Media Player .NET (MPDN) - D3D HQ GPU Video Renderer [v2.49.0/v1.31.0 27 Dec 2018]

[Zachs]

Originally Posted by Milardo

Hi is fluid motion the same as what svp can do or dmitrirender? Is it limited in anyway, or is limited by gpu? IF one uses it in combination with svp or dmitrirender, does that produce a better frame interpolation look to a video?

Fluid motion is blend FRC, except it uses extremely precise timing to get the best blends. SVP does more than that but does not use precise actual refresh rate timing. They can't be used together.

[Milardo]

So it doesn't matter about gpu? I thought it was amd tech-amd fluid motion? Does the fluid motion present in mediaplayer.net, does it make video look like its live? like your watching a real presentation?

[Zachs]

MPDN fluid motion uses the GPU for blend FRC. I'm aware AMD has their own hardware implementation but this isn't related to MPDN's fluid motion.

[Milardo]

Ok, but what is "blend frc"? How does it impact visuals on video playback?

[Zachs]

It's similar to avisynth's blend mode fps conversion, except it uses precise display refresh rate instead of a constant predefined fps.

Edit: essentially you should get more fluid playback for videos that don't match you display refresh rate.

[river1]

Originally Posted by Zachs

MPDN fluid motion uses the GPU for blend FRC. I'm aware AMD has their own hardware implementation but this isn't related to MPDN's fluid motion.

So is it possible to use AMD fluid motion video in MPDN?

[Zachs]

Nope.

[nekromantik]

Is it normal for black levels to be better with MPDN set to output Full range RBG even if TV and LAV is set to limited?
Also render times for 720 > 2160 is higher then 480 > 2160.
That cant be right can it? More work does need to be done on a SD video to upscale to 4k then a 720P video.

[Shiandow]

Originally Posted by nekromantik

Is it normal for black levels to be better with MPDN set to output Full range RBG even if TV and LAV is set to limited?
Also render times for 720 > 2160 is higher then 480 > 2160.
That cant be right can it? More work does need to be done on a SD video to upscale to 4k then a 720P video.

No idea about the black levels but scaling 720 -> 2160 and 480 -> 2160 takes about the same amount of work (or at least computations).

[nekromantik]

Originally Posted by Shiandow

No idea about the black levels but scaling 720 -> 2160 and 480 -> 2160 takes about the same amount of work (or at least computations).

Any idea why there is a big render time difference?
For 480 im getting 20 - 25 render time but with 720P im getting 50 - 55.

Edit:
Im using SuperChromaRes + SuperRES NNDI 3 passes.

[Zachs]

When you set MPDN to display full range but not have a corresponding display, you're essentially getting blacks and whites crushed. I.e. your TV displays 16-235 but MPDN has sent it 0-255. So anything below 16 becomes black, making it appear like it had better black levels. IOW you lose some bits per channel.

LAV tv output has no effect unless it's the one converting from YUV to RGB - MPDN does it in GPU.

[Shiandow]

Originally Posted by nekromantik

Any idea why there is a big render time difference?
For 480 im getting 20 - 25 render time but with 720P im getting 50 - 55.

Edit:
Im using SuperChromaRes + SuperRES NNDI 3 passes.

Well chroma scaling would take up some time. But the biggest difference is probably with SuperRes (which behaves differently from other scaling algorithms). One of the more recent changes makes it such that rendering with large scaling factors is done in multiple steps, I think 720p -> 2160p is done in 3 steps but 480p -> 2160p is done in about 4, anyway the result of this is that the processing for 480p -> 2160p is done at a lower resolution. This might seem like a disadvantage but it's actually necessary since SuperRes doesn't work well when you scale too much in one go. It's also faster, as you've noticed. With the old version of SuperRes scaling 480p -> 2160p might take up much longer, even longer than 720p -> 2160p currently does.

@all:
The newest MPDN Extensions include a small change to the default settings for debanding, the value of "margin" is changed from 1.00 to 0.25, this value will be used from now on unless you've enabled the advanced settings. This change should preserve more detail while (hopefully) still removing all banding. Feedback on this setting would be appreciated; I'm not 100% sure if this is the lowest possible setting that still removes all banding, or if this new value is already too low. If you think it's too low do keep in mind that the "Debanding" is only supposed to remove banding, not blocking or noise, you're free to use it for that purpose but those kind of settings aren't useful as a default.

[nekromantik]

Thanks guys.

@Zachs if I select limited in MPDN and have limited set on my TV I still get the blacks that turn grey issue. Is this a problem with the TV?

[Zachs]

If that's the case your graphics card drivers may be doing the conversion to limited range before sending it out. There's a registry hack to disable that.

Otherwise you could just use full range and have the driver convert it back to limited range again.

[nekromantik]

Originally Posted by Zachs

If that's the case your graphics card drivers may be doing the conversion to limited range before sending it out. There's a registry hack to disable that.

Otherwise you could just use full range and have the driver convert it back to limited range again.

OK thanks

[Anima123]

Shiandow, does deband has different behavior for different video resolutions?

When I played 576p files with deband -> SuperRes chain, there's no problem at all, no rendering time increasing, however, when playing a lower resolution file 640x480, the same render chain will causes the familiar rendering time increasing, at least to me.

The nVidia driver is the latest, setting to prefer performance.

Edit: BTW, I already set the NoIntermediates to True.

[Zachs]

Finally managed to have a play with an Optimus 750m system today (at a friend's place). Surely nVidia must've tested the product before they sold it!? Within 5 minutes, I found so many bugs with the driver that got me absolutely gobsmacked. There's even a setup that causes it to think it's running perfectly fine when in reality it's playing video as slide show (not more than 1 fps)! It sounds great in theory, until you get into the nitty gritty of what is really needed to make it work (i.e. it's nigh impossible for nVidia to get it fully working). If anyone's looking to buy a new laptop, my advise is avoid Optimus like the plague!

[TheBean]

Just signed up to say thanks for MPDN, it's great and will probably be my primary video player now.

There's a lot of customization available with the renderscripts so I'm a bit confused -- what settings have people found that the best IQ?

The content I primarily watch is live action (no anime), 720p with a pretty decent bitrate (usually x264 @ 4000kb/s+). I've been using VLC for the past couple of years with a small amount of film grain + sharpen and full debanding and this has given me pretty good IQ.

Thanks again.

[Zachs]

Depends on whether your GPU is fast enough but the following would give you very good IQ for upscaling.

SuperChromaRes
Debanding
SuperRes (NEDI enabled)

SuperRes is sharp enough so I wouldn't think you'd need more sharpening, but if you do, you could always throw an ImageProcessor with LumaSharpen (or fine sharp) in there after SuperRes.

Cheers.

[Blackfyre]

Originally Posted by Zachs

Depends on whether your GPU is fast enough but the following would give you very good IQ for upscaling.

SuperChromaRes
Debanding
SuperRes (NEDI enabled)

SuperRes is sharp enough so I wouldn't think you'd need more sharpening, but if you do, you could always throw an ImageProcessor with LumaSharpen (or fine sharp) in there after SuperRes.

Cheers.

That depends on SVP also... If you're using SVP I'd assume it'll be very difficult to do that:

SuperChromaRes
Debanding
SuperRes (Nedi Enabled)

For example for me I do source multiplied by 4x, so if you're doing the above you're actually doing:

SuperChromaRes
Debanding
SuperRes (Nedi Enabled)
SuperChromaRes
Debanding
SuperRes (Nedi Enabled)
SuperChromaRes
Debanding
SuperRes (Nedi Enabled)
SuperChromaRes
Debanding
SuperRes (Nedi Enabled)

So I realized that's the reason why I can't handle it on my 7970. SuperChromaRes-->Debanding-->SuperRes is fine, if I'm doing them alone. But with SVP 4x Every Frame, everything is multiplied by four. But this needs to be explained...

Even though your GPU is working 4 times as much with SVP, it doesn't mean sharpening is being applied four times to the same frame, so if you have luma sharpen, it doesn't multiply by four when you run SVP@4x. But it's applying luma sharpen for every "fake" or extra frame it's making (thus multiplying gpu usage by 4x or thereof)... For example with Luma Sharpen alone GPU usage goes from 15% to 60% after enabling SVP@4xEveryFrame with my HD7970.