madVR - high quality video renderer (GPU assisted)

[huhn]

you can add a hotkey to a profile.

[Damien147]

Originally Posted by HillieSan

I know that AMD changed something with DXVA in the previous driver. I would wait for the next driver.

Ok,clean install next time and lets see.


@tranfa

Took you some time but all good now,happy playback.




I bump this:

I have an old cpu(i5 760 oced 3800) which performance wise is like i5-2400(Sandy Bridge).My gpu(rx470) can handle 10bit hevc with DXVA2(copyback) as other users here(with same GPU) report but I can't.Could it be a bottleneck from cpu for some reason?Some instruction missing or something?I don't get high cpu load with DXVA2(copyback) to see the ''bottleneck'' if it exists but 10 bit HEVC is slideshow.

any opinions what happens?

[aufkrawall]

Originally Posted by Damien147

DXVA2(native) causes that?

Yes, at least with 8 bit H.264. I'm not aware if anyone tested HEVC 10 bit for this, but it might be affected as well. Probably you will also get more banding with DXVA2 native and 8 bit video than with software/copyback, but dunno if this was related to just DXVA2 scaling or also decoding.

[Damien147]

I guess the only solution then is dxva2(copyback) but I don't get why others with same GPU have smooth 10 bit HEVC playback and I don't. dxva2(copyback) works well and smooth with everything except 10 bit HEVC where it's slow.
Gonna settle with software decoding and lower bitrate 10 bit HEVC if nothing happens.I don't have a compromise with everyday usage,all good.

[aufkrawall]

Good question. afair nev's copyback in LAV uses SSE 4.1, but your CPU already supports 4.2. Maybe an undiscovered CPU architecture flaw?
Or some platform specific problem with PCIe?
Just wild guessing here.
You could try a video player that supports native CUDA input, which is probably just one (better not name it here).

[Damien147]

Yes,I have SSE 4.2 and performance wise I am like i5 2400S.
Mobo PCIe is x16 2.0.I can play battlefield 1 in 1920x1200 with 60fps+ on ultra.I don't have bottleneck while gaming and I have during playback?
Thanks for the suggestion but I am with AMD,I don't have cuda support.
Maybe it's the '' an undiscovered CPU architecture flaw?'' you said,I don't know.
I don't know if it counts but I tried passmark the other day to see and compare with same CPUs and GPUs to see if everything works fine.I get good/similar score,it doesn't show a bottleneck anywhere.

[wasper]

Originally Posted by madshi

madVR tries to avoid using DXVA processing if at all possible, because often the GPU drivers do fancy things to the video when using DXVA. Avoiding DXVA is possible when using film mode, but not when using DXVA deinterlacing, obviously. This is the cause of those color shifts. As huhn already suggested, check the GPU control panel. There are many options there. Probably one related to "dynamic range" or "content type" or similar might be responsible for this problem.

Ok, thank you, I figured out. Setting the video mode back to limited ranged fixed it. The output is still full range. I was mixed up between output range and input range. As DVD video are native limited range, I'm pretty much sure I'm fine! I'll try to find a full range video to see if it's working fine also in that case.

Originally Posted by madshi

madVR currently only supports film mode or DXVA deinterlacing. DXVA deint is *not* video mode. It's a black box. madVR sends the interlaced frames/fields into DXVA and gets progressive frames back. What DXVA does internally, madVR doesn't know. Most GPU manufacturers try to auto detect if the source is video or film, often even per pixel, and try to apply a proper deinterlacing method for that. Unfortunately due to how the DXVA interface is designed, the output is usually double frame rate, even for film sources, which means you can't play the result at 24Hz, and you often have 3:2 pulldown judder baked into the DXVA output.

Really sorry to bother you with something I should already understand... What is video mode exactly? For me, video mode, is just not film mode. So maybe, it does something more that could bring me other benefits. I searched forums and wasn't able to find a clear answer.

Originally Posted by madshi

Hardware DVD/Blu-Ray players have a processing chip built in which does the deinterlacing. Depending on the chip, deinterlacing quality can vary between good and bad. Better chips are often able to deinterlacing film content correctly, some can even IVTC for 24fps output. How good that works will vary from player to player.

Generally, applying IVTC can be very easy, or it can be very tricky, depending on the content. Some content is really 24fps, but just with flags for 60i output. Such content is extremely easy to output in 24fps. Other content is encoded as 60 different fields, either with or without flags that indicate which fields belong together. Yet other content might constantly switch between different encoding modes. madVR's film mode understands every of these variants and tries to produce perfect 24fps. It usually works pretty well. Hardware DVD/Blu-Ray players usually have no problem with the easier encoding variants, but might stumble with more difficult situations.

Ok, I'm trying to use film mode as much as possible. However, I have some videos that seems to be 3:2 pulldown but not perfectly. Playing them in film mode will bring a lot of cadence break and dropped frame. On the other side, if I'm deintrelacing the 29.97 video to play it at 59.9, it won't play well as there is a lot of pulldown judder like you said. Overall, I prefer film mode even with some glitch during playback. So, why my blu-ray is able to play this DVD correctly and I'm not. Is it because there is some frame interpolation or a kind of smooth playback like the one you built? I'm really thinking to use this feature for this video. Is a DVD video could get corrupted with scratch like a CD? In that case, the .vob file would be unusable, right?

Finally, thank you for your answer and time.

[Tarzan]

Originally Posted by Tarzan

This Wednesday I am finally getting a projector with lens memory function (jvc x35) and I am now making different profiles according to aspect ratios. However I wonder whether madvr after playback automatically changes the lens memory back to the whatever it was before playback.

Example:
I use memory 1 for viewing the windows desktop in 16:9. A cinemascope movie is started and madvr activates lens memory 2. After playback, will madvr be stuck on memory 2, or will it change back to memory 1?

Yes I know, I could wait until Wednesday, but I'm curious.

Ok, so now it is finally Wednesday and I know madvr does not change the lens memory back to whatever it was before playback. Is there any way of making it activate a certain lens memory while quitting madvr?

[aufkrawall]

Originally Posted by Damien147

but I am with AMD

Oops, I thought you were on Nvidia. Then DXVA2n might work without blurry chroma for you.

[Damien147]

Oh great!Full chroma with black screen.

[Georgel]

Tested the new NGU in depth and the performance is okay but heats up the GPU quite a bit. Those figures are after a 10 minute run of a video at medium medium for NGU , original video res 1920x1080, presented on 2560x1440



A bit of efficiency would be welcome, otherwise it will be mostly usable on low without turning my setup into a vacuum cleaner

[pose]

Georgel, whats in upscaling refinement? Also is that NGU pixart?

[Bloodred]

There must be something going on with your cooling, I use a GTX 1080 too and I haven't seen any of the NGU variants heat it up that much, mine stays around ~65C at ~80% load (pushing beyond that tends to make it start dropping frames for me). Not trying to dismiss your issue as a "works on my machine" type of deal, but a 1080 really shouldn't be hitting 85C at <200W peak power draw if adequately cooled.

[Georgel]

Originally Posted by Bloodred

There must be something going on with your cooling, I use a GTX 1080 too and I haven't seen any of the NGU variants heat it up that much, mine stays around ~65C at ~80% load (pushing beyond that tends to make it start dropping frames for me). Not trying to dismiss your issue as a "works on my machine" type of deal, but a 1080 really shouldn't be hitting 85C at <200W peak power draw if adequately cooled.

Laptop GTX 1080. Same power as desktop, cooling isn't as good so it heats more. Its safe temps are up to 92C where it becomes a real problem.

Those were the results with the fans on auto and a cooler turned at max underneath the laptop.

Turning the fans within at max level will make a noise like a vacuum cleaner, but it decreases the temperatures a bit, down to 70C - 75C.

Not that bad for laptops and taking into account that the power is actually similar to desktop.

Originally Posted by pose

Georgel, whats in upscaling refinement? Also is that NGU pixart?

Running the latest version of NGU pixart and nothing turned on inside Upscaling refinement since Pixart looks best without any additions to me.

[omarank]

Originally Posted by madshi

Here's a new NGU pixart test build:

http://madshi.net/nguPixArt.zip

Overall, I prefer NGU pixart 2/3 over the other options, but I can't decide if I prefer pixart 2 or 3. With some images I prefer 2, with some 3.

What are your thoughts?

NGU pixart 3 looks better than NGU pixart 2. Although these two versions look more natural than the first version of NGU pixart, yet I find the look of NNEDI3 even more natural. As I see it, all variants of NGU tend to sharpen the edges more than the texture, while NNEDI3 is more balanced. I will see if I can find any more samples which highlight the difference. I look forward to seeing NGU pixart 4.

[cork_OS]

Originally Posted by omarank

As I see it, all variants of NGU tend to sharpen the edges more than the texture

This is especially visible with NGU standard.

[Georgel]

Originally Posted by omarank

NGU pixart 3 looks better than NGU pixart 2. Although these two versions look more natural than the first version of NGU pixart, yet I find the look of NNEDI3 even more natural. As I see it, all variants of NGU tend to sharpen the edges more than the texture, while NNEDI3 is more balanced. I will see if I can find any more samples which highlight the difference. I look forward to seeing NGU pixart 4.

Now that you mention it, it does look like that is what is going on.

But NGU Pixart was basically made for low res content and especially anime, so using it with Real Life content might result in different results compared to normal NGU.

To my eyes, NGU Pixart looks considerably better with anime than typical NGU

What plans does madshi have for madVR that its development seems to have slowed down? What can we expect in madVR 1.0.0?

[huhn]

Originally Posted by Georgel

Now that you mention it, it does look like that is what is going on.

But NGU Pixart was basically made for low res content and especially anime, so using it with Real Life content might result in different results compared to normal NGU.

To my eyes, NGU Pixart looks considerably better with anime than typical NGU

the new NGU pixart only works at very high.

[Damien147]

@XMonarchY

Maybe free espresso coffee The best thing I can imagine is a shader which combines the best attributes of NGU pixart/NGU standard with the ''3D'' effect of old softcubic!That would be great.


@Madshi

I don't want to be tiring but I noticed somewhere that you have AMD HD7770.You don't get black screen with 10bit(hevc) in dxva2(native)?Is it only RX 4x0 issue?