madVR - high quality video renderer (GPU assisted)

[khanmein]

Originally Posted by BluesFanUK

NGU is very good, but to me it's a worthless function purely because of the noise it generates on the GPU. It's EXTREMELY annoying, sounds like the PC is frying.

this proven that your GPU is reaching the limit so i suggest upgrade it.

previous my GTX 970 have coil-whine but now i upgraded to GTX 1070 no more coil whine anymore.

[kitame]

Originally Posted by madshi

Are you talking about NGU (one of the official madVR builds)? Or the special "NGU pixart" test build? If it's the pixart test build, please make sure you say "NGU pixart", otherwise there might be confusion.

i'm using the official 0.91.5 build, so i'm pretty sure its not the pixart build.

[leeperry]

Originally Posted by madshi

Because NNEDI3 is very slow (and relies on unstable OpenCL). If I can make an algorithm which is both faster *and* better looking than NNEDI3 with the same sources NNEDI3 shines with, then that should be a benefit for everybody.

I hear you but options are good, NNEDI3 really shines at detecting edges in say 360*288 movies. I'm sure some ppl will always prefer NNEDI3 to anything else depending on their needs, taste and display. Anyway, please don't ditch sxbr coz I like it especially on SD that's too noisy for NGU.

[Georgel]

Originally Posted by khanmein

@Georgel what kinda graphic card u using now? cheers.

GTX 1080

Originally Posted by madshi

NGU or "NGU pixart"? Those are 2 different things!!

I copied latest madVR then copied the test files for NGU Pixart over the install like you said. Those settings are for that situation...

[Thunderbolt8]

Originally Posted by leeperry

360*288 movies

but how many people are watching movies in this resolution, let alone care about good scaling then?

[leeperry]

^ If you can't count on mVR no more for extreme hopeless upscaling then what has this world come to eh

[CruNcher]

Originally Posted by Thunderbolt8

but how many people are watching movies in this resolution, let alone care about good scaling then?

SD/D-1
HD
FHD

are the most typical inputs these days only handling sub that is crazy somehow to even take into account anymore, no TV advertises with sub SD improvements and few only with good SD Handling

Also if you think which codecs and/or analog sources have been used for those

But still DVD (Mpeg-2) is a viable target they still exist and want to be upscaled to up to 4K now

[Shiandow]

Originally Posted by Thunderbolt8

but how many people are watching movies in this resolution, let alone care about good scaling then?

Well, watching 360x288 on a 1080p screen is the same as watching 1440x1152 (i.e. >1080p) on an 8k screen, so it will become relevant eventually, probably. That said I don't think you can expect an image to look well when scaled more than 3x in each direction, eventually you just run out of detail. I am however somewhat optimistic that for smaller scaling factors you can make it hard to distinguish between a low quality high resolution image and a high quality low resolution image.

[CruNcher]

Most probably you first have to deartifacting before even thinking about the upscaling, every TV Hardware (MediaTek Based) these days has at least a simple MPEG improvement mode mostly even straight from the decoder part itself and Mali Txx onboard

And much faster as you can say A all the cheap Smart TVs will be Vulkan Render based

[cork_OS]

Originally Posted by madshi

FWIW, your zoom factor is very weird. It's not 2x and not 4x, but something in between. And I can say that with zoom factor, the current madVR builds happens to apply NGU pixart twice, but NNEDI3 only once, so it's somewhat of an apples to oranges comparison. If you pick a slightly smaller zoom factor, NGU pixart is applies only once, too, which makes the images nearer to each other. There still are differences, though.

My bad, I didn't checked that. I have 1680x1050 display and just used "touch window from outside" in player. Additionally, I didnt'n unchecked "add grain".
I redid the comparison with zoom factor around 4x and found that NGU pix output is much closer to NNEDI3 with some presharpen (like LumaSharpen 0.3) added to NGU.
NNEDI3 64+32 vs. NGU pix med+med
NNEDI3 64+32 vs. NGU pix med+med (LumaSharpen 0.3)
So I carefully assume people may prefer NNEDI3 due to its sharpness, not line thinning differences.

[Mistery 73]

which anti-bloating filter filter = 75% recommended me to use? I upscaling to 4K from 1080p bluray movies

[Asmodian]

Whichever one looks best to you. There isn't a universally recommended anti-bloating filter.

[amayra]

if i change madvr to high bitdepth output and my display does not support it
there any harm or bad effects for this ?
source : MPDN user + MPDN Options

[omarank]

Originally Posted by madshi

So now I need your FEEDBACK:

1) Do you think NNEDI3 looks better? Or the new NGU pixart variant?
2) How does performance compare for you?
3) Can I replace NNEDI3 with NGU pixart in the next official build (pretty please)?
4) Can I replace super-xbr (for luma doubling, only) with NGU pixart in the next official build?

1) NGU Pixart looks very nice and better than NNEDI3 overall, but I find NNEDI3 more natural looking. NGU Pixart gives a flat look to the images as I see it.

2) No performance issues here. On my system with AMD GPU, for 720p23 to 1080p23 scaling, NGU Pixart Very High and NNEDI3 64n show about the same rendering times.

3) I would urge you to not drop NNEDI3. There are some other users too who want NNEDI3 to stay.

4) I haven’t been using super-sbr for some time and don’t feel the need for it either. So it does not matter to me whether super-xbr is replaced or not.

[huhn]

Originally Posted by amayra

if i change madvr to high bitdepth output and my display does not support it
there any harm or bad effects for this ?
source : MPDN user + MPDN Options

someone else has to change the bit deep.

and just don't do it. i mean why even use high bit deep output if it isn't supported what a waste of time.

[Georgel]

Originally Posted by madshi

So, here's a little test build. Simply replace the v0.91.5 files with these:

http://madshi.net/madVRpixart.zip

I've created a new NGU variant which I'm going to call maybe "NGU pixart" or "NGU smooth", I'm not sure yet. This NGU variant will be optionally available in addition to the already existing NGU algorithm. The "pixart" variant has a look that is quite similar to NNEDI3. Which means it's very smooth and works great to reduce aliasing artifacts in the source. However, just as NNEDI3, "NGU pixart" is quite soft. My development goal was to create a NNEDI3 successor which has similar properties, but achieves a better "quality per watt" ratio.

Currently available are NGU pixart in "low", "medium" and "very high" quality settings. The "high" quality setting internally uses the "medium" algorithm for now. There'll be a separate "high" algo in the next official build.

In order to keep my development cost low, I haven't changed the settings dialog yet. So this test build replaces the main NGU algorithm with "NGU pixart". Which means this test build does not do the main NGU algorithm at all. The one and only purpose of this test build is so you guys can directly compare "NGU pixart" to NNEDI3, for both luma doubling and chroma upscaling.

So now I need your FEEDBACK:

1) Do you think NNEDI3 looks better? Or the new NGU pixart variant?
2) How does performance compare for you?
3) Can I replace NNEDI3 with NGU pixart in the next official build (pretty please)?
4) Can I replace super-xbr (for luma doubling, only) with NGU pixart in the next official build?

If you prefer NNEDI3, please let me know why. If you think NNEDI3 looks better, please provide comparison screenshots (for luma doubling comparisons always include the original unscaled source, please).

Thanks!

1 . NGU Pixart looks better overall.

3 . If you really want to - you could.

4. I think yes at this moment.

[atcmvr]

Hi, I had a few questions about what settings to use for playback, now that I have some hardware that can actually take advantage of more of the included settings. I apologize in advance if some of it comes off as very basic. Using the latest version of madVR 64, XySubFilter_3.1.0.746_x64_BETA3, display is an Asus pg279q, and GPU is an evga gtx 1080 ftw.

First off, in the devices tab, for hdr, is it possible/recommended to use the 3dlut used for calibration? Or are they 2 completely separate things, that each need to be generated on their own? In the scaling algorithms section, for image upscaling, if you choose ngu, what is the difference between the chroma here, and the actual chroma upscaling tab? Does one take precedence, or are they doing entirely different things? And is it sufficient to leave the chroma option in image upscaling, along with the algorithms to use after doubling to "let madvr decide"? Also, I think I read that you don't really have to bother with any of the upscaling refinements if using ngu, since many of them are already applied?

Also, is ngu high/very high "better" than using nnedi with a high neuron count? Another question I had was, does bumping the gpu queue size up past 8 do anything? I ask, because on a different machine + prior version of madvr, maybe 2 years ago, increasing that option to 16 from 8 helped me avoid frame drops for the settings I had then.

Finally, what combination of settings should I use, in terms of v-sync, g-sync, and smooth motion? The monitor I have is a g-sync monitor, can go up 165hz. Normally leave it at that, with g-sync enabled, and v-sync enabled through the nvidia console as well. But for viewing content that I guess is 23.976fps, at least going by what madvr says, should I turn of g-sync? Turn off v-sync? Drop the refresh rate to 144hz? Does doing any of that even matter/change anything And which setting under smooth motion would I choose, in combination with the above?

Also had what I suppose is a bug report, though it seems to have been from nvidia's side. Thought I would mention it, in case anybody else has the same issue. Using drivers from about a month ago, I was not able to get stable playback using any of the "demanding" algorithms. Had to drop settings down to NNEDI with 16 neurons just to get sub 30 ms render times, anything above had massive frame drops, render times at 70+. I noticed however that there would be periods of time where it would drop back down to normal levels, and the gpu queue would fill up, only for the queue to empty almost immediately, and times to shoot back up again. Looking at the gpu stats, using evga's tool, noticed that the gpu looked like it wasn't being utilized properly during the time when plaback was having an issue, ie. the clock speed would drop drastically, memory speed close to zero, along with power usage. Manually forcing max speeds on both clock and memory fixed the problem, was able to use ngu very high for 720p content, high for 1080p, but obviously something was wrong there. Decided to upgrade drivers today in the hopes it would fix it, and it actually seems to have done the trick, not noticing the problem anymore. Again, just an fyi.

[kitame]

question, how does the latest IGP perform on the current MadVR scalers?
e.g. AMD's Kaveri / Godavari APUs, or Intel's Skylake / Kabylake CPUs.

is NGU too "heavy" for these IGP?

Originally Posted by Shiandow

Well, watching 360x288 on a 1080p screen is the same as watching 1440x1152 (i.e. >1080p) on an 8k screen, so it will become relevant eventually, probably. That said I don't think you can expect an image to look well when scaled more than 3x in each direction, eventually you just run out of detail. I am however somewhat optimistic that for smaller scaling factors you can make it hard to distinguish between a low quality high resolution image and a high quality low resolution image.

not quite the same really, 360x288 source lacks way too much detail compared to 1080p source.
its like the gap detail between 16x16, 32x32, 64x64 and 128x128 pixel art or sprites.

[ryrynz]

Originally Posted by Georgel

1 . NGU Pixart looks better overall.

Depends on what you're viewing. The lines that NGU pixart creates are excellent but the overall look is bland, it's also sometimes not smart enough and removes detail which I've stated as it being a bit "too aggressive"

Originally Posted by Georgel

3 . If you really want to - you could.

At this point if that were to happen there would definitely be some unhappy people, me being one of them.

Here's NNEDI3 64 neurons vs NGU very high on a source I use a lot when evaluating upscalers and my own Anime Avisynth profies for low res content.

http://screenshotcomparison.com/comparison/199986

There is reasonably strong ringing and aliasing and quite soft lines, NNEDI3 enhances more of the image creating lines stronger and true to the source. Here the eyes are sharper and the image is more defined, although the lines are not as well enhanced as NGU pix very high. This is NGU's best quality and it bests even NNEDI3 256 neurons at creating these beautiful non aliased lines, notice the top shirt lines just below the neck for a good example of this. Unfortunately like I've said it also removes some things it doesn't consider to be lines, the wedge of hair just above the left eyebrow loses it's blurry left line for example.

Looking at the above you can see NGU pix's somewhat dull, processed appearance, for NGU pix to replace NNEDI3 it needs to do a better job in all areas and it's simply not doing this right now.

[kitame]

i'm seeing minor aliasing on NGU, but i can see errors on NNEDI3.

aliasing on NGU = top hair is most evident.
errors on NNEDI3 = sleeve cuff by the yellow ribbon has a broken line.

i'm using 10x zoom comparison.