madVR - high quality video renderer (GPU assisted)

[vivan]

IMO, Nvidia thought they doing us a favor by automatically dithering 16/10 bit content.
Or is it something LAV Filter does?LAV Filter should dither only if output colorspace is not the same as input (video). That means when it's not Y416 for that video.
However that dithering should be in lsb, so if you dither 16 to 10 and then truncate it to 8 it should be the same as truncate it from 16 to 8 directly.
UPD: that's not true. If you have, for example, 1000000 pixel and negative error it will be changed to 01111111...

Yes, although I'm not using hardware decoding.You can't use hardware decoding for anything other than High/Main/Baseline AVC profiles (means 8-bit 4:2:0 only) anyway.

[James Freeman]

Yes, although I'm not using hardware decoding.

Hardware (CUVID, DXVA2) or Software (None), all the same forced dithering in Y416 mode.

I'm on HDMI right now maybe its related?
I'll switch to DisplayPort and report.

EDIT:
No, same forced dithering on DP with Y416.

EDIT: Fixed, DeInterlacing was On.

[Ver Greeneyes]

I'm on HDMI right now maybe its related?
I'll switch to DisplayPort and report.I wouldn't expect it to make a difference - madVR outputs 8-bit per color channel regardless of the input color space.

Please generate the same videos (Grey and Color) at 16-bit 10-bit & 8-bit (4:2:0, 4:2:2 & 4:4:4).Should the input be RGB or YUV? My program currently outputs RGB48LE.

[mithra66]

A1 was a chock when Madshi released it last night. You do such a great job , Madshi!

I Had to wait for night as I couldn't tell the difference even with minimum daylight.

So I prefer A1 > A3 >= A2.

I agree that A1 is darker as if gamma was slightly raised. But it has no inconvenience such as oversaturated color or so on my display. On the contrary.

[James Freeman]

Should the input be RGB or YUV? My program currently outputs RGB48LE.

Same as the current videos.

I'll try to explain what I see and what I'm trying to test:

When I disable Y416, the decoder goes to the closes lower next one: v410.
When v410 is disabled it goes to Y410, and so on till YV24.
Each of these levels has DIFFERENT forced dithering pattern !!!

When all the 4:4:4 row is disabled it goes to the 16-bit RGB -> RGB48 which is not dithered at all.
That is what we are striving for, to let MadVR do the heavy lifting. ;)

I want to test if this forced dithering is somehow related to Nvidia or LAV filters with various formats of video.
These videos should be extremely useful to all video testing, not only this case.

I'll post all the different variants of forced dithering that I see with different output formats.

EDIT: Fixed, DeInterlacing was On.

[madshi]

James, does this have anything to do with madVR? If you want to analyze something about how LAV or NVidia do dithering, then this is not really the right thread for that.

[James Freeman]

Sorry for that madshi.

EDIT:
Nevermind, Solved.

It was the Automatic On DeInterlacing that dithered the 16-bit image.

[Ver Greeneyes]

Same as the current videos.Let me put it in the terms ffmpeg uses. There are YUV444, YUV422 and YUV420 variants for 8-bit, 10-bit and 16-bit. In total there are 9 possible combinations. Which of these do you want?yuv420p
yuv422p
yuv444p
yuv420p10le
yuv422p10le
yuv444p10le
yuv420p16le
yuv422p16le
yuv444p16leEdit: I'll move this to the new thread as well.

[James Freeman]

Ver Greeneyes,

I had DeInterlacing turned on automatically.
This what was dithering the picture.

Vivan was on to it as soon as I posted the first image. :D

[NicolasRobidoux]

Hmmmm... I like madVR's Jinc3 AR better. And NNEDI3 is another class, IMHO...
Note that except for the conversion to linear RGB (which I'm not sure is the best choice with video content when upsampling, but may be with pixel art) what I showed is a purely linear scheme: There is no AR applied whatsoever.

AR is your department. (Until I program my own; don't hold your breath :))

In any case, I tracked down the general purpose scheme I like with pixel art. This one: http://www.wizards-toolkit.org/discourse-server/viewtopic.php?f=22&t=19636&start=30#p78393

Result of runningconvert SNES.png -define filter:lobes=4 -define filter:blur=0.88451002338585141 -filter Lanczos -distort Resize 1234x lanczosSharpest4.sRGB.png
http://web.cs.laurentian.ca/nrobidoux/misc/SNES.lanczosSharpest4.throughRGB.png

Whether, with AR, it satisfies you or others, let alone beats NNEDI, is another story. Also, maybe it should only be applied to luma. But note how noticeably less jaggy it is on perfect diagonals, how tight the halo is, and how sharp it is, compared to the 3-lobe version I showed earlier. Also note that the corresponding disc is barely larger than EWA LanczosSharp, because the deblur is so strong. So it should run reasonably fast.

A slightly less sharp version, EWA LanczosRadius 4 (Jinc-windowed Jinc 4-lobe with support shrunk so it is a disc of radius 4):convert SNES.png -colorspace RGB -define filter:lobes=4 -filter LanczosRadius -distort Resize 1234x -colorspace sRGB SNES.lanczosRadius4.throughRGB.png
http://web.cs.laurentian.ca/nrobidoux/misc/SNES.lanczosRadius4.throughRGB.png

The jist is this: If you deblur EWA Lanczos (Jinc-windowed Jinc) a lot, and you want to do well with thin and artificially sharp diagonals (hello pixel art), 4 lobes is better than 3 lobes. (For natural images, I generally prefer 3 lobes and 5 lobes.)

Yes, these EWA LanczosSharpest/LanczosRadius don't antialias as strongly as, say, EWA LanczosSharp (one version of which is your Jinc3 without AR). But they are tight.

NNTR

[Ver Greeneyes]

I had DeInterlacing turned on automatically.
This what was dithering the picture.Mystery solved!

[wolfman2791]

Ver Greeneyes,

I had DeInterlacing turned on automatically.
This what was dithering the picture.

Vivan was on to it as soon as I posted the first image. :D So i'm assuming the "Auto" option is not what you want in LAV Video settings? So what is the setting that you want??

[James Freeman]

Auto is fine in LAV & MadVR,
but set MadVR to "if in doubt, DE-ACTIVATE deinterlacing".

[leeperry]

And what is your opinion of how A3 compares to NL6?
Well, kinda tricky question as all non-A builds look obsolete now due to the excessive grain but A2 & A3 are wasted potential and a waste of time once you've seen A1 IMHO.

Basically with A1 every moving object has its edges clearly defined but with A2/A3 it's just a big blurry mess that kills the depth impression(call it "3D" if you have to) and the picture dynamics...even RD looks better to my eyes than this big foggy mess.

A1 was a shock when Madshi released it last night. You do such a great job , Madshi!
[..]
I prefer A1 > A3 >= A2.
Wholeheartedly agreed, A1 is da bomb :cool:

[GCRaistlin]

Where can I download old versions of madVR (<0.86.1)?
It's about the issue I've reported above (http://forum.doom9.org/showthread.php?p=1668436#post1668436). MPC-HC developer has confirmed the issue with the latest madVR, ffdshow and MPC-HC but doubts that it's MPC-HC related bug. On the other hand he said that it was not reproducable with the older madVR and/or ffdshow builds though he has had overwritten both of them with the latest versions and he doesn't remember old version numbers. I've tried a lot of versions of ffdshow with madVR 0.86.1, 0.87.3 and 0.87.4 - MPC-HC (as well as MPC-BE) hangs on mouse moving over buttons in the sample DVD root menu.

[James Freeman]

Grey:

A2 is the closest to NL6 in terms of gamma, but it has better blacks.
A1 is similar to A2 but is a little brighter at the middle.
A3 is the brightest in the middle and has similar blacks as A1 and A2.

In terms of Gamma A2=NL6 (darkest or perfect?) > A1 > A3 (brightest).
Also, NL6 has more "holes" (undithered space) compared to the Adaptive builds.

A2 Wins.

Color:

A1 looks too fluffy, not dense enough.
A2 Nice and even.
A3 = A2, look exactly the same.
NL6 looks fluffy like A1 but with cut blacks.

A2, A3 Tie.


Overall A2 has the Positives from NL6 without the Negatives.

[Devrim]

What are the recommended settings with the new update?

Atm I have Jinc3 AR for Chroma and Image upscaling en C-R with AR en SLL for image downscaling but what about NNEDI?
How can we use that?

Also I noticed some banding on a BR I was watching, the artifac removal took care of that, beautifull feature! I set it to high.
Can I leave it at that? Should I set it to low/medium? Or just turn it off and only turn it on when I see banded image?

[Shiandow]

In my opinion artefact removal should be as low as possible, since it can remove some details. I use the default settings and use a hotkey to set it higher when I see banding.

[Devrim]

In my opinion artefact removal should be as low as possible, since it can remove some details. I use the default settings and use a hotkey to set it higher when I see banding.



http://screenshotcomparison.com/comparison/62937
http://screenshotcomparison.com/comparison/62938
http://screenshotcomparison.com/comparison/62939

There are only small differences. (Sorry for the not sharp screenshots with not much detail)

[6233638]

In any case, I tracked down the general purpose scheme I like with pixel art.
...
http://web.cs.laurentian.ca/nrobidoux/misc/SNES.lanczosSharpest4.throughRGB.png
http://web.cs.laurentian.ca/nrobidoux/misc/SNES.lanczosRadius4.throughRGB.pngHonestly, pixel art does not look good when you use video-style filters on it. It's just not how it was intended to look.

Something like this is more appropriate for gaming. (it is expected that you turn up your display brightness to compensate for the scanlines)
http://abload.de/thumb/crt-filterz0e7e.png (http://abload.de/img/crt-filterz0e7e.png)

We really need much higher resolution displays for proper CRT emulation, but some of the approximations are getting pretty good (to the point that I would actually prefer to use them rather than avoid them) and there are probably much better filters now than the one I used for that image.

P.S. I hate the vignetting and rounded-off corners seen here. Some aspects of CRTs are worth emulating (scanlines (http://abload.de/img/00745006-0f18-43ba-86mhj22.jpg), beam width (http://abload.de/img/castlevania3belmontrg8kczo.jpg), pixel structure etc.) others are not. (phosphor persistence, geometric distortion, vignetting, floating black level/low ANSI contrast, composite video connectors etc.)

But this is getting quite off-topic...

[madshi]

Ok, here's one last build, after this I'll probably not have time to do madVR development for 2-3 days. So there will probably be no more test builds in the next 2-3 days at least.

http://madshi.net/madVRadaptive4.rar

This build is a slight modification of the A1 build, in the same way that A3 was a slight modification of the A2 build. So basically A1 -> A4. And A2 -> A3. For the scientific crowd: The modification is that I'm applying a minimum random factor of 0.625f (identical to NL6). The max random factor is 1.0f, and depends on the exact gray level. The difference between A1/4 vs. A2/3 is that the exact random factor is calculated differently.

I think that A3/4 should hopefully combine the positive aspects of NL6 with the adaptive approach. At least I hope so. Whether A3 or A4 is better might be a matter of taste. I think A3 has a slightly more random look. Which could be positive or negative, depending on taste.

[Qaq]

Where can I download old versions of madVR (<0.86.1)?
1st post in the thread says its here: http://www.videohelp.com/tools/madVR/old-versions#download

[madshi]

Note that except for the conversion to linear RGB (which I'm not sure is the best choice with video content when upsampling, but may be with pixel art) what I showed is a purely linear scheme: There is no AR applied whatsoever.

AR is your department. (Until I program my own; don't hold your breath :))

In any case, I tracked down the general purpose scheme I like with pixel art. This one: http://www.wizards-toolkit.org/discourse-server/viewtopic.php?f=22&t=19636&start=30#p78393

Result of runningconvert SNES.png -define filter:lobes=4 -define filter:blur=0.88451002338585141 -filter Lanczos -distort Resize 1234x lanczosSharpest4.sRGB.png
http://web.cs.laurentian.ca/nrobidoux/misc/SNES.lanczosSharpest4.throughRGB.png

Whether, with AR, it satisfies you or others, let alone beats NNEDI, is another story. Also, maybe it should only be applied to luma. But note how noticeably less jaggy it is on perfect diagonals, how tight the halo is, and how sharp it is, compared to the 3-lobe version I showed earlier. Also note that the corresponding disc is barely larger than EWA LanczosSharp, because the deblur is so strong. So it should run reasonably fast.
Ok, with AR it might not look too bad. Not sure whether I would prefer it over Jinc3. Maybe yes for pixel art, but IMHO the NNEDI3 result is still in a whole other league.

Where can I download old versions of madVR (<0.86.1)?
See first post of this thread.

[GREG1292]

Adaptive 3 looks really good. Seems to maintain rich deep colors with good blacks and shadow detail. Clarity/realism is amazing.
Needi3 64 with doubling set to 32/16 always with lancos4 and jinc3. No
problem running 720p/1080p movies. All quality settings unticked. Thanks everyone looks better everyday.


Sent from my iPhone using Tapatalk

[NicolasRobidoux]

Mathias: I am pretty sure I have shown these schemes to you before. The "new" thing (hopefully) is pointing out they are reasonably antialiased for pixel art type objects given how sharp they are. I am not suggesting that they be used for upsampling "standard" video content: Jinc3 AR is about as good as this family of schemes gets.

[MistahBonzai]

I would like to verify a MadVR setting for viewing the various MadVR*.ax variants. It's been mentioned that dithering should be disabled in MadVR however I see no such setting. Under "Trade quality for performance" I can either select use random dithering.. or not. In that case OpenCL error diffusion has been selected. If so then "Use random dithering instead of OpenCL error diffusion" should not be selected. Correct?

[leeperry]

Ok, here's one last build, after this I'll probably not have time to do madVR development for 2-3 days. So there will probably be no more test builds in the next 2-3 days at least.

http://madshi.net/madVRadaptive4.rar

This build is a slight modification of the A1 build, in the same way that A3 was a slight modification of the A2 build. So basically A1 -> A4.
Actually it's more like A4 > A1 to my eyes :devil:

A1 felt a bit like unfinished business, I mean there was less noise but it still felt a bit out of control. With A4, the sensation of depth is back with a vengeance, there's even less noise visible and the picture is much clearer. You are setting new benchmarks everytime, GG :)

there will probably be no more test builds in the next 2-3 days at least.
so how am I gonna live without my daily drastic PQ improvement then?

[kerimcem]

new madvr(all test bulid) work in the old card ? 3650 dont work new madvr..

[XMonarchY]

Actually it's more like A4 > A1 to my eyes :devil:

A1 felt a bit like unfinished business, I mean there was less noise but it still felt a bit out of control. With A4, the sensation of depth is back with a vengeance, there's even less noise visible and the picture is much clearer. You are setting new benchmarks everytime, GG :)


so how am I gonna live without my daily drastic PQ improvement then?

I think he was trying to say that A1 evolved into A4 by typing A1 > A4, where > is a symbol for progression, rather than a symbol for "more/greater". I don't think he was trying to say that A1 is better than A4... but I was wrong so many times before that its likely I am wrong again...

[Asmodian]

I would like to verify a MadVR setting for viewing the various MadVR*.ax variants. It's been mentioned that dithering should be disabled in MadVR however I see no such setting. Under "Trade quality for performance" I can either select use random dithering.. or not. In that case OpenCL error diffusion has been selected. If so then "Use random dithering instead of OpenCL error diffusion" should not be selected. Correct?

IMO dithering should never be disabled in madVR, especially if you can use the great OpenCL (really DirectCompute now) error diffusion dithering. There is an option in the "Trade quality for performance" page under the "Do not use" section to disable all dithering.

If you disable OpenCL dithering in the various MadVR*.ax variants than there is no point to using the variants. We are testing different dithering techniques.

To test the various new dithering methods UNcheck all dither related "Trade quality for performance" options.

new madvr(all test bulid) work in the old card ? 3650 dont work new madvr..

The new stuff doesn't work on that old of a card, both raw GPU speed and DirectX 11 (DirectCompute 5.0) required:
FWIW, I'm playing with DirectCompute now, and it looks promising. But I will limit support to D3D11 capable GPUs, only, so Nvidia 2xx and AMD 4xxx GPUs will not be supported (only for DirectCompute features, of course).

NVidia 4xx should work. NVidia 2xx does not and AMD 4xxx does not. The reason for that is that although DirectCompute in theory supports DX10 GPUs, too, it does so only with a very limited feature set. I'd have to use totally different code to make it all work on DX10 GPUs, and honestly, I think NVidia 2xx and AMD 4xxx users probably don't have the raw speed to make use of the new features, anyway, so I don't feel like spending a lot of additional development resources just to support DX10 GPUs. So I've limited support to native DX11 GPUs.


I think he was trying to say that A1 evolved into A4 by typing A1 > A4, where > is a symbol for progression, rather than a symbol for "more/greater". I don't think he was trying to say that A1 is better than A4... but I was wrong so many times before that its likely I am wrong again...

I think you got it right.

I will have to back leeperry for my vote when deciding on which adaptive build to go for. ;)

[bugmen0t]

Sorry if this question has been asked before. If you're using LAV decoder are you using Ordered Dithering or Random Dithering? I.e. is random dithering redundant and just raising the noise floor or is it dithering something else?

[huhn]

Sorry if this question has been asked before. If you're using LAV decoder are you using Ordered Dithering or Random Dithering? I.e. is random dithering redundant and just raising the noise floor or is it dithering something else?

this option is not used with madvr you can ignore it.

madr supports ll colorspaces lv can output and this option is only used when a yCbCr colorspace is change to rgb and this is not going to happen. except you are playing around with settings just leave it as it is and it fine and not used.

[James Freeman]

A4 looks good in terms of randomness but its also the brightest and has the most abrupt transition to black.

A1 looks good but it lacks dithering in the darkest shades, good gamma curve.
A2 It has the smoothest transition between shades and looks like A1 in terms of gamma.
A3 is a tad brighter than A2.
A4 is even brighter than A3, but it also fixes A1 "holes" lack of dithering.

For me its still A2.

P.S
The Adaptive builds are a HUGE improvement over the NL builds (from what I see).

[ryrynz]

For me its still A2.

Sounds reasonable. Screenshots would be great.

[James Freeman]

Comparison of all the NL1-8 and AD1-4, Grey & Color: ED Builds Test.zip (http://www.mediafire.com/download/ckwo7r4jfd37026/ED_Builds_Test.zip)

* Ver Greeneyes's 16-bit videos are used (Color & Grey).
* Organized in folders according to frame.
* "NL Tests" are frame 88 of Color.

IMO, You should look for smoothness in transition and no abrupt changes or big brightness (gamma) changes.
Also, the dithering pattern should look even and without "holes"** in the dithering pattern.
**(I'm not talking about the undithered places which should be undithered).

Now more than 3 people can decide the future of MadVR. :D

[cyberbeing]

This build is a slight modification of the A1 build, in the same way that A3 was a slight modification of the A2 build. So basically A1 -> A4. And A2 -> A3.

At this point, I'll only say that the A4 is a major improvement over the A1 build on linear gradients. I won't know my preference among all the adaptive builds, until I expand my tests to examine the dithering behavior on real-world videos.

[MistahBonzai]

IMO dithering should never be disabled in madVR, especially if you can use the great OpenCL (really DirectCompute now) error diffusion dithering.

I think we are in 100% agreement when I stated "If so then "Use random dithering instead of OpenCL error diffusion" should not be selected." BTW: I had overlooked the "Don't use dithering" option at the bottom of the "Trade quality for performance" section - not that it was selected.

Bottom Line is that I have spent the whole day attempting to find a combination which yields screen shots demonstrating the effects of the various MadVR*.ax files and have not succeeded. In the past simply deselecting random dithering, capturing the screen with MadVR or Printscreen has worked.

Now for an unknown reason(s) no combination of NV12, YV12, adaptive4, linearlight, MadVR image capture or printscreen (with and without FSE) shows any indication that dithering (error diffusion) is changing. Based on past experience with the MP4-2c gray scale-ramps and test files provided here they have never looked better - while utilizing whatever method they are using (or not) :confused:

Shared folder @Mediafire (https://www.mediafire.com/folder/56j5i4z6vfpto/DirectComputeEDtests)

I'm stumped and this doesn't happen all that often when it comes to 'pixel peeping'. I'm open to all specific suggestions relating to MadVR/LAV settings (I disabled ffdshow and associated AVIsynth scripts right up front).

[James Freeman]

MistahBonzai,

Everything is fine, ED is changing.
The difference is so refined between the latest builds, that its almost invisible with the naked eye (without software intervention).

Its not the big difference we had between, DC3, Noisy & NL builds.
We are at the top of the pyramid (so to speak) now.

[leeperry]

A4 looks good in terms of randomness but its also the brightest and has the most abrupt transition to black.
It might very well be a good thing because when transitions to black aren't "abrupt" this might very well kill contrast and clarity on edges, just a thought as I care a lot more about subjective impressions than overblown screenshots.

All I'm saying is that subjective contrast is amazing with A4(Oblivion looks so darn good, yay!) so this might very well match your findings, 0.2 cents territory here so don't throw rocks at me just yet ;)

The difference is so refined between the latest builds that it's almost invisible with the naked eye (without software intervention).
FWIW I can DBT them anytime you want.

I think you got it right.
It was meant as a joke but sarcasm notoriously doesn't work too well over the internet :p

[James Freeman]

It might very well be a good thing because when transitions to black aren't "abrupt" this might very well kill contrast and clarity on edges,
just a thought as I care a lot more about subjective impressions than overblown screenshots.

All I'm saying is that subjective contrast is amazing with A4 so this might very well match your findings, 0.2 cents territory here so don't throw rocks at me just yet ;)

Me too, but I can't really see the difference without software intervention, so I had to base my conclusions on the overblown images (like everyone else).

Subjective contrast can mean Enhanced Contrast and that's not a good thing for video purists or what MadVR is trying to achieve.
I think we should go for accuracy instead of subjective perception.

What do you think?

FWIW I can DBT them anytime you want.

Hmm....without enhancements?
You seem to be the only one... :)

On my hardware calibrated (i1 Display Pro) monitor, I can't see the difference with my naked eyes.
On what hardware do you compare the builds?
Are you sure your TV/Monitor doesn't do some funky enhancements or something?
Maybe you have the Brightness (TV) set too high so you can clearly see what supposed to be very dark (where the dithering is most prominent)?
Wait, don't tell me you're sitting a foot from a 100" TV? :D

[leeperry]

Subjective contrast can mean Enhanced Contrast and that's not a good thing for video purists or what MadVR is trying to achieve.
I think we should go for accuracy instead of subjective perception.

What do you think?
Accuracy as compared to what? The more it goes the more ED's come with a lower noise floor and consequently an improved sense of depth as dynamics are increased. Don't shoot the messenger, A2/A3 look foggy to me as moving objects seem too blended-in and edges aren't clearly defined anymore. I'll even go for RD over A2/A3, edges are far more discernable than in that big syrupy noise feast.

Please be so kind as to define "video purist", I'm new to this hobby.

Hmm....without enhancements?
I quit doing weed ages ago if that's what you meant, it gave me the munches ;)

You seem to be the only one... :)
Mind you, many ppl see have previously seen differences between those algorithms, but apparently only one person bothered sitting down and scrutinize changes between the A builds so far:
I prefer A1 > A3 >= A2.

On what hardware do you compare the builds?
1GHz HD7850, W7SP1, NNEDI 64x/ED all the way, dead-on REC709 gamut mapped 3500:1 32" A-MVA panel LED TV with BFI turned on from a 80 cm distance in a pitch black room.

[James Freeman]

Please be so kind as to define "video purist", I'm new to this hobby.
Audiophiles, Videophiles...

What comes in, comes out, as the mastering engineers intended.
Properly calibrated equipment (to a standard) without enhancements.
I quit doing weed ages ago if that's what you meant, it gave me the munches ;)
I never did. :D
Mind you, many ppl see have previously seen differences between those algorithms, but apparently only one person bothered sitting down and scrutinize changes between the A builds so far:
What am I not a factor? I gave my best shot.
I voted A2 after a comprehensive test.
1GHz HD7850, W7SP1, NNEDI 64x/ED all the way, dead-on REC709 gamut mapped 3500:1 32" A-MVA panel LED TV with BFI turned on from a 80 cm distance in a pitch black room.
Looks good.

[ryrynz]

Comparison of all the NL1-8 and AD1-4, Grey & Color: ED Builds Test.zip (http://www.mediafire.com/download/ckwo7r4jfd37026/ED_Builds_Test.zip)


Thank you.

Color 250, AD4 wins by a mile, that center bit just get fixed man.
The graduation is much better than the rest IMO. I'm all over AD4.
Only reason I could possibly see passing it up is if it's brighter than it should be..


FWIW I can DBT them anytime you want.


Anyone wanna put together a pot for this? XD

Mean signature. "I'll DBT anytime you want"

[James Freeman]

Thank you.

Color 250, AD4 wins by a mile, that center bit just get fixed man.
The graduation is much better than the rest IMO. I'm all over AD4.
Only reason I could possibly see passing it up is if it's brighter than it should be..



The 250 Color undergone a slightly different procedure than 88 & 150, to see the difference clearer.

Yes, A4 is better than A1.
I wouldn't mind if A4 is selected as the final build, but NOT A1.
I still like A2 the most, but A4 will also do.

[huhn]

1GHz HD7850, W7SP1, NNEDI 64x/ED all the way, dead-on REC709 gamut mapped 3500:1 32" A-MVA panel LED TV with BFI turned on from a 80 cm distance in a pitch black room.

on a 4:2:2 display yeah sounds reasonable.

[James Freeman]

@Ver Greeneyes

Can you please make an, 8-bit yuv444p video of Color & Grey, to compare the dithered 16-bit vs undithered 8-bit at the same frame?
True 8-bit video should not be dithered at all, right?
The purpose of this is to test the Gamma or any other deviations between the builds.

Thanks.

[fairchild]

IMO dithering should never be disabled in madVR, especially if you can use the great OpenCL (really DirectCompute now) error diffusion dithering.

Not entirely true. In my case any added dithering through MadVR raises my black level floor slightly where it doesn't match with the Brightness setting which was calibrated to through a different device such as a pattern generator and/or blu-ray player. So if someone out there has a plasma or any other device that generates it's video through dithering in the panel itself, and they plan to calibrate through MadVR for say an HTPC, then it may be beneficial to disable dithering entirely so that all sources in the display chain match. (OTA TV broadcasts, cable/dish broadcasts, blu-ray player, gaming systems, etc...)

But whenever you are going to watch any video content through MadVR, then of course enable dithering as the banding that is produced without it can be visible in some content. I have tried to tell the difference between all the new ED builds, but frankly I guess my eyes aren't keen enough to see the difference. I'm currently using EDAD4 and have seen no problems with anything I've watched through it. Same goes for all the previous builds which I have tried with regular content.

[James Freeman]

The clearest Comparison: Color 110 AD1-4.zip (http://www.mediafire.com/download/mpvg3llvm8r4hh6/Color_110_AD1-4.zip)

EDIT:
The last comparison had an error, this zip (fixed) should accurately reflect whats going on.

[leeperry]

What am I not a factor? I gave my best shot.
I voted A2 after a comprehensive test.
My bad, I thought you were only nitpicking about screenshots.

I should have also stated that I run Reclock & have "low" debanding forced in mVR.

I wouldn't mind if A4 is selected as the final build, but NOT A1.
A1 is obsolete compared to A4, both technically and visually IMHO.

Anyway, this quote might save the day:
Whether A3 or A4 is better might be a matter of taste. I think A3 has a slightly more random look. Which could be positive or negative, depending on taste.
madshi will not try to convince either of us that strawberry tastes better than banana, maybe A4 + A2(or 3) might be the final ED's http://forum-images.hardware.fr/images/perso/zigg.gif

Anyone wanna put together a pot for this? XD

Mean signature. "I'll DBT anytime you want"
It's rather exhausting to compare them as it takes a lot of concentration but make it interesting and we got a deal ;)

on a 4:2:2 display yeah sounds reasonable.
Yep, chroma is such a big deal when it comes to sharpness. I'll take 4:2:2 BFI over blurry 60Hz-only 4:4:4 LCD anytime of the day. Especially as the TV runs a 10/12bit video engine and I feed it J3AR chroma to begin with. Necessary evil - perfect world, you know the drill.

[cyberbeing]

Using my special method.
...
A1 & A4 are a tad brighter than A2 & A3.

That isn't just a tad brighter, but rather multiple levels brighter. I'd question your special brightening method for causing this discrepancy. You must have brightened A1 & A4 more than A2 & A3 because of noise level differences. In untouched form, all these adaptive routines should have identical overall brightness with 1/100th of a level step accuracy.