madVR - high quality video renderer (GPU assisted) [Archive] - Page 408

ryrynz

13th October 2013, 10:07

Detail loss in "mid/high" should be slightly reduced, too, hopefully without impacting debanding quality too much.

Slightly reduced is a bit of an understatement, it's quite a bit better. I consider the medium mode better than my current f3kd settings.
6233638's custom settings I find are a slight improvement and I'd recommend adopting those for the medium preset.

The "localContrast" parameter can be turned on/off. Turning it on results in doubled rendering times. Please check if this parameter is needed at all. Maybe we can drop it? That would be good for rendering speed.

Remember when I said the picture was noticeably softer in the previous build with deband enabled? Well disabling localcontrast is what's doing just that.
I wouldn't want this removed or disabled, it affects a fair amount of detail in the whole picture.
Personally I'd prefer to have this enabled in all presets, maybe it could be toggled? That softening might suit some sources.. and then there's that performance hit.

Edit: Actually, it seems the new "mid/high" have a lower debanding strength compared to the previous test build. My algorithm changes seem to have lowered overall debanding strength.

Can't win em all, it's not a huge difference and the detail improvement makes up for it. I think the low mode could use some extra deband strength and perhaps something extra could be eeked out from high..
There's always an option of an extra high preset :D

Image anti ringing shows noticeable detail loss though.

If there was noticeable detail loss Madshi, myself or one of the others would have picked it up early on.
Just for curiosities sake I took a 712x384 anime upscaled it with Lanczos 8 tap to 1920x1200 and compared with and without anti-ringing, there is a ever so slight softening of edges, otherwise nothing to really comment on.
I performed an image comparison using Beyond Compare for confirmation and it backs up what my eyes told me.

Knight77

13th October 2013, 11:38

Just FYI The nVidia full RGB fix doesn't work with the last drivers :-(((

bacondither

13th October 2013, 11:43

To my eyes local contrast makes almost no difference and the cost of rendering times (~20% slower) seems to high to justify it at lower(default) settings.

I found the settings avgDif 0,8 maxDif 1,6 gradient 1,5 to be the minimum settings to remove the most of banding without sacrificing any detail.

Some testimages (quite heavy banding)

No debanding: http://postimg.org/image/ld3zit0td/full/
Custom debanding: http://postimg.org/image/k0ydl0mnx/full/
Custom debanding with maximum local contrast : http://postimg.org/image/435quvvk1/full/

mikser

13th October 2013, 12:34

Madshi, I would like to suggest something. I don't know if it is considered to be an actual feature request as it relates to the debanding feature you are already working on. What do you think about the following idea?

With three settings you talked about there are several important problems/points:
1. I (and I am speaking for the average user here) have to know beforehand what kind of video I am going to watch, how much banding it will contain and what kind of debanding algorithm I need. But I don't know this information. Maybe I will know after I watch it, but then I wouldn't need that information anymore.
2. Even if user knows his need, he needs to toggle this setting for each video separately = extra hassle.
3. I usually have a couple of videos started (some podcast to listen to when I am doing something mentally unchallenging like being on facebook or eating; some educational/technique/instructional video when I feel like learning; a couple of movies/TV-shows of which I choose whatever I like to start/continue at that day). Currently it's impossible to have different settings for each video.
4. It's clear that different videos have different debanding needs. If I am watching excellent quality Hi10 video in half of the original resolution - low setting is ok. If I am watching 240p shitty mp2 quality video full-screen on my 1080p, then maybe even high setting is not enough. You get my point.

But how about, additionally to having low, medium and high settings, to also have a "automatic/dynamic" setting. It doesn't have to be perfect. But I believe that it's easy enough to make it way over 90% perfect for your average user.

I won't tell you how exactly to factor in all the ingredients, because you know your stuff and will make it work if you see it feasible. All I am saying is that a simple algorithms which would automatically change the debanding algorithm settings based on "Bits/(Pixel*Frame)" or "Bits/(Pixel*Frame*(upscale/downscale ratio))" would make it reasonably perfect for pretty much every content for the vast majority of users.

What do you think?

madshi

13th October 2013, 13:48

With three settings you talked about there are several important problems/points:
1. I (and I am speaking for the average user here) have to know beforehand what kind of video I am going to watch, how much banding it will contain and what kind of debanding algorithm I need. But I don't know this information. Maybe I will know after I watch it, but then I wouldn't need that information anymore.
2. Even if user knows his need, he needs to toggle this setting for each video separately = extra hassle.
3. I usually have a couple of videos started (some podcast to listen to when I am doing something mentally unchallenging like being on facebook or eating; some educational/technique/instructional video when I feel like learning; a couple of movies/TV-shows of which I choose whatever I like to start/continue at that day). Currently it's impossible to have different settings for each video.
4. It's clear that different videos have different debanding needs. If I am watching excellent quality Hi10 video in half of the original resolution - low setting is ok. If I am watching 240p shitty mp2 quality video full-screen on my 1080p, then maybe even high setting is not enough. You get my point.
I know the problem. That's why I originally only wanted to have 2 settings: One you could always enable without fear of losing detail. And another one which would only be used for especially bad sources.

But how about, additionally to having low, medium and high settings, to also have a "automatic/dynamic" setting. It doesn't have to be perfect. But I believe that it's easy enough to make it way over 90% perfect for your average user.
I wish that were possible, but I see no way to do that. madVR doesn't know which bitrate the video was encoded with. madVR also doesn't know whether dithering was used before encoding or not. I simply don't have any useful information that I could use to automatically choose a debanding setting.

So my recommendation is: Use the "low" setting for all sources, it should not harm. And switch to mid/high only if you find a movie which is especially bad.

Local Contrast definitely seems to help preserve details near black.

From some very limited testing so far, I'm quite liking:
avgDif: 1.0
maxDif: 3.0
gradient: 2.5 (p.s. you have a typo there)
local contrast: 1.5
This seems to have slightly better debanding than medium, and preserves detail much better.

With local contrast set any lower than 1.5, I cannot achieve totally smooth debanding on many of the samples I've looked at. 1.5 seems to be about the threshold for that, but it may have to be set a bit higher.
Thanks, I've modified the "mid" settings to your settings. However, I've still disabled local contrast by default, reason for that see below.

Remember when I said the picture was noticeably softer in the previous build with deband enabled? Well disabling localcontrast is what's doing just that.
I wouldn't want this removed or disabled, it affects a fair amount of detail in the whole picture.
I've just noticed that when you disable "localContrast", the "gradient" check wasn't working correctly. Ooops, my fault. So I'm very sorry, but I fear the test about whether "localContrast" is useful or not needs to be repeated... :(

I found the settings avgDif 0,8 maxDif 1,6 gradient 1,5 to be the minimum settings to remove the most of banding without sacrificing any detail.
Thanks, I've modified the "low" setting accordingly.

-------

So here's a new test build:

http://madshi.net/madVRdeband3.rar

This build comes with the following changes:

(1) increased parameters for "low" slightly
(2) increased parameters for "mid" slightly
(3) fixed bug: "gradient" check didn't work when "localContrast" was turned off
(4) modified algorithm again to get some more debanding strength back, hopefully without losing too much detail
(5) increased max parameters from 4.0 to 40.0

Would be great if you guys could:

- recheck whether "localContrast" is really needed/useful
- double check if the new "low"/"mid" settings are alright now
- find optimal settings for "high"

Since I modified the algorithm again slightly, and due to the gradient bug (with localContrast disabled), unfortunately everything needs to be rechecked, I'm sorry about that.

Werewolfy

13th October 2013, 15:06

I did some testing with your latest build. I find that the low setting is just a little bit too high, it blurs some details sometimes. It's especially the avgDif setting that blurs some details. I set it to 0.5 and there is not harmful, you can increase it to 0.6 but not higher. And yes, debanding is still effective even if it's low.

The mid setting is a good balance between debanding and details loss.

The high setting could be higher for heavy compressed videos. But I don't really have this kind of videos.

The locaContrast setting is maybe not needed. Sometimes it's better with it, sometimes it's worse but the difference is very minimal.

Here some screenshots where you can see that the low setting blurs some details on the car hood and the wipers.

http://imageshack.us/a/img841/1970/ra8z.th.png (http://imageshack.us/photo/my-images/841/ra8z.png/)

http://imageshack.us/a/img69/8526/4a4q.th.png (http://imageshack.us/photo/my-images/69/4a4q.png/)

http://imageshack.us/a/img820/9863/0o6j.th.png (http://imageshack.us/photo/my-images/820/0o6j.png/)

http://imageshack.us/a/img716/1236/24i6.th.png (http://imageshack.us/photo/my-images/716/24i6.png/)

A video sample if it's needed : http://www.mediafire.com/?jyjk691ddd4jpbh

Here just to see the debanding differences between all the settings.

http://imageshack.us/a/img707/1746/p2x4.th.png (http://imageshack.us/photo/my-images/707/p2x4.png/)

http://imageshack.us/a/img407/1619/si2p.th.png (http://imageshack.us/photo/my-images/407/si2p.png/)

http://imageshack.us/a/img43/1277/b79h.th.png (http://imageshack.us/photo/my-images/43/b79h.png/)

http://imageshack.us/a/img266/9758/ztn6.th.png (http://imageshack.us/photo/my-images/266/ztn6.png/)

http://imageshack.us/a/img30/3391/jsb.png

Ver Greeneyes

13th October 2013, 15:18

There is no setting in collink to generate a 3dlut with Rec709 color space, and preserve my calibration gamma curve from my .icm profile.
Ah, it's in the .icm file? That makes much more sense! So what you could try is to use the ICCXmlTools to convert both your icm profile and Rec709.icm to xml, then replace the rTRC, gTRC and bTRC section in Rec709.xml with the ones from your profile - then convert the result back to .icm and use that to generate the 3DLUT.

6233638

13th October 2013, 16:12

I've just noticed that when you disable "localContrast", the "gradient" check wasn't working correctly. Ooops, my fault. So I'm very sorry, but I fear the test about whether "localContrast" is useful or not needs to be repeated... :(It definitely still helps with shadow detail at higher deband settings. I'm not sure whether lower settings without localContrast would produce the same results though, I've only done a small amount of testing so far.

- double check if the new "low"/"mid" settings are alright nowThe minimum settings I found which produced good debanding results were:
avgDif 0.8, maxDif 2.5, gradient 1.5, localContrast 1.5

http://www.abload.de/thumb/no-debandfvpe3.png (http://www.abload.de/img/no-debandfvpe3.png) http://www.abload.de/thumb/min-debandhhrux.png (http://www.abload.de/img/min-debandhhrux.png) http://www.abload.de/thumb/low-debandgho17.png (http://www.abload.de/img/low-debandgho17.png)

I would say that this pattern is probably the smallest amount of banding you are likely to encounter as it seems to be truly limited by 8-bit.
These settings are right on the edge of eliminating the banding, and still being able to make out bands.
If you get close, you can still make out very slight amounts of banding that remain with these settings.

If you want the absolute lowest settings which show improvement, they are:
avgDif 0.6, maxDif 2.4, gradient 1.2, localContrast off (I have not yet checked to see if it helps with these settings)

http://www.abload.de/thumb/absolute-minimumqwold.png (http://www.abload.de/img/absolute-minimumqwold.png)

As you can see, they do not eliminate all banding though, just reduce it.
If you reduce any of these parameters by 0.1, you will then see a large increase in banding, which is why I would say these are the absolute minimum useful settings.

I really need to collect a number of samples which show detail loss though. I have a few that I'm checking, but need to go through my collection to see if there are any good tests.
I don't know that I have any more time to spend on it today though.
I did some testing with your latest build. I find that the low setting is just a little bit too high, it blurs some details sometimes. It's especially the avgDif setting that blurs some details. I set it to 0.5 and there is not harmful, you can increase it to 0.6 but not higher. I can't view your full images (imageshack never works here) but from looking at your sample, all you're seeing between 0.5 and 0.6 is distinct macroblocking and debanded macroblocking. I don't know that I would call it a loss of detail.

Knight77

13th October 2013, 16:21

Just FYI The nVidia full RGB fix doesn't work with the last drivers :-(((

Guys, any solution to this?

bacondither

13th October 2013, 16:44

madshi

13th October 2013, 17:07

Guys, any solution to this?
Try using a custom resolution, that has at least worked in the past.

-------------------------

@Guys, please continue testing debanding, and post your favorite settings. I'll probably not have much time for madVR until the next weekend. So by then I'll browse through your reports and try to find a common ground for the debanding settings.

Nobody has yet recommended new settings for the "high" setting. I think debanding strength should be increased for "high", but I don't have good settings so far.

Also useful to know would be if "localContrast" is maybe only needed at "high" settings, or whether we also need it for "mid" and "low". Of course I could add an option to enable/disable "localContrast", but I'd really like to keep the options to a minimum. So if I could e.g. skip "localContrast" for at least "low", maybe also for "mid", that would make things easier for the average user, and also produce a nice rendering speedup. But of course if the quality loss is too big, I can't do that...

Werewolfy

13th October 2013, 17:10

I would not call that detail loss, seems to be compression artifacts and banding due to low bitdeph that is slightly blurred, i would call that a positive effect.

And with a setting of avgDif of 0.5 you can still se alot of banding in your test image http://imageshack.us/f/43/b79h.png/ compared to http://imageshack.us/f/407/si2p.png/
The lowest avgDif could work satisfactory is >0.7

Some testing images that are contrast boosted to easier spot changes in dark areas:

No debanding: http://postimg.org/image/x56k1bxcf/full/
avgDif 0.5: http://postimg.org/image/tu9nft0m9/full/
avgDif 0.8 "low": http://postimg.org/image/ruslgm83l/full/

I agree with you, there is a little more banding with avgDif on 0.5 but on 0.7 the shadows of the wipers are blurred, it's not compression artifacts.
Personnaly I think that if you want a setting that doesn't blur ANY detail even the small one, it's 0.6.

0.7 seems to be an acceptable trade-off for most people but 0.8 is too high for me.

THX-UltraII

13th October 2013, 20:12

Gagorian

13th October 2013, 21:00

question for madVR + Intel 2000HD gpu users:

someone on the avs community claims that the intel gpu outputs ycbcr and not rgb. He says this affects pq in a bad way.

Can someone confirms if Intel gpu's do indeed not output rgb and if this is indeed bad for pq?

You shouldn't post the same question multiple times. Anyway, have you tried the instructions here?

http://www.avsforum.com/t/1477460/theory-about-intels-hdmi-quantization-range-setting-full-0-255/30#post_23789526

or here

http://hardforum.com/showpost.php?p=1039260834&postcount=325

But considering the quality of your projector I would suggest you consider getting a discrete GPU so you have enough processing power for proper upscaling etc.

bacondither

13th October 2013, 21:56

I have played around with the settings in the new test build(V3) and i think that avgDif: 0.7 maxDif: 1.5 gradient: 1.4 localContrast: off is the lowest setting that removes most of the banding in most material i tested and keeps blurring to a minimum/nonexistence.

It should make a good low "default" setting for most material i think, if no more changes are made to the algorithm.

turbojet

13th October 2013, 22:58

I'm not saying there's an issue with anti-ringing, just that the detail loss in anti-ringing which I've seen referred to as taking a little 'fuzz on the peach' is more noticeable to me then the loss of detail from debanding when sandwiched between lumasharpen. It's much more noticeable on grainy video and at 24 fps instead of a still frame but here's some pics:
http://postimg.org/image/h6puk4rzr/
http://postimg.org/image/int5m8e3t/
http://postimg.org/image/vrhybssa5/

The deband-samples are good but I think bluray, digital factory and focus features is intentional 'contouring' I think it's called in the video world much like https://en.wikipedia.org/wiki/Contour_line, given the perfect curves. This is another example of it: http://www.sendspace.com/file/0yw6pd How is something like this supposed to be debanded without added grain or removed completely?

Zachs

14th October 2013, 00:34

Guys, any solution to this?

I'm using 331.40 Beta. Mad's PC level 0-255 registry tweak works fine for me.

6233638

14th October 2013, 07:14

I've spent a bit more time with this now, and so far these are the settings I'm fairly happy with:

avgDif maxDif grad LC
Min: 0.6 2.5 1.5 off
Low: 0.8 2.5 1.5 1.5
High: 1.9 3.7 4.3 off

The minimum settings are the lowest values for each parameter that I could find which really did any kind of effective debanding.

The slight differences between Min/Low increase the effectiveness of debanding quite a lot, without much change to detail as local contrast is enabled.

For the High settings, I was looking at quite highly compressed sources such as YouTube videos.
These are the lowest settings I found which had effective debanding for this type of content.

I need to spend a bit more time with a variety of sources to see what would be effective medium settings - perhaps they're good enough as they are.

Of course these values may end up changing, but I did try quite a lot of content when testing these, which takes a long time when you can't save your custom settings.

Also useful to know would be if "localContrast" is maybe only needed at "high" settings, or whether we also need it for "mid" and "low". Of course I could add an option to enable/disable "localContrast", but I'd really like to keep the options to a minimum. So if I could e.g. skip "localContrast" for at least "low", maybe also for "mid", that would make things easier for the average user, and also produce a nice rendering speedup. But of course if the quality loss is too big, I can't do that...I actually find that Local Contrast is effective at low/medium settings, and not at higher settings.

Sources which would use higher settings need stronger levels of debanding, and Local Contrast works against that.
When high amounts of debanding are used, Local Contrast has to be set so high that there's really no difference between it being enabled or not.

Perhaps it belongs in the "trade quality for performance" section?

To my eyes local contrast makes almost no difference and the cost of rendering times (~20% slower) seems to high to justify it at lower(default) settings.

I found the settings avgDif 0,8 maxDif 1,6 gradient 1,5 to be the minimum settings to remove the most of banding without sacrificing any detail.

Some testimages (quite heavy banding)
No debanding: http://postimg.org/image/ld3zit0td/full/
Custom debanding: http://postimg.org/image/k0ydl0mnx/full/
Custom debanding with maximum local contrast : http://postimg.org/image/435quvvk1/full/With Local Contrast, lower values increase the amount of shadow detail preserved. Higher values like 4.0 will have little effect.

That said, with carefully tuned settings, local contrast is rarely needed, and if it's set too low, it undoes much of the debanding effect.

The deband-samples are good but I think bluray, digital factory and focus features is intentional 'contouring'I don't think it's intentional at all, just graphics which were not created with high enough precision. But now that we have the option of custom settings, I don't think there's anything which does effective debanding of those samples with acceptable image quality.

http://www.sendspace.com/file/0yw6pd How is something like this supposed to be debanded without added grain or removed completely?If your priority is debanding/deblocking, even at the expense of detail, 5.0/14.5/12.0 seems to take care of most of it.
Local Contrast settings of 1.1 or 2.3 will both preserve some amount of detail at the expense of increased blocking/banding. I'd probably leave it off.
It's always a trade-off when dealing with bad sources.

I agree with you, there is a little more banding with avgDif on 0.5 but on 0.7 the shadows of the wipers are blurred, it's not compression artifacts.
Personnaly I think that if you want a setting that doesn't blur ANY detail even the small one, it's 0.6.If you don't want any detail loss, then you have to disable debanding. It's all a balance between detail loss and banding.

The "wipers" themselves are just big macroblocks in your sample though, all one solid color. Here it is brightened up:

http://www.abload.de/thumb/wipersclk8z.png (http://www.abload.de/img/wipersclk8z.png) http://www.abload.de/thumb/wiper-debanda4jl5.png (http://www.abload.de/img/wiper-debanda4jl5.png)

I wouldn't call it detail loss, just debanding at work, smoothing the edges.
You might also want to check your display's calibration if those are easy to see without being brightened like that.

ryrynz

14th October 2013, 07:38

more noticeable to me then the loss of detail from debanding

I'm looking at your PNG images and the "detail loss" from anti-ringing is approaching non existent, especially on the left side there's so little change that it's almost identical.
The right side has some extremely minor edge softening which in my experience is as expected.
Looking one foot away from a 24" monitor comparing these two pics and while I can spot some minor differences they are so minor I honestly I see no issue at all.

Regarding debanding, you should not use the high preset on this sort of material also you should not expect zero loss of detail when debanding with any quality preset.
The trick here is to gain more (debanding) than you give up (sacrifice of non banded detail)

THX-UltraII

14th October 2013, 08:50

You shouldn't post the same question multiple times. Anyway, have you tried the instructions here?

http://www.avsforum.com/t/1477460/theory-about-intels-hdmi-quantization-range-setting-full-0-255/30#post_23789526

or here

http://hardforum.com/showpost.php?p=1039260834&postcount=325

But considering the quality of your projector I would suggest you consider getting a discrete GPU so you have enough processing power for proper upscaling etc.

These threads are only involving FULL or LIMITED output range. I m just asking (politely :)) if there is someone here with a Intel HD graphics card that can confirm if the Intel cards do only output ycbcr and cannot output rgb.

See also this post by user huhn:
http://forum.doom9.org/showpost.php?p=1647304&postcount=20278

huhn says:
''madvr outputs rgb and your intel card transforms this to ycbcr with is bad...''

So my madVR+intel HD grpahics (maybe noob) conclusion is that this is a permanent no-go option.

But maybe I m missing something here.

baii

14th October 2013, 08:54

madshi

14th October 2013, 10:59

Hi, I am trying to get smooth playback for 1080i30 material scale to 1440p using
hardware deinterlace jinc3 chroma + jinc3/ar upscale

I assume the total time spend for a frame need to below 1000ms/60 =16ms?

About what level of gpu can do this task?
Yes, max 16ms per frame. I can't tell you which GPU would be needed for this. FWIW, chroma is not as important as luma/image upscaling. Maybe someone else can chime in with information about which GPU would be needed for this?

-------

In order to make deband testing more efficient, I've implemented a few small changes:

http://madshi.net/madVRdeband4.rar

(1) There's no "low", "mid" or "high" in this build. Only "off" and "custom".
(2) The "custom" parameter settings are automatically stored (HKCU\Software\madshi\madVR), so you can switch video files without losing your custom settings.
(3) You can now turn on/off every check separately. Basically Ctrl+Alt+SomeNumber increases/decreases a specific parameter and Ctrl+Alt+Shift+SomeNumber enables/disables the same check. This way you can check whether maybe some checks are not needed at all, anymore, after tweaking the other parameters.
(4) I've added one more check, named "nonGradientPenalty". A value of 1.0 means the check is "off". Higher values decrease debanding and increase detail preservation. This check is related to the "gradient" check. So if you turn off the "gradient" check, the "nonGradientPenalty" check is not working, either. I'm not sure if this new check is useful or not. Maybe it's not useful. Could you please check whether increasing this from 1.0 to some higher values improves the "debanding vs detail" equation or not?

Thanks!

This is probably the last test build for this week.

P.S: A good way to check the new "nonGradientPenalty" would be to increase both "gradient" and "nonGradientPenalty" at the same time. Doing so might eventually result in a similar debanding strength, but maybe less detail loss. Or maybe not. These two (gradient and nonGradientPenalty) belong together, though. So you could limit testing to changing these two, while keeping the others the same. This way you don't have to test so many different combinations...

huhn

14th October 2013, 11:18

These threads are only involving FULL or LIMITED output range. I m just asking (politely :)) if there is someone here with a Intel HD graphics card that can confirm if the Intel cards do only output ycbcr and cannot output rgb.

See also this post by user huhn:
http://forum.doom9.org/showpost.php?p=1647304&postcount=20278

huhn says:
''madvr outputs rgb and your intel card transforms this to ycbcr with is bad...''

So my madVR+intel HD grpahics (maybe noob) conclusion is that this is a permanent no-go option.

But maybe I m missing something here.
you can change this in the intel options i have my display connected through vga (limited issue...) so i can't see this option at the moment.

but for me this was allways limited even with my pc monitor.

i have not tryed this :

Anyone, who want FullRange RGB HDMI output on Intel Graphic may try this:
1. Find your current Intel Graphic Adapter software registry instance under:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E968-E325-11CE-BFC1-08002BE10318}\xxxx
(0000 on my PC).
2.Create DWORD value:
EnableRGBFullRange
and set it to 1
3. Reboot.
4. Enjoy.

i simply gave up and got a new gpu

mindz

14th October 2013, 12:17

I just bought an nVidia videocard after having owned ATI for the most part, last nVidia card i owned was the first GeForce (cool!). Ive read and seen the 0-255 registry tweak software madshi ships with madVR for nVidia users. Do i need this? To my eyes it looks exactly the same as my ATI 5770 did before this, and i had everything on 0-255, just like now. Any info on this? Why do i need this?

bacondither

14th October 2013, 12:24

(3) You can now turn on/off every check separately. Basically Ctrl+Alt+SomeNumber increases/decreases a specific parameter and Ctrl+Alt+Shift+SomeNumber enables/disables the same check. This way you can check whether maybe some checks are not needed at all, anymore, after tweaking the other parameters.

I cannot find any key to disable localContrast, so i have to disable it in the registry every time i want to disable it.

6233638

14th October 2013, 12:28

In order to make deband testing more efficient, I've implemented a few small changes:

http://madshi.net/madVRdeband4.rar

(1) There's no "low", "mid" or "high" in this build. Only "off" and "custom".
(2) The "custom" parameter settings are automatically stored (HKCU\Software\madshi\madVR), so you can switch video files without losing your custom settings.
(3) You can now turn on/off every check separately. Basically Ctrl+Alt+SomeNumber increases/decreases a specific parameter and Ctrl+Alt+Shift+SomeNumber enables/disables the same check. This way you can check whether maybe some checks are not needed at all, anymore, after tweaking the other parameters.
(4) I've added one more check, named "nonGradientPenalty". A value of 1.0 means the check is "off". Higher values decrease debanding and increase detail preservation. This check is related to the "gradient" check. So if you turn off the "gradient" check, the "nonGradientPenalty" check is not working, either. I'm not sure if this new check is useful or not. Maybe it's not useful. Could you please check whether increasing this from 1.0 to some higher values improves the "debanding vs detail" equation or not?

Thanks!

This is probably the last test build for this week.Thanks for the new build, this should definitely make testing a lot easier. I had been setting parameters to 40 and making adjustments from there, and now I can just turn a parameter off for the same effect. (actually seems to be higher than 40 in some cases)

Saving between videos, and having an on/off toggle rather than cycling through everything will make things much faster as well.

ryrynz

14th October 2013, 12:28

I cannot find any key to disable localContrast, so i have to disable it in the registry every time i want to disable it.

Ctrl+Alt+Shift+SomeNumber enables/disables the same check.

Ctrl Alt Shift 9.

bacondither

14th October 2013, 12:37

Ctrl Alt Shift 9.
Thanks! I blame myself for not finding it on my deficiency of coffee :o

THX-UltraII

14th October 2013, 13:00

you can change this in the intel options i have my display connected through vga (limited issue...) so i can't see this option at the moment.

but for me this was allways limited even with my pc monitor.

i have not tryed this :

i simply gave up and got a new gpu
I can choose the quantization AUTO, LIMITED and FULL in the intel display options for my digital display.

However, I thought that you are saying that the Intel cards cannot output RGB and can only output YCBCR which you say is bad for PQ.

Edit:
Huhn, on the AVS forum in the intel quantization topic people claim that there is no 'RGB problem' with Intel gpu's.....:
http://www.avsforum.com/t/1477460/theory-about-intels-hdmi-quantization-range-setting-full-0-255/120#post_23834415

andybkma

14th October 2013, 14:48

i have not tryed this :

Anyone, who want FullRange RGB HDMI output on Intel Graphic may try this:
1. Find your current Intel Graphic Adapter software registry instance under:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E968-E325-11CE-BFC1-08002BE10318}\xxxx
(0000 on my PC).
2.Create DWORD value:

EnableRGBFullRange
and set it to 1
3. Reboot.
4. Enjoy.

i simply gave up and got a new gpu

I can confirm this Intel Graphics registry tweak works :-) I now have full color range HDMI to HDMI 0-235 using my Intel HD4000 without having to make that custom resolution with CVT-RB. Happy days

Thanks to SweetLow's solution at this thread:

http://www.avsforum.com/t/1477460/theory-about-intels-hdmi-quantization-range-setting-full-0-255

Sorry if off-topic, madshi. Just think that some mVR users might not know about this yet (and I see your fingerprints on that thread too ;-)...

THX-UltraII

14th October 2013, 14:54

I m not talking about the problem FULL and LIMITED. This already works out-of-the-box with the latest Intel driver (you can simply select AUTO, FULL or LIMITED in the Intel Control Panel and it works perfect).

I am talking about huhn saying that Intel cards output ycbcr and not rgb and that this is supposed to affect PQ in a bad way when using madVR.

andybkma

14th October 2013, 16:31

I m not talking about the problem FULL and LIMITED. This already works out-of-the-box with the latest Intel driver (you can simply select AUTO, FULL or LIMITED in the Intel Control Panel and it works perfect).

I don't have these options in my Intel driver control panel. Where exactly are they located? I am already using the latest Intel drivers for my HD4000 with Win7 64bit that I could find on Intel's site (9/11/2013, 15.31.17.64.3257). I don't have any auto, full or limited options or I need clear cut directions to find them ;-)

Also, my post wasn't directed at you but mVR users in general who have struggled as I have to achieve proper blacks with their HTPC....

Werewolfy

14th October 2013, 17:29

If you don't want any detail loss, then you have to disable debanding. It's all a balance between detail loss and banding.

The "wipers" themselves are just big macroblocks in your sample though, all one solid color. Here it is brightened up:

http://www.abload.de/thumb/wipersclk8z.png (http://www.abload.de/img/wipersclk8z.png) http://www.abload.de/thumb/wiper-debanda4jl5.png (http://www.abload.de/img/wiper-debanda4jl5.png)

I wouldn't call it detail loss, just debanding at work, smoothing the edges.
You might also want to check your display's calibration if those are easy to see without being brightened like that.
You're right, some "details" will always be lost with debanding. I was trying to find a setting without any detail loss but in that case, the debanding strenght is very limited.

Here my current settings :
low : avgDif: 0.7 maxDif: 2.5 gradient: 1.5 nonGradientPenalty : 1.2
mid : avgDif: 1.2 maxDif: 3.2 gradient: 2.8 nonGradientPenalty : 1.3
high : avgDif: 2.3 maxDif: 4.0 gradient: 4.3 nonGradientPenalty : 1.3

For low, the debanding is effective and there is almost no detail loss. At higher values, I find the picture too smooth on certain places. Lower values exhibit a lot more banding.
For mid, the debanding is very effective on most videos. There is some detail loss but for Anime it's a perfect setting. There is less details generally in Anime than movies.
For high, I aim a very effective debanding on bad sources. But I need more bad videos to test it. If anyone has links to that kind of videos, I'd appreciate it.

I like the new nonGradientPenalty setting. It allows to maintains some details that are usually blurred by the gradient setting and maintaining good debanding at low values. But some people might desactive it for high because it maintains sometimes compression artifacts.
I still find the LocalContrast quite useless. There is almost no difference with or without and when there is a difference, it's very minimal.

madshi

14th October 2013, 17:51

Interesting, you guys are getting nearer to each other in your parameter choices. @6233638, what is your opinion about Werewolfy's current settings? They're relatively near to yours, but not totally identical. I would really like to get rid of localContrast, but I'm not sure if I can. I do wonder whether I should have 3 or 4 settings. It seems that Werewolfy prefers "avgDif 0.7" and 6233638 prefers "avgDif 0.8" for the "low" setting. Maybe a "very low" setting might make sense. I'm thinking that if we get native 4K content sooner or later with native 10bit decoding then maybe a "very low" setting might make sense to improve 10bit gradients slightly. For clean native 10bit sources "low" might be overkill. Of course we don't have clean native 10bit sources at the moment (do we?), so maybe it's hard to find good parameter values for that...

bacondither

14th October 2013, 19:06

I agree on avgDif 0.7 on a low setting, but values >1.4-1.5 for maxDif and >1.3-1.4 for gradient seems to have quite a steep diminishing return on debanding. Atleast for a low setting.

/errata
For some low resolution material it seems that maxDif would have to be atleast 1.8 to have sufficient effect.

With nonGradientPenalty over 1.1 i can see some miniscule blocking and slight banding sometimes. 1.2 would be max. Atleast at a low setting. I would use 1.1 or disabled.

Test scene 1:
No debanding (http://postimg.org/image/i1qel7cg1/full/)
avgDif 0.7 maxDif 1.5 gradient 1.4 (http://postimg.org/image/ier64z5i9/full/)
avgDif 0.7 maxDif 2.5 gradient 1.5 (http://postimg.org/image/f19p39n83/full/)
avgDif 0.7 maxDif 2.5 gradient 1.5 nonGradientPenalty 1.2 (http://postimg.org/image/ofxzu7gi3/full/)

Test scene 2:
No debanding (http://postimg.org/image/3qkoliyf5/full/)
avgDif 0.7 maxDif 1.5 gradient 1.4 (http://postimg.org/image/i2j31v6tr/full/)
avgDif 0.7 maxDif 2.5 gradient 1.5 (http://postimg.org/image/nggw1swk3/full/)
avgDif 0.7 maxDif 2.5 gradient 1.5 nonGradientPenalty 1.2 (http://postimg.org/image/9y0nzfjk1/full/)

madshi

14th October 2013, 20:55

@bacondither,

I think the other guys have also started with relatively low "maxDif" values, but have increased them recently. Of course a lot depends on the source content you test with. With some material low values work ok, with other material higher values are needed. Basically the flatter a banded gradient in the image is, the lower "maxDif" can be. If there is a rather strong gradient in the image, "maxDif" needs to be bigger for debanding to work correctly.

About "nonGradientPenalty": If you see too much blocking/banding with 1.2 or 1.3, then you can either turn "nonGradientPenalty" back to 1.1 or even 1.0. Or alternatively you could leave it at 1.2 or 1.3 and instead turn "gradient" further up, since "gradient" and "nonGradientPenalty" are related. Blocking/banding caused by "nonGradientPenalty" can be removed by increasing "gradient". The big question is which combination of "gradient" and "nonGradientPenalty" produces the better overall result (best debanding with the lowest detail loss). I'm not sure about that.

Werewolfy

14th October 2013, 22:11

maxDif doesn't seem to harm the picture so I don't see why reducing it. Of course, maybe it harms the picture but not on the samples I have so feel free to test it on your samples and tell if it harms the pictures with the values I gave for example.

nonGradientPenalty seems to help at low values, I'd have not so high values for Gradient otherwise. The only side effect is that the compression artifacts are not so well hidden when it's activated so maybe it's not suited for the high setting.

Interesting, you guys are getting nearer to each other in your parameter choices. @6233638, what is your opinion about Werewolfy's current settings? They're relatively near to yours, but not totally identical. I would really like to get rid of localContrast, but I'm not sure if I can. I do wonder whether I should have 3 or 4 settings. It seems that Werewolfy prefers "avgDif 0.7" and 6233638 prefers "avgDif 0.8" for the "low" setting. Maybe a "very low" setting might make sense. I'm thinking that if we get native 4K content sooner or later with native 10bit decoding then maybe a "very low" setting might make sense to improve 10bit gradients slightly. For clean native 10bit sources "low" might be overkill. Of course we don't have clean native 10bit sources at the moment (do we?), so maybe it's hard to find good parameter values for that...

I thought also you could add a very low setting but now it seems useless to me. It's true that we can find a lower setting than the low setting I gave for example but in that case the debanding effect is quite minmal so yes maybe for native 4K content 10bit sources it can ben useful but for now, I don't really see the point.

Skankee

14th October 2013, 23:57

I compared Werewolfy's and 6233638's low-debanding settings on high-quality videos to see if (low) debanding has a negative influence on picture-quality.

There was not much banding in my videos, so i couldn't see a difference between Werewolfy's and 6233638's low-debanding settings. In almost all examples the picture-quality was as good as without debanding, + noticeable debanding in some cases.

But in really dark areas i noticed something i dont like.

For me it looks like the "bright" areas stay bright and the dark areas become much brighter, so that the video seems partly brighter in some areas with active debanding.
I expected that the "bright" areas would be a little bit darker, so that the average luminosity stay the same (visual).

No debanding (part):
http://s1.directupload.net/images/user/131015/6p8eeupz.png

low-debanding (part):
http://s1.directupload.net/images/user/131015/74cb7nwh.png

I looked at the luminosity-histogram and the average luminosity is almost identical. But especially the lower part in the middle looks too bright for me with debanding.

Full-size:
No debanding:
http://s1.directupload.net/images/user/131015/t6uej6ba.png

Werewolfy low-debanding:
http://s14.directupload.net/images/user/131015/v45eysk9.png

6233638 low-debanding:
http://s14.directupload.net/images/user/131015/cc3qxczb.png

6233638

15th October 2013, 04:47

@6233638, what is your opinion about Werewolfy's current settings? They're relatively near to yours, but not totally identical. I would really like to get rid of localContrast, but I'm not sure if I can.Unfortunately, I'm still finding sources where localContrast seems to be useful, and it doesn't seem like nonGradientPenalty will completely replace it.

I must say, I'm finding it difficult to see anything but very subtle changes when adjusting nonGradientPenalty with the sources I've looked at, but 1.2 does seem to have some benefit at low settings.
I'm not sure that I would want to use high levels of gradient and nonGradientPenalty together, because high levels of gradient seem to add a lot of dither/noise to the image.

I really need to keep better track of my testing, because I'm not sure what video(s) it was that caused me to raise the low avgDif to 0.8 before
So unless I can find another example of that, 0.7 seems acceptable.
Using 0.7 as the avgDif, 2.4 seems like the logical choice for maxDif, and a gradient of 1.6 seemed to be as low as I could go with those before debanding effectiveness was reduced.

I'm somewhat flexible on the nonGradientPenalty and localContrast settings, but 1.2/1.4 (possibly 1.5) seemed to have little impact on debanding effectiveness, while helping preserve some shadow detail.
avgDif maxDif grad penalty contrast
low: 0.7 2.4 1.6 1.2 1.4
maximum: 1.9 3.7 4.3 off off

I need to do more testing to find the optimal settings for a "medium" preset, but I'm thinking something around an avgDif of 1.2 seems like a good choice. I would not want any lower.

I'm still quite happy with the max settings I posted before, and haven't found a source where Werewolfy's stronger settings are required.

I don't know that you would need separate presets which use localContrast, just have presets with localContrast values (and this would likely only be low/medium) and a "trade quality for performance" that disables the localContrast portion if required.

I agree on avgDif 0.7 on a low setting, but values >1.4-1.5 for maxDif and >1.3-1.4 for gradient seems to have quite a steep diminishing return on debanding. Atleast for a low setting.I think it depends on what you consider a "low" setting to be useful for.
If we are going to use an avgDif of 0.7, I think maxDif should be 2.4. Going any higher does not result in much improvement for debanding, and going lower is much less effective.

http://www.abload.de/thumb/maxdif-15zesk1.png (http://www.abload.de/img/maxdif-15zesk1.png) http://www.abload.de/thumb/maxdif-24wtsqq.png (http://www.abload.de/img/maxdif-24wtsqq.png)

Test scene 1:
Test scene 2:
Can you please post video samples for these. I do not find your example images to have sufficient debanding, and would really like to experiment and see what settings would be effective with them.

I compared Werewolfy's and 6233638's low-debanding settings on high-quality videos to see if (low) debanding has a negative influence on picture-qualityCan you post a video sample of this? It seems like a good test for medium settings.

bacondither

15th October 2013, 09:03

Unfortunately, I'm still finding sources where localContrast seems to be useful, and it doesn't seem like nonGradientPenalty will completely replace it.
Could you provide an example of that please?
If we are going to use an avgDif of 0.7, I think maxDif should be 2.4. Going any higher does not result in much improvement for debanding, and going lower is much less effective.

http://www.abload.de/thumb/maxdif-15zesk1.png (http://www.abload.de/img/maxdif-15zesk1.png) http://www.abload.de/thumb/maxdif-24wtsqq.png (http://www.abload.de/img/maxdif-24wtsqq.png)

Are those images from the AVS HD 709 grayscale ramp?, and if so i too found maxDif 2.4 to remove most bands in that syntetic test image. But you almost have to be pixel peeping to see it.
It seems that the detail reduction by using maxDif 2.4 over 1.5-1.6 is so miniscule it might be a good vaule for a low setting.
Can you please post video samples for these. I do not find your example images to have sufficient debanding, and would really like to experiment and see what settings would be effective with them. You cannot eliminate all banding all the time without destroying the image, especially if you try to tune to remove every trace of banding in clips of quite high banding that i will post below.
In a low setting it "should" try to hide the banding from "eeeew banding" to "this looks acceptable", or atleast that is what i think. :)

The clips:
Test scene 1 (http://www.sendspace.com/file/4td7ab)
Test scene 2 (http://www.sendspace.com/file/kd119r)

Additional clips:
Test scene 3 (http://www.sendspace.com/file/6ik5m2)
Test scene 4 (http://www.sendspace.com/file/55yctr)

ryrynz

15th October 2013, 09:39

Unfortunately, I'm still finding sources where localContrast seems to be useful, and it doesn't seem like nonGradientPenalty will completely replace it.

Agreed, even with localContrast fixed I'd still prefer it to stay, I know it's minor but I'm still seeing a difference.

Using 0.7 as the avgDif, 2.4 seems like the logical choice for maxDif, and a gradient of 1.6 seemed to be as low as I could go with those before debanding effectiveness was reduced.

I've had a play with the low setting using 6233638's parameters as a basis and I've found the following to provide a smoother deband transition. I think we're basically there as far as the low preset goes, please provide feedback.

avgDif maxDif grad penalty contrast
low: 0.7 2.4 2.5 1.5 1.5

nonGradientPenalty if it's costly could be removed, the differences it makes are terribly minor (at these settings) and are hard to identify without some detailed image comparison. I do prefer it enabled but I could live without it.

bacondither

15th October 2013, 10:36

madshi

15th October 2013, 10:41

LOW SETTINGS FINDING:

maxDif doesn't seem to harm the picture so I don't see why reducing it. Of course, maybe it harms the picture but not on the samples I have so feel free to test it on your samples and tell if it harms the pictures with the values I gave for example.
Are those images from the AVS HD 709 grayscale ramp?, and if so i too found maxDif 2.4 to remove most bands in that syntetic test image. But you almost have to be pixel peeping to see it.
It seems that the detail reduction by using maxDif 2.4 over 1.5-1.6 is so miniscule it might be a good vaule for a low setting.
nonGradientPenalty has seems to have very low impact on rendering times and does improve detail.

This is my reevaluated settings for low.
avgDif maxDif grad penalty localcontrast
low: 0.7 2.4 1.5 1.2 off

I really need to keep better track of my testing, because I'm not sure what video(s) it was that caused me to raise the low avgDif to 0.8 before
So unless I can find another example of that, 0.7 seems acceptable.
Using 0.7 as the avgDif, 2.4 seems like the logical choice for maxDif, and a gradient of 1.6 seemed to be as low as I could go with those before debanding effectiveness was reduced.
I've had a play with the low setting using 6233638's parameters as a basis and I've found the following to provide a smoother deband transition. I think we're basically there as far as the low preset goes, please provide feedback.

avgDif maxDif grad penalty contrast
low: 0.7 2.4 2.5 1.5 1.5

Ok, good, so it seems everybody agrees with "avgDif 0.7" and "maxDif 2.4" for the low setting? That's progress! So let's consider those two fixed. For the low settings we still need to decide on 2 things:

(1) How to set "gradient" and "nonGradientPenalty". Remember: A higher "nonGradientPenalty" requires a higher "gradient".

gradient / nonGradientPenalty:

Werewolfy: 1.5 / 1.2
bacondither: 1.5 / 1.2
6233638: 1.6 / 1.2
ryrynz: 2.5 / 1.5

ryrynz' suggestions are higher than the rest, but he has *both* values raised, so it makes sense, too. Can we somehow get an agreement on these values, too?

(2) We also still need to decide on "localContrast". See next section.

------------------------

LOCAL CONTRAST:

I still find the LocalContrast quite useless. There is almost no difference with or without and when there is a difference, it's very minimal.
I think local contrast could be made optional or removed. Makes almost no difference at removing banding at the cost of ~20% longer rendering times.
Unfortunately, I'm still finding sources where localContrast seems to be useful, and it doesn't seem like nonGradientPenalty will completely replace it.

I don't know that you would need separate presets which use localContrast, just have presets with localContrast values (and this would likely only be low/medium) and a "trade quality for performance" that disables the localContrast portion if required.
Agreed, even with localContrast fixed I'd still prefer it to stay, I know it's minor but I'm still seeing a difference.
I could live with a "trade quality for performance" setting for "localContrast". But could you guys, 6233638 and/or ryrynz, please provide a small test sample which shows the benefit of "localContrast"? Please let us know at which runtime in the sample a difference is visible. Thanks. Maybe that would convince Werewolfy, bacondither and me of its use. And that would also help finding optimal values for "localContrast". It seems that you guys favor a setting of 1.4 or 1.5 for "low"?

------------------------

RANDOM COMMENTS:

I must say, I'm finding it difficult to see anything but very subtle changes when adjusting nonGradientPenalty with the sources I've looked at, but 1.2 does seem to have some benefit at low settings.
I'm not sure that I would want to use high levels of gradient and nonGradientPenalty together, because high levels of gradient seem to add a lot of dither/noise to the image.
That sounds weird to me. The amount of dithering/noise should be totally independent of any of the debanding parameters. The dithering strength used is always the same. If large image areas are exactly between two 8bit values (e.g. 80.5), then the dithering noise might be more visible. But it's really quite random which debanding values put which specific areas to which 8bit values.

I thought also you could add a very low setting but now it seems useless to me. It's true that we can find a lower setting than the low setting I gave for example but in that case the debanding effect is quite minmal so yes maybe for native 4K content 10bit sources it can ben useful but for now, I don't really see the point.
Yeah, maybe we should forget about "very low" for now. If you guys are all happy with whatever settings we find for "low", and agree on that detail isn't harmed (in any way worth mentioning) by those settings, then "very low" can wait until we see a need for it.

I'm still quite happy with the max settings I posted before, and haven't found a source where Werewolfy's stronger settings are required.
Let's first decide on the final "low" settings, then we can concentrate more on "mid" and high".

nonGradientPenalty if it's costly could be removed, the differences it makes are terribly minor (at these settings) and are hard to identify without some detailed image comparison. I do prefer it enabled but I could live without it.
The cost is pretty small.

But in really dark areas i noticed something i dont like.

For me it looks like the "bright" areas stay bright and the dark areas become much brighter, so that the video seems partly brighter in some areas with active debanding.
I expected that the "bright" areas would be a little bit darker, so that the average luminosity stay the same (visual).

No debanding (part):
http://s1.directupload.net/images/user/131015/6p8eeupz.png

low-debanding (part):
http://s1.directupload.net/images/user/131015/74cb7nwh.png

I looked at the luminosity-histogram and the average luminosity is almost identical. But especially the lower part in the middle looks too bright for me with debanding.
Interesting. A key part of the debanding algorithm is to build an average of 4 surrounding pixels. This is one of those cases where linear light might help. I thought that the value difference between those pixels would be small enough to make linear light superfluous, but maybe I was wrong. So here's a new test build which has an additional "linear light" switch. Keyboard shortcuts Ctrl+Alt+Shift+F9/F10:

http://madshi.net/madVRdeband5.rar

On a quick check it doesn't seem to make any difference in most situations, but maybe it helps in your specific case to avoid the brightening effect you mentioned?

ryrynz

15th October 2013, 12:15

Double post.

ryrynz

15th October 2013, 12:16

I think local contrast could be made optional or removed. Makes almost no difference at removing banding at the cost of ~20% longer rendering times.
nonGradientPenalty has seems to have very low impact on rendering times and does improve detail.

This is my reevaluated settings for low.
avgDif maxDif grad penalty localcontrast
low: 0.7 2.4 1.5 1.2 off

Optional yes, removal I most certainly hope not. It makes more of a difference than changing many of the settings we have at our disposal.. I'll happily provide screenshots to show you, it's all in the blending transition. Believe me if I could generate the same quality results with it disabled I'd push for it's removal.

localcontrast makes more of a difference than nonGradientPenalty, again I can happily back this up.

madshi

15th October 2013, 12:24

@ryrynz, a short video sample would be great!

bacondither

15th October 2013, 12:53

ryrynz' suggestions are higher than the rest, but he has *both* values raised, so it makes sense, too. Can we somehow get an agreement on these values, too?
Well, we could take the median of our collective findings.
Variable N Median
gradient 4 1,550
nongradientpenalty 4 1,2000

ryrynz

15th October 2013, 13:23

It's going to be tough to push such an minor change in image quality with localContrast considering it's impact but the difference is in the softness of the image.
It find the differences easier to see when switching from no deband to LocalContrast and then to no LocalContrast, brightness and contrast has been increased to help highlight the differences.

The fix you implemented Madshi diminished the difference somewhat, I've used screenshots from the video I linked you earlier.

It's this slight softening of detail that would be removing the shadow detail that 6233638 talks about, it's really a case of how pedantic you/we want to be about this light detail.
Since I'm examining things at the pixel level.. it's hard for it not to do unnoticed and I prefer it being there.. by default? maybe not.

No deband
http://imageshack.us/a/img833/129/s54.th.png (http://imageshack.us/photo/my-images/833/s54.png/)

My settings with LocalContrast=1.5
http://imageshack.us/a/img43/2883/er95.th.png (http://imageshack.us/photo/my-images/43/er95.png/)

My settings with LocalContrast=disabled
http://imageshack.us/a/img11/2254/akpt.th.png (http://imageshack.us/photo/my-images/11/akpt.png/)

Well, we could take the median of our collective findings.

I'd rather compare and choose based on content, I'm striving for the absolute best deband quality for this because let's face it we don't have a lot of strength to work with and ideally I think we'd want low on by default.
I'm relying on LocalContrast here to maximize the amount of detail we get out of low whilst still maintaining a reasonable blending.

We've already tightened the two largest screws, so just about anything now is good enough really, but I've checked out Nongradientpenalty 1.2 and I'm happy with that so I'm in with everyone else.
I still think the gradient needs to be higher than 1.5-1.6, I've settled at 2.0 for this ATM. LocalContrast at 1.5 is almost identical to 1.4 but 1.5 from what I see has the every so slight edge, 6233638 could you also confirm this for LocalContrast?