU
so I think it would be ok for 4k.

Here are some bench marks information leaked
https://wccftech.com/intel-kaby-lake-g-series-integrated-radeon-gpus-first-benchmarks-specifications/

The performance is about 3.3 TFLOPs.

Seems at least much much stronger than GTX750TI or GTX960.

[ryrynz]

Good testing. But annoying test results, because the source code changes between 0.91.11 and 0.92.1 are pretty large.

What is the easiest way to reproduce this problem?


Play anything, pause, change screen power off to 1 min and wait. Let screen blank then wait a bit, most times the video but not audio playback is frozen upon resuming.
For anyone else having this issue (is it Nvidia only?) try setting LAV to software decoding (Hardware decoding "none") until a fix is found.

[huhn]

some of Kabylake-G use 4GB HBM2 memory for the GPU
so I think it would be ok for 4k.

Here are some bench marks information leaked
https://wccftech.com/intel-kaby-lake-g-series-integrated-radeon-gpus-first-benchmarks-specifications/

The performance is about 3.3 TFLOPs.

Seems at least much much stronger than GTX750TI or GTX960.

it is still not clear if the IGPU get's replaced by the amd GPU or if both are on the chip. i don't have to remind anyone here how terrible nvidia "optimus" is.

and most important these chip are going to be expensive.
HBM2 is very expensive.
it a faster 560 which is decent no question. but really wait until they are out there is so much missing informations.

[heiseikiseki]

Hello mashi,
I have some small problem confused me a long time with the osd which I don't know is it a bug of madvr or not.
I thought that is caused by my computer is optimus spec.
But recently I buy a new desktop with GTX1060 the problem is still not gone.

Here are some screenshot.

First I set my presented frames for 16
https://i.imgur.com/YVBXBb8.png

And here is the OSD
https://i.imgur.com/DX4UFwC.png

the decoder queue and upload queue seems work well is about 17-18 (I set 18)

Some questions here.

1. The render queue always not reach 18.
2. The present queue I set it 16 ,but the maximum here shows 15. (if I sent N, it would be N-1)
3. The Hardware Decoder I set D3D11 on LAV , but the OSD shows DXVA11.

Is it all normal?

[Nyago123]

My notes on nVidia driver 388.31 + madVR 0.92.9 & playing HDR on a Win10 1709+LG E6+GTX 1080+Denon X4300H AVR:

Originally after the Fall Creator's update, the default Windows-included nVidia drivers wouldn't set HDR for me (MPC-BE). With the reports here on the problems, I went with a 385.xx version which was OK.

But I decided to play around with the most recent driver release (388.31). What I found is:

1. HDR works correctly with Zoom Player - it kicks in when starting the player, and disables when the player exits at least 90% of the time. One time it started with incorrect colors and one time the player hung. For the former, I just double toggled HDR in the Win 10 Display Settings; for the latter I just killed the process and tried again (to be honest, ZP freezes from time to time on me in general - nothing to do with HDR - so I can't tell if this was just "one of those times").

2. HDR works with MPC-HC but stays on when the player exits. I have to use the Win 10 Display Settings dialog and toggle on/off to get back to SDR.

3. HDR doesn't start at all with MPC-BE

I should probably spend a little more time looking at how LAV might be playing into this, but for now Zoom Player is my player of choice for HDR content until nVidia+MS can get this fixed.

[chros]

For SDR screenshots, please compare F5 with PrintScreen. Is there a difference? You can configure F5 behaviour in the new madVR "screenshots" settings page.
Yes, there is, using F5 for different HDR->SDR conversion:
- pixel shader conversion is OK
- 3dlut conversion is Not (image is much darker)

PrintScreen in Windowed mode works fine for both of them.

[madshi]

that would hurt my gtx 150 ti boost :D, i realy do not know.
Please don't remove this option, RCA high is too slow for GTX 1060 and below.
Unfortunately, I don't think so because the GPU load is quite higher with high quality. Even on my Geforce GTX1080 I can see a difference so I supsect weaker GPU will struggle with this.
Ok! :)

Thank you for taking your time to reply. Does Sharpen Edges really do any bloating on 1080p+ sources? I mean objectively and in technical terms, by like a specific amount or %?

I run tests on various sets and sources and I am unable to see any kind of bloating while running Sharpen Edges at 4 (like other sharpeners do) on up to 65 inches. Is there something I am missing or am I not looking at what I should be looking at? Does it bloat in a different way other sharpeners do, one which I cannot identify, like a different kind of video distortion?
Sharpen Edges internally does some supersampling which reduces bloating. It's all somewhat subjective, please trust your eyes and pick values that look good to you. This is all not really scientific. If you want scientific, you'd have to use deconvolution instead of sharpening, but even then, which deconvolution kernel would you be using. Gaussian or other? Linear light or gamma light?

I’ll try with 385.2x driver and check the “HDR and Advanced Color” tourned off.
Hope it will work for you!

It's strange thus, I am actually using NGU-sharp at medium as a chroma, why the crash would report NNEDI3?

It's because I left the option "Let MadVR decide" on somewhere?
Let madVR decide will never switch to NNEDI3 if you haven't selected it. In your bug report there were 2 crashes. One was not clear. The other one pointed to a crash in the Nvidia OpenCL driver. madVR uses OpenCL only for NNEDI3, so my conclusion was that you probably used NNEDI3. Please check the OSD (Ctrl+J) to confirm that NNEDI3 is really not used, neither in chroma upscaling nor image upscaling. If you're really not using NNEDI3, then I'm at a loss why the Nvidia OpenCL driver crashed!

newbie questions: I have a 65" 4k TV & so, upscaling 720p anime, results in a heavy GPU load - direct x4 luma upscaling to 2880p & then, downscaling to 2160p. Why is it not possible to just upscale 720p to 2160p (x3 instead of x4)?
It's possible, but difficult to do, and I'm not sure it would actually produce competetive results. It might look worse than x4 + downscaling. Furthermore, it would probably be only barely faster than x4 upscaling + downscaling. Anyway, x3 is still on my list of things to look at, but it's a low priority atm, because I doubt its usefulness.

some of Kabylake-G use 4GB HBM2 memory for the GPU
so I think it would be ok for 4k.
Yes, that looks promising, actually! We'll have to wait and see how it works in reality, but it *could* be suitable for madVR!

it is still not clear if the IGPU get's replaced by the amd GPU or if both are on the chip. i don't have to remind anyone here how terrible nvidia "optimus" is.
FWIW, the key problem with Optimus is that there are really 2 totally separate GPUs, each of which has their own HDMI/DVI driver inside. And what is worse: The actual driving of the display (and HDMI output) is done by the Intel GPU, while the rendering is done by the Nvidia GPU, so the 2 GPUs have to work together. Basically Nvidia has to transport the rendered frames to the Intel GPU, to be sent to the display. Practically this means, both drivers have to work together. That's like driver bugs ^ 2!

If the new Intel CPUs with integrated AMD GPUs only have one GPU instead of two, most (or even all) of the Optimus problems should not occur.

Play anything, pause, change screen power off to 1 min and wait. Let screen blank then wait a bit, most times the video but not audio playback is frozen upon resuming.
For anyone else having this issue (is it Nvidia only?) try setting LAV to software decoding (Hardware decoding "none") until a fix is found.
So it only occurs with hardware decoding? Does it occur with DXVA2 copyback *and* native? How about D3D11?

Hello mashi,
I have some small problem confused me a long time with the osd which I don't know is it a bug of madvr or not.
I thought that is caused by my computer is optimus spec.
But recently I buy a new desktop with GTX1060 the problem is still not gone.

the decoder queue and upload queue seems work well is about 17-18 (I set 18)

Some questions here.

1. The render queue always not reach 18.
2. The present queue I set it 16 ,but the maximum here shows 15. (if I sent N, it would be N-1)
3. The Hardware Decoder I set D3D11 on LAV , but the OSD shows DXVA11.

Is it all normal?
That's all perfectly normal. The queues don't have to be 100% full. As long as they're nearly full, everything's fine. The key thing to look at is that you don't get frame drops/repeats or presentation glitches increasing all the time during playback.

My notes on nVidia driver 388.31 + madVR 0.92.9 & playing HDR on a Win10 1709+LG E6+GTX 1080+Denon X4300H AVR:

Originally after the Fall Creator's update, the default Windows-included nVidia drivers wouldn't set HDR for me (MPC-BE). With the reports here on the problems, I went with a 385.xx version which was OK.

But I decided to play around with the most recent driver release (388.31). What I found is:

1. HDR works correctly with Zoom Player - it kicks in when starting the player, and disables when the player exits at least 90% of the time. One time it started with incorrect colors and one time the player hung. For the former, I just double toggled HDR in the Win 10 Display Settings; for the latter I just killed the process and tried again (to be honest, ZP freezes from time to time on me in general - nothing to do with HDR - so I can't tell if this was just "one of those times").

2. HDR works with MPC-HC but stays on when the player exits. I have to use the Win 10 Display Settings dialog and toggle on/off to get back to SDR.

3. HDR doesn't start at all with MPC-BE

I should probably spend a little more time looking at how LAV might be playing into this, but for now Zoom Player is my player of choice for HDR content until nVidia+MS can get this fixed.
Interesting. But why not stick to 385.xx until the issue if fixed?

Yes, there is, using F5 for different HDR->SDR conversion:
- pixel shader conversion is OK
- 3dlut conversion is Not (image is much darker)

PrintScreen in Windowed mode works fine for both of them.
Can you give me a few more details. How is 3dlut conversion not ok? Is the 3dlut not applied for the screenshot? Or is it applied incorrectly? What happens if you disable the 3dlut? Are F5 and PrintScreen identical then for SDR movies?

[ryrynz]

So it only occurs with hardware decoding? Does it occur with DXVA2 copyback *and* native? How about D3D11?


Occurs with all of those, haven't been able to reproduce on Intel.

[madshi]

In case you guys wonder how NGU Sharp compares to mpv's latest FSRCNN(X), here's a little comparison:

Blu-Ray screenshot (http://madVR.com/doom9/spiritedAway/HiRes.png)
downscaled (PNG) (http://madVR.com/doom9/spiritedAway/LowRes.png) | (JPG, 100 quality) (http://madVR.com/doom9/spiritedAway/LowRes.jpg)
latest FSRCNN32 (http://madVR.com/doom9/spiritedAway/FSRCNN32.png)
latest FSRCNNX32 (http://madVR.com/doom9/spiritedAway/FSRCNNX32.png)
NGU Sharp - High (http://madVR.com/doom9/spiritedAway/NGUSharpHigh.png)
NGU Sharp - Very High (http://madVR.com/doom9/spiritedAway/NGUSharpVeryHigh.png)

To make things as fair as possible I've downscaled the image with Bicubic/Catrom, which is exactly what FSRCNN and FSRCNNX were trained for.

Here are benchmark numbers, for 720p doubling:

Nvidia 1070:
FSRCNN16: 15.270 ms
FSRCNNX16: 26.397 ms
FSRCNN32: 46.290 ms
FSRCNNX32: ? (estimated: 80.021 ms)
NGU-Sharp High: 3.940 ms
NGU-Sharp Very High: 11.800 ms

AMD 560:
FSRCNN16: 14.289 ms
FSRCNNX16: 24.412 ms
FSRCNN32: 45.235 ms
FSRCNNX32: ? (estimated: 77.282 ms)
NGU-Sharp High: 12.970 ms
NGU-Sharp Very High: 37.100 ms

These are very weird benchmark results, to say the least. We know that NGU doesn't run as well as it should on AMD Polaris GPUs. But FSRCNN(X) running (ever so slighty) faster on my AMD 560 than on my Nvidia 1070 is just plain weird.

[ABDO]

In case you guys wonder how NGU Sharp compares to mpv's latest FSRCNN(X)


well, no Surprise here, NGU Sharp look, best, crisp and closer to
Blu-Ray.

edit
i see that FSRCNN bring some colours shift.

[thighhighs]

Always open to suggestions for alternative settings.
What you think about NGU Sharp without any extra sharpness? IMHO:
NGU Sharp look very sharp by default, edges looks thin and sharp in focus. Another situation when you use NGU AA for image upscaling. NGU AA + post sharp can be better than NGU Sharp for low quality sources, because NGU AA can fix some edges (https://imgur.com/LBEncfq), where NGU Sharp\Standard just sharp it (https://imgur.com/OZvsRSA).

Try something like this:

Artifact removal - Debanding: Low\Low or med\med
Artifact removal - Deringing: Off
Artifact removal - RCA: What you like
Image enhancements: Off
Chroma: NGU AA High
Image doubling: NGU Sharp High (looks like you prefer sharp image)
Chroma doubling: Off\ let madvr decide
Upscaling refinement: Off
Dithering: Orderer

edit: well, no Surprise here, NGU Sharp look, best, crisp and closer to
Blu-Ray +

[huhn]

In case you guys wonder how NGU Sharp compares to mpv's latest FSRCNN(X), here's a little comparison:

Blu-Ray screenshot (http://madVR.com/doom9/spiritedAway/HiRes.png)
downscaled (PNG) (http://madVR.com/doom9/spiritedAway/LowRes.png) | (JPG, 100 quality) (http://madVR.com/doom9/spiritedAway/LowRes.jpg)
latest FSRCNN32 (http://madVR.com/doom9/spiritedAway/FSRCNN32.png)
latest FSRCNNX32 (http://madVR.com/doom9/spiritedAway/FSRCNNX32.png)
NGU Sharp - High (http://madVR.com/doom9/spiritedAway/NGUSharpHigh.png)
NGU Sharp - Very High (http://madVR.com/doom9/spiritedAway/NGUSharpVeryHigh.png)

To make things as fair as possible I've downscaled the image with Bicubic/Catrom, which is exactly what FSRCNN and FSRCNNX were trained for.

are you sure everything is correct i mean FSRCNN has chroma bleeding...

but in term of anti alasing (ignoring chroma) i give FSRCNN more points but i would nearly for sure use NGU AA on this source.
i'm not a fan of NGU sharp on anime in general lines can look really strange.

[madshi]

are you sure everything is correct i mean FSRCNN has chroma bleeding...
Well, I don't know how to activate FSRCNN(X) for chroma doubling. I don't know if it's possible. If anyone knows, let me know and I'll redo the FSRCNN(X) screenshots.

but in term of anti alasing (ignoring chroma) i give FSRCNN more points but i would nearly for sure use NGU AA on this source.
i'm not a fan of NGU sharp on anime in general lines can look really strange.
Can you point me to where in the image exactly you see FSRCNN having lower aliasing than NGU Sharp?

[huhn]

FWIW, the key problem with Optimus is that there are really 2 totally separate GPUs, each of which has their own HDMI/DVI driver inside. And what is worse: The actual driving of the display (and HDMI output) is done by the Intel GPU, while the rendering is done by the Nvidia GPU, so the 2 GPUs have to work together. Basically Nvidia has to transport the rendered frames to the Intel GPU, to be sent to the display. Practically this means, both drivers have to work together. That's like driver bugs ^ 2!

If the new Intel CPUs with integrated AMD GPUs only have one GPU instead of two, most (or even all) of the Optimus problems should not occur.

using the intel iGPU for desktop and co makes sadly a lot of sense. the hole system is clearly made for battery time.
i mean the price of HBM is about 3x GDDR5 and currently the major benfit of HBM is powerconsumption.
AMD evne has a zero power GHPU function and not using this woudl be quite a waste.
i'm just guessing here so whatever will take a some time until we see these devices anyway.

These are very weird benchmark results, to say the least. We know that NGU doesn't run as well as it should on AMD Polaris GPUs. But FSRCNN(X) running (ever so slighty) faster on my AMD 560 than on my Nvidia 1070 is just plain weird.

do they use openCL.
i eman even with the handbrake copyback issues on AMD side these cards where usually still a lot faster than nvidia cards back in the nnedi 3 days.

[huhn]

Can you point me to where in the image exactly you see FSRCNN having lower aliasing than NGU Sharp?

the bigs babys head for example.

the red hat top center (where FSRCNN has the huge chroma issue)

BTW. the chroma issue is visible on the JPG screen.

[Neo-XP]

Well, LL on/off sharpen different kinds of edges with different strengths, so in some parts of the image LL off will sharpen more than LL on, and vice versa. I'm surprised you got halos, though. Do you happen to have a couple good test images where I can see why you prefer LL off? That might help with development.

A quick comparison with The Neon Demon ( Bluray / UHD Bluray ):

https://thumb.ibb.co/jeysJR/The_Neon_Demon_Bluray.png (https://image.ibb.co/mM06dR/The_Neon_Demon_Bluray.png) https://thumb.ibb.co/hmKck6/The_Neon_Demon_Bluray_UHD.png (https://image.ibb.co/jTMq56/The_Neon_Demon_Bluray_UHD.png)

Let's try to upscale the Bluray ( NGU very high / NGU very high + AS 0.5 / NGU very high + AS 0.5 LL ):

https://thumb.ibb.co/crZMCm/The_Neon_Demon_NGUvery_High.png (https://image.ibb.co/fLuMCm/The_Neon_Demon_NGUvery_High.png) https://thumb.ibb.co/fBQ6dR/The_Neon_Demon_NGUvery_High_AS05.png (https://image.ibb.co/hVJRdR/The_Neon_Demon_NGUvery_High_AS05.png) https://thumb.ibb.co/jjgCJR/The_Neon_Demon_NGUvery_High_AS05_LL.png (https://image.ibb.co/mAHV56/The_Neon_Demon_NGUvery_High_AS05_LL.png)

Zoomed (AS with LL off and on) : http://screenshotcomparison.com/comparison/124629

I prefer with LL off, because it is sharper for the same value. Higher values for LL only accentuate artifacts around the edges so it is not fair to compare with different values.

The old NGU is also closer to the UHD Bluray :

https://thumb.ibb.co/enAWdR/The_Neon_Demon_Old_NGUvery_High.png (https://image.ibb.co/eeSYsm/The_Neon_Demon_Old_NGUvery_High.png)

Zoomed : http://screenshotcomparison.com/comparison/124639

Old NGU looks a lot better also for me for this movie, but I do not know if it was downscaled with LL or not.

[madshi]

are you sure everything is correct i mean FSRCNN has chroma bleeding...
Updated mpv screenshots, they have less chroma bleeding now.

do they use openCL.
No, OpenGL/GLSL.

[madshi]

the bigs babys head for example.

the red hat top center (where FSRCNN has the huge chroma issue)
I'm sorry, I'm having trouble seeing it. Where do you see aliasing there? Do you mean the top line of the big baby's head? If so, there's some minor aliasing there, but it's also in the Hi-Res image! Shouldn't you give NGU bonus points for reproducing the Hi-Res image more faithfully, warts and all?

Talking about the big baby's head: How about the more obvious differences, like the big baby's mouth, the eye brows, the nose?

A quick comparison with The Neon Demon ( Bluray / UHD Bluray )

Old NGU looks a lot better also for me for this movie, but I do not know if it was downscaled with LL or not.
Thanks!

[leeperry]

Is that 4x NGU Sharp in one step, followed by SR? Or is it 2X NGU Sharp + SR + 2x NGU Sharp + SR?
Direct 4X NGU sharp > SR@2 LL for both luma & chroma > SSIM 2D 100% AB25%, hard to believe how good VCD looks now with all the bells and whistles enabled huh!

I still wish you would properly align chroma for MPEG-1 especially as nevcariel said that he was willing to cooperate but it still looks great when outputting properly aligned RGB32 off ffdshow so no biggy I guess :o

I wonder if I should maybe remove the "quality" option and auto pick, based on the strength?
Please don't, I can afford "high" on some videos but not on others and I have some blocky ones that actually appear to look better in "medium". I see "high" as the icing on the cake if I have GPU cycles to spare and PQ improves.

For the record I only use RCA@1, anything higher is too strong to my taste.

And somewhat unrelated to mVR but as much as we care about shielding of A/V cables, it seems that power cables are vastly overlooked and a triple-shielded garden-hose one has tremendously improved sharpness on my Sammy TV. Identical cable actually did the same on my G3-550 powered PC but then again I pickup +40 wifi networks here so not having them interfering with my A/V gear does appear to do the magic :eek:

[huhn]

I'm sorry, I'm having trouble seeing it. Where do you see aliasing there? Do you mean the top line of the big baby's head? that line If so, there's some minor aliasing there, but it's also in the Hi-Res image! Shouldn't you give NGU bonus points for reproducing the Hi-Res image more faithfully, warts and all?
well that topic again.
well i guess we want the image close to the original but one bonus of upscaling was anti alaising. NGU is just so sharp that it is hard to totally avoid alaising.
Talking about the big baby's head: How about the more obvious differences, like the big baby's mouth, the eye brows, the nose?

NGU is generally sharper and the line thickness is closer to the original.
but i just gave FSRCNN(and i through waifux2 is a bad name...) more points for anti aliasing and nothing else and the new image isn't that much better now the chroma bleeding was hiding the aliasing quite a bit. but the thin lines from NGU have a bad part to at the bottom line of the red hat top center. the line is nearly completely gone now while FSRCNN makes it relative big now but the source has this line.

is there a reason you have to use a JPG with mpv don't tell me they don't support PNG? but why else should you add a jpg and a PNG as source...

[Soulnight]

I agree. That would be great to make the option "command line to execute when this profile is activated/deactivated "works

Madshi: any idea when you could make this function work? :thanks:

http://bugs.madshi.net/view.php?id=210

[Fabulist]

Sharpen Edges internally does some supersampling which reduces bloating. It's all somewhat subjective, please trust your eyes and pick values that look good to you. This is all not really scientific. If you want scientific, you'd have to use deconvolution instead of sharpening, but even then, which deconvolution kernel would you be using. Gaussian or other? Linear light or gamma light?

I understand, thank you!

[madshi]

well i guess we want the image close to the original but one bonus of upscaling was anti alaising.
Yes. The problem is that any change you do to an algorithm usually improves something, but makes something else worse. So if I designed NGU to intentionally modify the image to reduce aliasing, it would also result in a softer look and loss of detail at the same time. The whole design approach of NGU Sharp is that it tries to get as close to the original Hi-Res image as possible. And if the original Hi-Res image has aliasing, then NGU Sharp unfortunately sometimes also reproduces that aliasing faithfully.

That said, most aliasing that was introduced through downscaling is removed by NGU Sharp very effectively. Which all the conventional algos like Bicubic, Lanczos or even Jinc don't do.

NGU is generally sharper and the line thickness is closer to the original.
And NGU is virtually ringing free, while FSRCNN(X) have visible ringing.

but the thin lines from NGU have a bad part to at the bottom line of the red hat top center. the line is nearly completely gone now while FSRCNN makes it relative big now but the source has this line.
Yes. Unless one algo totally sucks, when comparing 2 different algos you'll almost always find something which algo 1 does better than algo 2 and vice versa. The important thing is which algo does better overall, of course.

is there a reason you have to use a JPG with mpv don't tell me they don't support PNG? but why else should you add a jpg and a PNG as source...
mpv does support PNG, but it doesn't activate FSRCNN(X) when I try to "play" a PNG or BMP file. I'm not completely sure why, maybe the source has to be YCbCr instead of RGB, I don't know. It's weird, but that's the way it is. I've used the highest possible JPG quality level, though, which is without any visible compression artifacts, even when zooming in. So it should be ok. I would have compared with PNG, if I could, of course.

[huhn]

Yes. Unless one algo totally sucks, when comparing 2 different algos you'll almost always find something which algo 1 does better than algo 2 and vice versa. The important thing is which algo does better overall, of course.

i noticed in general that NGU sharp has a tendency to thin really thin lines a little bit to much.

and it looks like NGU sharp has problems with aliased sources too not only "bad" sources.

And NGU is virtually ringing free, while FSRCNN(X) have visible ringing.

if you mean the holoing around lines and discolouring of small image parts than yes i can see this ringing too else i don't really notice it.

another issue i notice with FSRCNN is the "noise" reproduction.
this old source has some colored noise in it and this noise is sometimes reproduced in a checkboard pattern which is really ugly.

[Patrik G]

madVR does what you ask it to do. If you use "convert HDR content to SDR by using pixel shader math", and if you also set the "calibration" tab to "this display is already calibrated", then madVR will convert the HDR content to the gamut and transfer function you selected in the calibration tab. That can be BT.709 or DCI-P3 or BT.2020, whatever you choose.


thanks!
it must be a pain in the ass to answer all questions i guess :)

so that was easy fix
since this Kuro have a native gamut close to DCI P3 i just select this display is already calibrated and choose DCI P3 then.
thats the closest.

but even then
downconverting uhd hdr content to srgb colors with a display that has a wider gamut doent look bad.
colors gets slight more oversaturated but not in a bad way as you would have with blu ray content

[madshi]

Madshi: any idea when you could make this function work? :thanks:
I'll move it up in the priority. It's really annoying that the functionality *seems* to be there, but simply doesn't work.

if you mean the holoing around lines and discolouring of small image parts than yes i can see this ringing too else i don't really notice it.

another issue i notice with FSRCNN is the "noise" reproduction.
this old source has some colored noise in it and this noise is sometimes reproduced in a checkboard pattern which is really ugly.
It's interesting. Ringing and bloated lines stick out like a sore thumb to my eyes. I can spot that from a mile away. As does banding (at least in motion). I don't like aliasing, either, but it's not as dramatic for me as it seems to be for you. And I didn't even notice the checkboard pattern. So it really seems every person has different sensivities to different types of artifacts.

so that was easy fix
since this Kuro have a native gamut close to DCI P3 i just select this display is already calibrated and choose DCI P3 then.
thats the closest.

but even then
downconverting uhd hdr content to srgb colors with a display that has a wider gamut doent look bad.
colors gets slight more oversaturated but not in a bad way as you would have with blu ray content
So everything's fine now?

[Gopa]

What you think about NGU Sharp without any extra sharpness? IMHO:
NGU Sharp look very sharp by default, edges looks thin and sharp in focus. Another situation when you use NGU AA for image upscaling. NGU AA + post sharp can be better than NGU Sharp for low quality sources, because NGU AA can fix some edges (https://imgur.com/LBEncfq), where NGU Sharp\Standard just sharp it (https://imgur.com/OZvsRSA).

Try something like this:

Artifact removal - Debanding: Low\Low or med\med
Artifact removal - Deringing: Off
Artifact removal - RCA: What you like
Image enhancements: Off
Chroma: NGU AA High
Image doubling: NGU Sharp High (looks like you prefer sharp image)
Chroma doubling: Off\ let madvr decide
Upscaling refinement: Off
Dithering: Orderer

edit: +

2160p: Tried all of above + thin edges (refinements x2), because this refinement is always my go-to preference for anime. Not sure yet if debanding helped. Will look more closely. Chroma NGU AA High DID help! Reduce random noise (RRN?), instead of RCA, provided same results, with lower rendertime (mostly, I think, because of chroma NGU AA high). Still not quite sharp enough, though & getting very close to rendertime limit. Experimenting, with ffdshow sharpeners. I think, msharpen helps quite a bit. Lots of other possibilities in ffdshow. Not very user-friendly, though. Have yet to figure-out how to add shader pre-sets, in mpc-hc, so that is on my to-do list. Will try lower chroma and/or luma settings & see if I can use another madVR sharpener, instead of ffdshow, because I can easily see realtime results & also, because I can use anti-ringing and anti-bloating, as needed. And yes, I know, less is more, when it comes to sharpeners. I can use NGU Sharp Direct x4, so medium, might be okay, instead of high. Also, experimenting, with 1440p, instead of 2160p. I can use very high settings (anything, everything), at 1440p & images, look a little better, to me, than 2160p (significantly lower madVR settings). Thank you for your input. Almost there!

[madshi]

So guys, sorry to say, but no new madVR build this week. I tried, but didn't manage to complete a new build... :(

To wet your appetite, here's what I'm working on: I'm trying to fuse "NGU Sharp" + "Reduce Compression Artifacts" into one. Well, not really, but partially. Uhm. Ok. So if you don't use NGU Sharp, everything will stay as it is now. But if you activate both NGU Sharp + Reduce Compression Artifacts at the same time, basically RCA will come for free. So you get a nice speed reward for combining NGU Sharp + RCA. However, you can still use both separately, of course.

First results seem promising. I'll need your tests to check if there's any quality loss when using the fused algorithm instead of doing both separately. I hope there'll be no quality loss, but I'll let you test it, once it's ready to go. Maybe ready next weekend, maybe the weekend after that, but probably no later then 14 days from now.

For a start, it will only be available for NGU Sharp, maybe I'll also offer it for NGU Standard/Soft/AA, don't know yet. It's all a lot of work, so...

---------------------

In the meanwhile, I'd like your FEEDBACK on the following 3 questions:

1) Do we need "double again/twice" for NGU Sharp/Standard/Soft? Or is everybody happy with "direct quadruple"? I'd really like to remove the "double again" option, because it's dramatically bad for performance and IMHO also looks worse. But here's your chance to object, before I remove it.

2) Anybody still using NNEDI3 for chroma upscaling? If so, is there a good reason why you prefer it over NGU AA? I'd really like to remove NNEDI3 chroma upscaling, because it's simply worse than NGU AA, IMHO, and slower, so removing the option should be all good, and make the settings dialog simpler.

3) Anybody still using NNEDI3 for image upscaling? If so, is there a good reason why you prefer it over NGU AA? I've recently seen an image comparison where NGU AA Medium was significantly better than NNEDI3-128 (sharper, more detailed *and* more natural), so I really wonder if there's any need for NNEDI3, anymore?

If I could remove NNEDI3 completely, that would be a relieve, because I could completely get rid of OpenCL, with all the driver issues etc.

[ashlar42]

My opinion is that NNEDI3 can go the way of the dodo. :)

[3x3cUt0r]

I stopped using NNEDI3, so i vote yes, go ahead and remove it.

[foozoor]

You can clearly throw this opencl based sh*t in the trash.

[Gopa]

To wet your appetite, here's what I'm working on: I'm trying to fuse "NGU Sharp" + "Reduce Compression Artifacts" into one. Well, not really, but partially. Uhm. Ok. So if you don't use NGU Sharp, everything will stay as it is now. But if you activate both NGU Sharp + Reduce Compression Artifacts at the same time, basically RCA will come for free. So you get a nice speed reward for combining NGU Sharp + RCA.

Do we need "double again/twice" for NGU Sharp/Standard/Soft? Or is everybody happy with "direct quadruple"? I'd really like to remove the "double again" option, because it's dramatically bad for performance and IMHO also looks worse. But here's your chance to object, before I remove it.

Anybody still using NNEDI3 for chroma upscaling? If so, is there a good reason why you prefer it over NGU AA? I'd really like to remove NNEDI3 chroma upscaling, because it's simply worse than NGU AA, IMHO, and slower, so removing the option should be all good, and make the settings dialog simpler.

Anybody still using NNEDI3 for image upscaling? If so, is there a good reason why you prefer it over NGU AA? I've recently seen an image comparison where NGU AA Medium was significantly better than NNEDI3-128 (sharper, more detailed *and* more natural), so I really wonder if there's any need for NNEDI3, anymore?

If I could remove NNEDI3 completely, that would be a relieve, because I could completely get rid of OpenCL, with all the driver issues etc.

Free RCA!!!
I never use double again/twice. Prefer: Direct Quadruple.
I never use NNED13 for anything. Prefer: NGU AA.
NGU Sharp + Free RCA!!!

[Thunderbolt8]

I dont use nnedi3 so imho it can go

[ABDO]

So guys, sorry to say, but no new madVR build this week. I tried, but didn't manage to complete a new build... :(
No problem madshi, take your time as you need.

I'm trying to fuse "NGU Sharp" + "Reduce Compression Artifacts" into one. Well, not really, but partially. Uhm. Ok. So if you don't use NGU Sharp, everything will stay as it is now. But if you activate both NGU Sharp + Reduce Compression Artifacts at the same time, basically RCA will come for free. So you get a nice speed reward for combining NGU Sharp + RCA. However, you can still use both separately, of course.

First results seem promising For a start, it will only be available for NGU Sharp, maybe I'll also offer it for NGU Standard/Soft/AA, don't know yet. It's all a lot of work, so...
WOW that wouLd be very great work :D:D:D

---------------------


1) Do we need "double again/twice" for NGU Sharp/Standard/Soft? Or is everybody happy with "direct quadruple"? I'd really like to remove the "double again" option, because it's dramatically bad for performance and IMHO also looks worse. But here's your chance to object, before I remove it.
i think this need some test to Check it out, i will return after some tests.

2) Anybody still using NNEDI3 for chroma upscaling? If so, is there a good reason why you prefer it over NGU AA? I'd really like to remove NNEDI3 chroma upscaling, because it's simply worse than NGU AA, IMHO, and slower, so removing the option should be all good, and make the settings dialog simpler.


3) Anybody still using NNEDI3 for image upscaling? If so, is there a good reason why you prefer it over NGU AA? I've recently seen an image comparison where NGU AA Medium was significantly better than NNEDI3-128 (sharper, more detailed *and* more natural), so I really wonder if there's any need for NNEDI3, anymore?

If I could remove NNEDI3 completely, that would be a relieve, because I could completely get rid of OpenCL, with all the driver issues etc.I never ever use NNEDI3 since NGU AA come to madvr, so for me it is time to go.
thank you madshi for your effort and time :)

[JarrettH]

Why is RCA so important to be built into NGU?

[ashlar42]

Why is RCA so important to be built into NGU?I think what madshi stated is slightly different. You'll be able to use NGU without RCA.

But if you'll use NGU with RCA, that won't incur significant performance penalties, for whatever reason.

[Polopretress]

I never ever use NNEDI3.
No double/again upscaling also.


I've implemented this functionality for Sony and JVC projectors, but not yet for Epson, because I don't have an Epson projector and there didn't seem to be any Epson projector user interested in this feature. If you have an LS10000 and would like to have this feature, I might be willing to implement it, but you'd have to be available for tests, ideally via email, to make things faster/easier. I don't know when I'll find the time for this, though, could be a while.

Thank you. That would be great. I am going to contact my friend that has the Epson.


Another way to fix it, is also to be able to use the function implemented in profile "command line to execute when this profile is activated/de-activated" and launch an external *.bat that will select the good zoom memory.
So as Soulnight has highlighted , the fix of the following bug is really interresting : http://bugs.madshi.net/view.php?id=210
(and not only for the zoom memory selection of the Epson)
I'll move it up in the priority. It's really annoying that the functionality *seems* to be there, but simply doesn't work.

[Ver Greeneyes]

Why is RCA so important to be built into NGU?
Aside from getting it for free, if you're asking why NGU is expected to benefit from RCA, it's probably because upscaling tends to emphasize compression artifacts because the algorithm doesn't realize they're noise. Removing them before/during upscaling can make the scaler behave much better.

[cork_OS]

To wet your appetite, here's what I'm working on: I'm trying to fuse "NGU Sharp" + "Reduce Compression Artifacts" into one. Well, not really, but partially. Uhm. Ok. So if you don't use NGU Sharp, everything will stay as it is now. But if you activate both NGU Sharp + Reduce Compression Artifacts at the same time, basically RCA will come for free. So you get a nice speed reward for combining NGU Sharp + RCA. However, you can still use both separately, of course.
Sounds great!

1) Do we need "double again/twice" for NGU Sharp/Standard/Soft? Or is everybody happy with "direct quadruple"?

If I could remove NNEDI3 completely, that would be a relieve, because I could completely get rid of OpenCL, with all the driver issues etc.
1) "Double again" is too sharp and slow, never used it since direct quadruple is appears.

2) 3) NGU AA made NNEDI3 obslolete. No regrets.

Why is RCA so important to be built into NGU?
A lot of real life sources (especially online videos like youtube, twitch etc.) contains compression artefacts. Even if its not visible at first sight, upscaling & sharpening will often reveal and amplify it. And NGU is a sharp scaler, plus both NGU and RCA are performance hungry.

[ryrynz]

Double again and NNEDI3 can go.