madVR - high quality video renderer (GPU assisted)

[madshi]

madVR v0.88.7 released

http://madshi.net/madVR.zip

* fixed: D3D11 render queue didn't fill in Windows 7
* fixed: FineSharp didn't like "use 10bit image buffer instead of 16bit"
* fixed: "madVR_Load3dlutFile" ignored the attached "cal1" gamma ramps
* updated to latest Shiandow deband script
This might be the last update for a while, unless new critical bugs show up. If you find non-show-stopper-bugs, which are 100% reproducable and definitely madVR's fault and not a misconfiguration or a GPU driver bug, then please enter them into the madVR bug tracker. Thanks.

Still need more feedback about Shiandow's latest deband algorithm version. Thanks!

[omarank]

I still have the same opinion as before: images look better with madVR’s old debanding algorithms irrespective of the debanding strength. Instead of replacing the old “high” preset with Shiandow’s algorithm, I believe it would be better to modify the old “high” preset algorithm such that it is able to remove an equivalent amount of banding as Shiandow’s algo. However, I have no issues if you even replace the “high” preset as Shiandow’s debanding has really improved after he made those modifications.

By the way, you haven’t added that trade quality option for disabling Error Diffusion dithering for 10 bit output? It’s not a high priority thing, I am just mentioning it so that this thing doesn’t get missed.

[detmek]

Madshi, render queue gets filled with this new version on my Intel IGPU in full screen windowed mode. Now I only have a second or two frame out-of-order during switching Full Screen - Windowed mode and vice versa. Same with switching Windowed - FSE. It maybe better for me to stay with D3D9 for now.

[huhn]

So, if 8 bit are selected in the CCC, it will output a signal with this bitdepth, even though applications in FSE used 10 bit?

yes i tested this it is the case. i disabled dithering for AMD in the registry and feed it a 10 bit in put and saw banding like undithered 8 bit madVR.

Well, this was exactly my question, which remains open. :(
i was talking about adobes openGL part. as nev said windows DWM can't do that at the moment.

[zveroboy]

You probably have D3D11 enabled? If you do that, madVR cannot disable desktop composition.
I tried both (D3D11 enabled & D3D11 disabled). Still not working.
For a moment, turned off desktop composition, and immediately turned back on.

0.88.7 - the same behavior.

[huhn]

As mentioned in the v0.88.6 release notes, octupling is only performed if you activate one of the "image refinement" options at the same time. If you don't do that, octupling would be a waste of power.

ok it found the problem it only works with "refine the image after every ~x2 upscaling step".
"refine the image only once after upscaling" doesn't work.

[aufkrawall]

@aufkrawall
Do you have the trade for quality option "don't analyze gradient angles for debanding" un-ticked?

Yes, I do.

Seen from the status quo, I'd also stick with the classic option set to high and I agree that I'd even like it to be stronger.

[luk008]

madVR v0.88.7 released

http://madshi.net/madVR.zip

This build is running pretty fine here. I haven't noticied any bug until now. :thanks:

[huhn]

Yes, I do.

Seen from the status quo, I'd also stick with the classic option set to high and I agree that I'd even like it to be stronger.

the "ultra" option. doesn't sound to wrong to me.

the problem is this option would clearly harm high quality source.

[Dogway]

This might be the last update for a while, unless new critical bugs show up. If you find non-show-stopper-bugs, which are 100% reproducable and definitely madVR's fault and not a misconfiguration or a GPU driver bug, then please enter them into the madVR bug tracker. Thanks.

Can't you image double in YUV space? Currently there are 2 extra color conversions (RGB->YCbCr->RGB).

This is worrying:
When doing image doubling, the "input" into the doubling algorithm is always RGB. When using NNEDI3 for doubling, this RGB image is then split into Luma and Chroma. Luma is doubled with NNEDI3. Chroma is doubled by using the "image upscaling" algorithm - but only if it has less than 3 taps. If the "image upscaling" algorithm has 3 taps or more, madVR doubles Chroma by using Catmull-Rom instead. I've implemented these changes because the algorithm used for Chroma doubling is very very unimportant, so it's no use wasting precious GPU processing time there.

Luma and Chroma doubling should use the same kernel otherwise mismatches will arise specially on aliased borders.

[James Freeman]

I like Shiandow's debanding, it does an excellent job and it is very versatile with only 2 settings.
I love how the Threshold determines the strength and Detail can fine tune everything in between.

I find that both algorithms are doing what they're meant to do perfectly... so it is a tough choice.
A debanding filter is a "set it and forget it" for some people and some will want a more in depth control.
I think that Shiandow's debanding is great for both types of people, it can do exactly what the default debanding does with everything in between (and way more if you want) and only with 2 knobs to twist.

Some people like a setting above High (I read Ultra somewhere), and some will want lower than Low.
Personally I find the custom mode with the default debanding algo too complicated so I never used it, but with shiandow's algo a single Detail knob does just that.

Again, it's a though choice but I would go with the one that can be fine tuned easily, and comes out as a one size fits all.

[MysteryX]

Wow, a lot of updates came out recently!

I'm trying the new version. DirectX 11 presentation path isn't working on my computer. The audio plays fine but there is no video (black). I had the same issue when testing Media Player.NET + DirectX 11

As for image doubling, it still isn't working on my computer. Applying NNEDI3 frame quadrupling doesn't affect rendering times. The option "use alternative interop hack" isn't helping either. However, NEDI doubling works.

I have a Radeon HD 7670M with Windows 7 x64.

[huhn]

madVR v0.88.7 released

http://madshi.net/madVR.zip

* fixed: D3D11 render queue didn't fill in Windows 7
* fixed: FineSharp didn't like "use 10bit image buffer instead of 16bit"
* fixed: "madVR_Load3dlutFile" ignored the attached "cal1" gamma ramps
* updated to latest Shiandow deband script
This might be the last update for a while, unless new critical bugs show up. If you find non-show-stopper-bugs, which are 100% reproducable and definitely madVR's fault and not a misconfiguration or a GPU driver bug, then please enter them into the madVR bug tracker. Thanks.

Still need more feedback about Shiandow's latest deband algorithm version. Thanks!


this version has broken D3D11 10 bit output.

i get now "always" tearing with 8 present queue and my other more stable tested present queue 2-3.
don't waste your time fixing this now. it's my problem using windows 10.
i just want you to know that windows 10 doesn't like your newest changes.

edit: ignore this for now nothing works right anymore sorry. i have to investigate way more why.

[aufkrawall]

the problem is this option would clearly harm high quality source.
If it was just about rounding issues, why would madshi implement Shiandow's algorithm? :)

Btw: The latest two madVR versions are running absolutely issue-free for me. :thanks:

[huhn]

If it was just about rounding issues, why would madshi implement Shiandow's algorithm? :)
high does damage already. that's my point
Btw: The latest two madVR versions are running absolutely issue-free for me. :thanks:

can you please test d3d11 10 bit FSE? you are on AMD right?
would be nice to know if my system is just on a rampage.

[James Freeman]

FSE mode fails to engage. Black Screen in 8bit. "exclusive mode failed" message in 10bit.
Nvidia Windows 7 64

:(

[leeperry]

Actually, no. The "octuple" option had nothing to do with that. You can't uncheck "double/quad chroma resolution" if you have NEDI selected because NEDI handles luma+chroma at the same time.
My bad, then what's confusing is that whenever luma NEDI is checked, ideally NEDI chroma should also be automatically ticked in order to make very clear that NEDI feeds RGB...but that you could still use NNEDI3 for luma(together with chroma if you so wish) if you don't run NEDI simultaneously.

I also wish I could enable 2X for >1.01 and 4X for 2.01, I kinda find 1.2 & 2.4 too high and pretty arbitrary.

[tobindac]

"No dithering" is not a viable choice *at all*. Testing with 1bit is not a good test for this. Try 2bit. See here:

http://madshi.net/2bitNoDithering.png
http://madshi.net/2bitOrderedDithering.png

(CAUTION: Please zoom the images in the browser to 100%, otherwise the dither pattern will look ugly.)

I think no sane human could possibly prefer the image with no dithering there.

Hehe. I know the visual information is less without dithering in those examples. I meant the motion noise is more.

What if the tradeoff favors no noise at all given enough colors already available, for some users or cases?

[Ver Greeneyes]

Working well in windowed mode, queues fill up now.

In the previous version, I had an odd issue with seeking backward where the video slowed to a crawl while the audio continued to play, badly desyncing them. The video eventually resumed playing at full speed, but still very delayed. But it's only happened once, and for all I know the queue fix also fixed that, so don't worry about it for now.

[huhn]

My bad, then what's confusing is that whenever luma NEDI is checked, ideally NEDI chroma should also be automatically ticked in order to make very clear that NEDI feeds RGB...but that you could still use NNEDI3 for luma(together with chroma if you so wish) if you don't run NEDI simultaneously.
when i check NEDI luma NEDI chroma is checked too.

[Shiandow]

I like Shiandow's debanding, it does an excellent job and it is very versatile with only 2 settings.
I love how the Threshold determines the strength and Detail can fine tune everything in between.

I find that both algorithms are doing what they're meant to do perfectly... so it is a tough choice.
A debanding filter is a "set it and forget it" for some people and some will want a more in depth control.
I think that Shiandow's debanding is great for both types of people, it can do exactly what the default debanding does with everything in between (and way more if you want) and only with 2 knobs to twist.

Some people like a setting above High (I read Ultra somewhere), and some will want lower than Low.
Personally I find the custom mode with the default debanding algo too complicated so I never used it, but with shiandow's algo a single Detail knob does just that.

Again, it's a though choice but I would go with the one that can be fine tuned easily, and comes out as a one size fits all.

Hmm, I think I discovered a better method, with only 1 (useful) parameter, which would control the balance between "detail" and debanding. I'm still testing it though, not sure how fast I can finish it.

Also, thanks to everyone for the feedback, it helps a lot to know where the algorithm can be improved.

[luk008]

high does damage already. that's my point


can you please test d3d11 10 bit FSE? you are on AMD right?
would be nice to know if my system is just on a rampage.

Here it's working normally.
Windows 8.1 x64 + AMD R9 270

PS: I don't have a 10bit display. Doing the 10 bit test I can only see the difference between 8 and 10 using AMD dithering (registry), like in the older 0.88.xx releases. Don't know if this is your point, but nothing is crashing. I hope this helps.

[MysteryX]

Something is confusing.

SuperRes can be applied either in Upscaling Refinement or in Chroma Upscaling. What's the difference?

FineSharp and LumaSharpen can either be applied in Image Enhancements or Upscaling Refinement. What's the difference?

And because SuperRes includes anti-ringing, can I safely remove Anti-Ringing with Jinc if combined with SuperRes?

[MS-DOS]

FSE mode works fine here in both 8 and 10 bit modes. 2 hours of playback without a single drop\glitch, "present a frame for every Vsync" unchecked.
All queues are full in windowed mode, while present queue is at 6-7\8 in FSE.

[huhn]

Something is confusing.

SuperRes can be applied either in Upscaling Refinement or in Chroma Upscaling. What's the difference?

the superres from Upscaling Refinement isn't used on chroma upsacaling.
don't forget chroma upscaling does only scale to RGB and nothing else

FineSharp and LumaSharpen can either be applied in Image Enhancements or Upscaling Refinement. What's the difference?
the "Image Enhancements" is applied before scaling AFAIK.

[madshi]

I still have the same opinion as before: images look better with madVR’s old debanding algorithms irrespective of the debanding strength. Instead of replacing the old “high” preset with Shiandow’s algorithm, I believe it would be better to modify the old “high” preset algorithm such that it is able to remove an equivalent amount of banding as Shiandow’s algo. However, I have no issues if you even replace the “high” preset as Shiandow’s debanding has really improved after he made those modifications.
Thanks for the feedback! Seems there'll be another Shiandow deband iteration coming. <sigh> Of course it's all good, the algorithm getting better all the time. But I hope you users/testers won't be bored at some point, retesting revised algos all the time?

By the way, you haven’t added that trade quality option for disabling Error Diffusion dithering for 10 bit output?
It's on my list.

Madshi, render queue gets filled with this new version on my Intel IGPU in full screen windowed mode. Now I only have a second or two frame out-of-order during switching Full Screen - Windowed mode and vice versa. Same with switching Windowed - FSE. It maybe better for me to stay with D3D9 for now.
If you see no advantage for D3D11 than there's no problem at all staying with D3D9. I think I didn't answer to your previous comment: If "use a separate device for presentation" causes the same problem with D3D9 then that points to a driver issue. I've had several of those with Intel already. Many of them have to do with sharing textures over multiple devices, sadly.

I tried both (D3D11 enabled & D3D11 disabled). Still not working.
For a moment, turned off desktop composition, and immediately turned back on.

0.88.7 - the same behavior.
Did you restart the media player after disabling D3D11? If so, did previous madVR versions not have this problem? If so, could you please check which exact madVR version introduced the problem for you? That would help me figuring out which change may have broken this for you. You can find the old madVR builds here:

http://www.videohelp.com/software/madVR/old-versions#download

ok it found the problem it only works with "refine the image after every ~x2 upscaling step".
"refine the image only once after upscaling" doesn't work.
Ah, I see. But that does make sense: If you refine the image only once after upscaling is complete, octupling won't bring any benefit over e.g. Jinc3 AR upscaling. Luma doubling brings a nice quality improvement. Quadrupling only a small improvement. Octupling will be invisible, unless you sharpen the image in between the doubling steps.

Can't you image double in YUV space? Currently there are 2 extra color conversions (RGB->YCbCr->RGB).

This is worrying:

Luma and Chroma doubling should use the same kernel otherwise mismatches will arise specially on aliased borders.
Artifacts from different algorithms will be rather low, because the chroma channel is usually very soft. But if you're worried about this kind of problems, you can use both luma+chroma doubling, then this problem should be solved.

Multiple colorspace conversions are not "nice", but I'm using 16bit integer textures which have a LOT of precision, and I'm not cutting away BTB/WTW, either, so there's really not much to worry about. The colorspace conversions in madVR are not 100% lossless, but they should be nearly lossless.

I like Shiandow's debanding, it does an excellent job and it is very versatile with only 2 settings.
I love how the Threshold determines the strength and Detail can fine tune everything in between.

I find that both algorithms are doing what they're meant to do perfectly... so it is a tough choice.
A debanding filter is a "set it and forget it" for some people and some will want a more in depth control.
I think that Shiandow's debanding is great for both types of people, it can do exactly what the default debanding does with everything in between (and way more if you want) and only with 2 knobs to twist.

Some people like a setting above High (I read Ultra somewhere), and some will want lower than Low.
Personally I find the custom mode with the default debanding algo too complicated so I never used it, but with shiandow's algo a single Detail knob does just that.

Again, it's a though choice but I would go with the one that can be fine tuned easily, and comes out as a one size fits all.
There will be no fine tuning (by users). After all is said and done, there will be low/med/high, just as it is now. No other controls. Regardless of which algorithm will be chosen. So "configurability" is not something I'm interested in hearing your feedback about. What I want to hear about is quality differences, maybe quality vs performance considerations, nothing else. Thanks!

(Maybe, in some far away time, there might be an advanced options page with more controls, but this is not something that is of any decision making interest to me right now. I only care about the algo which provides the best quality (and performance) right now. Everything else simply DOESN'T MATTER.)

I'm trying the new version. DirectX 11 presentation path isn't working on my computer. The audio plays fine but there is no video (black). I had the same issue when testing Media Player.NET + DirectX 11

As for image doubling, it still isn't working on my computer. Applying NNEDI3 frame quadrupling doesn't affect rendering times. The option "use alternative interop hack" isn't helping either. However, NEDI doubling works.

I have a Radeon HD 7670M with Windows 7 x64.
Do you have the Windows 7 Platform Update installed? If yes, this sounds like a driver issue to me. Don't know what else it could be.

can you please test d3d11 10 bit FSE? you are on AMD right?
would be nice to know if my system is just on a rampage.
Works fine on two different Windows 7 PCs with two different GPUs for me. And works also fine on my Windows 8.1 x64 dev PC. All in both 8bit and 10bit.

FSE mode fails to engage. Black Screen in 8bit. "exclusive mode failed" message in 10bit.
Nvidia Windows 7 64
Works fine here. Is that a new problem for you? If so, could you please try to find out which exact madVR build introduced this problem for you?

My bad, then what's confusing is that whenever luma NEDI is checked, ideally NEDI chroma should also be automatically ticked
Exactly that should already happen, as huhn already reported.

I also wish I could enable 2X for >1.01 and 4X for 2.01, I kinda find 1.2 & 2.4 too high and pretty arbitrary.
Isn't "always - if upscaling is needed" exactly the same as 1.01/2.01? I'm confused.

Something is confusing.

SuperRes can be applied either in Upscaling Refinement or in Chroma Upscaling. What's the difference?

FineSharp and LumaSharpen can either be applied in Image Enhancements or Upscaling Refinement. What's the difference?
I've explained this in detail in the first couple of pages following the initial v0.88.0 release.

FSE mode works fine here in both 8 and 10 bit modes. 2 hours of playback without a single drop\glitch, "present a frame for every Vsync" unchecked.
All queues are full in windowed mode, while present queue is at 6-7\8 in FSE.
That's what I want to hear... :)

[aufkrawall]

can you please test d3d11 10 bit FSE? you are on AMD right?
would be nice to know if my system is just on a rampage.
I'm on NV (GTX 980, Win 10074, driver 350.12, x64).
Seems like 10 bit D3D11 FSE is working fine (tested in 10 bit mode with 8 and 10 bit video), but I have to enable frame for every sync to avoid massive presentation glitches. No tearing etc.

[XMonarchY]

How do you know it uses the BT.709 3dlut? If that is really the case, it's a bug and then please enter it to the madVR bug tracker.

Its difficult to explain, but I do know it for a fact. When I made Rec.709 3DLUT I had my TV's setting for Flesh Tone set to 0, but I accidentally left Flash Tone set to -15 when I created Rec.601 3DLUT. That creates an obvious difference when I switch between Rec.709 and Rec.601 3DLUT's using identical TV settings. Rec.709 line is used for SMPTE C content for sure, at least in 88.5.

[MysteryX]

Shouldn't "Upscaling Refinement" be within profile groups instead of being on its own for all profiles?

[aufkrawall]

Thanks for the feedback! Seems there'll be another Shiandow deband iteration coming. <sigh> Of course it's all good, the algorithm getting better all the time. But I hope you users/testers won't be bored at some point, retesting revised algos all the time?

It's not very time consuming, and if one wants better quality it's not really a sacrifice.

However, have we seen yet example images that show a real advantage for Shiandow's current implementation?
I for myself haven't found such an example yet. Higher threshould values with bad sources can be nice because of that deblocking effect, but it vanishes too much detail with good sources.

[XMonarchY]

88.7 works wonderful for me! I get no frame drops, even when I use Chroma Upscaling SuperRes Filter and also enable SuperRes in Image Refinement (High setting). Hell, I even enabled Octuple settings and everything is peachy! Shiandow's debanding and new setting (0.2 for threshold and 1 for detail level) are also better. I used to get obvious big shimmering squares (NOT dithering) on darker grays in Mid/Low-HQ content that were not present when only the original debanding was used. With 88.7 and new Shaindow's settings, those shimmering squares are now a lot smaller. Sometimes I think the original de-banding is still better, not as noisy.

[huhn]

I'm on NV (GTX 980, Win 10074, driver 350.12, x64).
Seems like 10 bit D3D11 FSE is working fine (tested in 10 bit mode with 8 and 10 bit video), but I have to enable frame for every sync to avoid massive presentation glitches. No tearing etc.

could be limited to the AMD WDDM 2.0 driver.
thanks.

[tickled_pink]

Stupid(?) question - When testing deband algos, we should check one or the other, correct? Leaving both checked is effectively using both algos simultaneously, or am I missing something?

[cyberbeing]

madVR v0.88.7 released

* fixed: D3D11 render queue didn't fill in Windows 7

This version appears to have broken D3D11 entirely on my Win7 GTX770 system, or rather it never activates and madVR always uses D3D9. It works when I reset madVR to defaults though, so it must be something related to my settings/profiles, but so far I've not been able to figure out which causes it.

https://www.mediafire.com/?x1fgya1148ecsm4

[x7007]

I have couple of questions.

Currently by your eyes if I have 970 GTX I have Philips 7007 3D it's 16-235. Should I use native display bitrate 8bit or 10bit ? I'm using the D3D11.

What should I choose for debanding Debanding Strength - medium / fade in/out high ? or to enable shaindow's deband threshold 0.5 detail level 2 ? I don't remember if that's the default

Should I use quadruple luma resolution ?

How come without the Settings.bin file in the folder Madvr keep my settings ? and after that it makes the file, how it remembers all my settings without the settings.bin ?

Those settings bothering me cause I can't be sure what's the best.

Thanks

[aufkrawall]

could be limited to the AMD WDDM 2.0 driver.
thanks.
Funny, I had the exactly same problem with the Nvidia WDDM 2.0 driver. :eek:
Edit: But this was not related to 10 bit, but to the D3D11 renderer in general.

[daert]

Hi madshi. First I want to thank you for your hard work. I've been using madvr for a long time without any problem but the last few builds doesn't work well for me.
In the latest build (0.88.7) d3d11 is not enabled at all even if I check the option. The OSD shows always "(new path)" both in windowed mode and FSE.
In 0.88.6 the refresh rate fix doesn't work. With D3d9 FSE both 24p and 60p work but with d3d11 FSE only 60p works, 24p doesn't work yet.
Let me explain what happens: my monitor has a 23p with a blurred image and a timing of 23.970 so I created a 24p custom resolution with the exact timing of 23.976 and a clear image. When MPC enters in FSE the resolution is switched back to the blurred 23p even if I set my 24p in the nvidia control panel.
I'm using MPC-HC 1.7.8.162 x64, Lav 0.65-2 and Windows 8.1 x64. I have a gtx 660 with 350.12 and an intel q9550.

[JarrettH]

And because SuperRes includes anti-ringing, can I safely remove Anti-Ringing with Jinc if combined with SuperRes?

I'd like to know too, and what is "error upscaling quality"? It's either medium or high. Sorry if I'm asking about SuperRes too soon.

[MysteryX]

I'd like to know too, and what is "error upscaling quality"? It's either medium or high. Sorry if I'm asking about SuperRes too soon.

If you set it to high, rendering times go up considerably, so it is simply the quality setting.

[Asmodian]

If you set it to high, rendering times go up considerably, so it is simply the quality setting.

It probably changes the error upscale step from bilinear to something else, maybe Jinc3 based on the rendering times?

[huhn]

Funny, I had the exactly same problem with the Nvidia WDDM 2.0 driver. :eek:
Edit: But this was not related to 10 bit, but to the D3D11 renderer in general.

i could get problem with windowed D3D11 too. but 10 bit was on a whole different level. and 88.6 worked pretty good after some tweaks.

but i come back to this when windows 10 is "final".

[BetA13]

hmm,

i just tested teh new version. so far so good, enjoying teh new settings.
i only have 2 problems..

1. the artefacts are back when using nnedi3, both, in chroma upscaling and image doubling.
It was fine with v0.87.21. (was about kepler gtx670 , win 764bit and 32bit..)

2. when using d3d11 mode in FSE/windowed my monitor resolution resets and stuff is cut off..it changes my desktop resolution also and changes teh hz form 60 to 25...
with the normal d3d9 (new path) mode it works fine doh, all 60 hz...

greetz

[bozokaydin]

i amusign windows 8 with gtx 690 nvidia driver 347.88 when i tried to install madvr 88.7 as usually right click and run as administrator to install.bat nothing happens. Just blank command screen. when i hit enter it says you have to right click to install bat and run as administrator. but i already did that. version 87.21 run with no problem no problems. how can i install 88.7 thank you. sorry for my bad english

[XRyche]

It seems Direct3D11 FSE stopped showing an image during playback for me in this release. Direct3D11 FS Windowed works as well as all modes of Direct 3D9. Direct3D11 FSE was working fine in the previous release.

[Moony349]

I don't think DX11 is working for me after updating to 88.7.

FSE mode says "fullscreen exclusive mode new path" instead of FSE 10 bit.

Going back to 88.5 with same settings fixed issue.

I'm on Windows 7 x64, GTX 780, x64 MPC HC.

[MysteryX]

Now I see that "Upscaling Refinement" had to be added to the profile group at the time it was created. Is there a way to add it to existing profile groups without having to delete all existing profiles and restarting all configuration from scratch?

[andybkma]

This version appears to have broken D3D11 entirely on my Win7 GTX770 system, or rather it never activates and madVR always uses D3D9. It works when I reset madVR to defaults though, so it must be something related to my settings/profiles, but so far I've not been able to figure out which causes it.



I would like to check this myself if D3D11 is being used on Win7 but I don't know how. When I press Cntl+J I don't see any OSD for that?

[Balthazar2k4]

I am really appreciating all of the tweaks that madVR offers, but my GTX970 is struggling with 4K upscaling of 1080p BD content. I am now considering a Titan X, but the compute performance concerns me. Should I be looking at AMD cards? I need something quiet and powerful enough to crank up madVR options with 4K upscaling. Suggestions?

[Asmodian]

I am having no trouble with 0.88.7 (32-bit) using D3D11. GK110 (Titan), Win 8.1 x64, 350.12

Now I see that "Upscaling Refinement" had to be added to the profile group at the time it was created. Is there a way to add it to existing profile groups without having to delete all existing profiles and restarting all configuration from scratch?

No, copy your profile rules to a text file first is best you can do.

I would like to check this myself if D3D11 is being used on Win7 but I don't know how. When I press Cntl+J I don't see any OSD for that?

If the line "fullscreen exclusive mode" has a bit depth after it you are using D3D11. If you are in a windowed mode it will say [D3D11] after it.

I am really appreciating all of the tweaks that madVR offers, but my GTX970 is struggling with 4K upscaling of 1080p BD content. I am now considering a Titan X, but the compute performance concerns me. Should I be looking at AMD cards? I need something quiet and powerful enough to crank up madVR options with 4K upscaling. Suggestions?

AMD still has interop issues which slow them down so I don't think AMD is a better option for NNEDI3 image doubling. I think a Titan X is your only upgrade option right now but the 980 Ti and 390X might be released pretty soon, if you can wait a month or two.

I am also pretty sure you could find some settings that look great and still run fine on a 970. :)

[Balthazar2k4]

Thanks Asmodian. I am happy with the settings I am using, but knowing I can do more with a more powerful card is killing me. I am using a Sony VW600ES 4K projector to a 110" screen so every bit of improvement is noticeable. I will most likely go with the Titan X, but wanted to see if there was anything else out there to consider. Guess not.