Improving Madvr HDR to SDR mapping for projector - Page 15 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 5892Likes
Reply
 
Thread Tools
post #421 of 7035 Old 04-07-2018, 04:46 PM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,042
Mentioned: 491 Post(s)
Tagged: 0 Thread(s)
Quoted: 6927 Post(s)
Liked: 6637
Quote:
Originally Posted by mkohman View Post
Sounds amazing and already excited about it.. Thank you to you and all the others including the man himself @madshi for making such an amazing platform for us all to enjoy and experience.. Any idea when this will be complete for all to experience? Thank you..

Sent from my Moto G (4) using Tapatalk
I didnt do anything.

It will be ready when its ready.

Just keep an eye on the thread.
mkohman likes this.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
Sponsored Links
Advertisement
 
post #422 of 7035 Old 04-07-2018, 04:57 PM
Advanced Member
 
mkohman's Avatar
 
Join Date: Jun 2009
Location: London, United Kingdom
Posts: 793
Mentioned: 22 Post(s)
Tagged: 0 Thread(s)
Quoted: 417 Post(s)
Liked: 344
Quote:
Originally Posted by Javs View Post
I didnt do anything.

It will be ready when its ready.

Just keep an eye on the thread.
Will do.. Thanks

Sent from my Moto G (4) using Tapatalk
mkohman is offline  
post #423 of 7035 Old 04-08-2018, 12:00 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Quote:
Originally Posted by Javs View Post
Florian, do you have any real HDR capable displays in your house or is this all looking at an LS10k?

Just curious, because some things become a bit clearer when comparing to another HDR display as well as MadVR.
I only have the ls10000 at home with 100% dci coverage. I did compare planet earth 2 jungle episod on my parents LG 4k hdr oled tv. But the tv could not achieve 100% dci and it was clearly visible. Otherwise, the full green bird bird seem from above and the luminiscent mushroom looked more like dumb V2 than scientific.

You have to be careful with comparing with other tv because you are comparing tone mapping. And the tone mapping on your tv seems as wrong as on the jvc.

I will post some examples where there is no denying that dumb v2 is vastly inferior.

Problem is: when someone is used to a non calibrated picture and you suddenly calibrate it, sometimes they prefer the previous picture because they got use to the look of it. (Far too cold for example)

Last edited by Soulnight; 04-08-2018 at 12:06 AM.
Soulnight is online now  
Sponsored Links
Advertisement
 
post #424 of 7035 Old 04-08-2018, 12:17 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,042
Mentioned: 491 Post(s)
Tagged: 0 Thread(s)
Quoted: 6927 Post(s)
Liked: 6637
Quote:
Originally Posted by Soulnight View Post
You have to be careful with comparing with other tv because you are comparing tone mapping. And the tone mapping on your tv seems as wrong as on the jvc.

I will post some examples where there is no denying that dumb v2 is vastly inferior.
I have three HDR displays in my house. A JVC, a Samsung and a Hisense. So you can get a pretty good indication from things when looking at a broad range like that all different manufacturers.

Keen to see where you think v2 is vastly inferior. I prefer it in most cases personally even the lave shot you posted yourself.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #425 of 7035 Old 04-08-2018, 01:33 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Quote:
Originally Posted by Soulnight View Post
What I mean is: desaturate LESS if it should be very saturated in the original.

Of course you need to desaturate, the question is how much. And the how much should be a function of original brightness and saturation. That would be a improvement of your current dynamic logic.

- if original is very saturated, then allow only a slight desaturation to give best compromise between highlights details and color saturation fidelity.
- if original is not that saturated, then you can desaturate even more and regain brightness and details.
I'm already taking both brightness and original saturation into account. That's the very key of the algorithm. E.g. 50%/50%, splits the modifications to brightness and saturation exactly in half, so that brightness is changed in a similar amount to saturation. However, as you can see in the spear screenshot, 50%/50% simply doesn't work very well in many scenes with extremely bright pixels. Because of that "dynamic" adjusts the ratio between luminance and saturation adjustment dynamically.

What you're *really* asking for is some sort of intelligent algorithm which takes the whole image context into account to judge how much desaturation is needed to not lose highlight detail. But that's not how the algorithm works. The algorithm looks only at one pixel at a time, ignoring what the neighbor pixels do. As such, I don't know if I lose image detail if I don't desaturate, or not. Looking at the image context would be extremely difficult, because we wouldn't even have to look just at very near neighbor pixels, but we'd have to look a lot further than that. It's not something I can do at acceptable performance levels.

Please understand that to a human being, it's immediately obvious whether applying a specific tone mapping algo produces highlight loss in a specific scene or not. So it's easy for you to say "in the spear scene the algo should desaturate more than in the swordsmith scene". But that's simply not the way tone mapping algos work. I can tell you right now that whatever approach you like for scene A will not produce perfect results for scene B, unless you design the algorithm to try to intelligently judge the whole scene. So due to the limitions of the algorithm, I have to design the algo in such a way that it does not lose highlight detail in scenes like the spear scene, but that means the algo might desaturate more than a human being would in the swordsmith scene.

In the swordsmith scene, the whole glowing blob has a somewhat similar brightness and saturation with not much texture to it (except for small much darker areas). If that thing had much more overbright HDR highlight variation, like the green spear has, it would look very bad with 50%/50%. The lack of highlight texture is the only reason why the swordsmith scene looks acceptable in 50%/50% at all.

You're making it easy for yourself, by judging one scene at a time. But if you tune the algo for that scene, it will look worse elsewhere.

Quote:
Originally Posted by Javs View Post
At least to me, the JVC style looks easily the best, and interestingly it also looks the most 'hot' which is to be expected...
But you do understand that the scene it not supposed to be yellow? It doesn't bother you at all that the scene is rendered in a way that doesn't match the colorist's intent? We've recently learned that studio mastering monitors usually clip. Which means they don't use a hue shifting tone mapping algo like the JVC or Samsung TVs/projectors seem to use. Clipping will introduce some hue shifts, too, but only if the content goes above the clipping point, and it's *much* less strong than the hue shift produced by naive tone mapping, especially because the studio monitors have relatively high peak luminance capability, compared to what we're testing with (projectors).
stevenjw likes this.
madshi is online now  
post #426 of 7035 Old 04-08-2018, 02:16 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Some examples where dynamic is clearly better than JVC's dumb approach:

The whole demo The World in 4K HDR looks better with dynamic:

dynamic:
(red is more intense and sky is bluer)


dumb:




dynamic:
(snake is greener and flower are way more red)


dumb:




dynamic:
(here the orange, red colors are way more intense and pop. Also sky is more blue)


dumb:




dynamic:
(the landscape is more orange saturated)


dumb:




dynamic:
(the tongue is more pink)


dumb:




The same is true for "The secret life of pets". It looks better with the scientific dynamic tone mapping.


dynamic:
(here all the colors are way more intense with scientific than dumbV2. The HDR pop also)


dumb:




dynamic:
(here all the colors are way more intense with scientific than dumbV2. The HDR pop also)


dumb:




dynamic:
(The sun is there with scientific and is gone with dumb-V2)


dumb:




In "Planet earth II" episode 3 almost everything looks better with the scientific dynamic tone mapping, except for the 2 examples I already mentioned (deeper green with dumb tone mapping), which is why I attached them here:


dynamic:
(The mushroom has less intensive green)


dumb:




dynamic:
(Bird looks less intensive green)


dumb:




Another screenshot from this episode, where you could argue about which one is better. I still prefer the scientific tone mapping, because of the detailed golden feaethers.


dynamic:



dumb:



Also I noticed that in all these examples dynamic, 50-50 and even 0-100 all looked identical or had very few differences. For me it closes this topic, because such examples like the spear or the swordsmith are very rare and dynamic does a good job on them.
Let's move on.

EDIT:
best is to download those picture and compare them back and forth in a picture viewer. The differences are very clear doing so. One above the others, it is way more difficult to spot the differences.
Attached Thumbnails
Click image for larger version

Name:	Planet-Earth-II-E03_00.38.36_dynamic.jpg
Views:	355
Size:	3.11 MB
ID:	2386806   Click image for larger version

Name:	Planet-Earth-II-E03_00.38.36_dumbV2.jpg
Views:	356
Size:	3.19 MB
ID:	2386808   Click image for larger version

Name:	Planet-Earth-II-E03_00.47.38_dynamic.jpg
Views:	350
Size:	3.27 MB
ID:	2386810   Click image for larger version

Name:	Planet-Earth-II-E03_00.47.38_dumbV2.jpg
Views:	359
Size:	3.26 MB
ID:	2386812   Click image for larger version

Name:	Planet-Earth-II-E03_00.44.01_dynamic.jpg
Views:	361
Size:	2.82 MB
ID:	2386814  

Click image for larger version

Name:	Planet-Earth-II-E03_00.44.01_dumbV2.jpg
Views:	341
Size:	2.92 MB
ID:	2386816   Click image for larger version

Name:	The World In HDR 4K Demo_00.39_dumbV2.jpg
Views:	375
Size:	2.59 MB
ID:	2386818   Click image for larger version

Name:	The World In HDR 4K Demo_00.07_dynamic.jpg
Views:	382
Size:	2.82 MB
ID:	2386820   Click image for larger version

Name:	The World In HDR 4K Demo_00.07_dumbV2.jpg
Views:	372
Size:	2.79 MB
ID:	2386822   Click image for larger version

Name:	The World In HDR 4K Demo_00.39_dynamic.jpg
Views:	374
Size:	2.69 MB
ID:	2386824  

Click image for larger version

Name:	The World In HDR 4K Demo_00.45_dynamic.jpg
Views:	379
Size:	5.16 MB
ID:	2386826   Click image for larger version

Name:	The World In HDR 4K Demo_00.45_dumbV2.jpg
Views:	363
Size:	4.83 MB
ID:	2386828   Click image for larger version

Name:	The World In HDR 4K Demo_02.06_dynamic.jpg
Views:	371
Size:	3.57 MB
ID:	2386830   Click image for larger version

Name:	The World In HDR 4K Demo_02.06_dumbV2.jpg
Views:	370
Size:	3.46 MB
ID:	2386832   Click image for larger version

Name:	The World In HDR 4K Demo_01.37_dynamic.jpg
Views:	363
Size:	3.81 MB
ID:	2386834  

Click image for larger version

Name:	The World In HDR 4K Demo_01.37_dumbV2.jpg
Views:	367
Size:	3.74 MB
ID:	2386836   Click image for larger version

Name:	The-secret-life-of-pets_00.01.58_dynamic.jpg
Views:	362
Size:	3.24 MB
ID:	2386838   Click image for larger version

Name:	The-secret-life-of-pets_00.01.58_dumbV2.jpg
Views:	371
Size:	3.20 MB
ID:	2386840   Click image for larger version

Name:	The-secret-life-of-pets_00.01.49_dynamic.jpg
Views:	367
Size:	3.06 MB
ID:	2386842   Click image for larger version

Name:	The-secret-life-of-pets_00.01.49_dumbV2.jpg
Views:	363
Size:	3.04 MB
ID:	2386844  

Click image for larger version

Name:	The-secret-life-of-pets_00.02.19_dynamic.jpg
Views:	348
Size:	3.35 MB
ID:	2386846   Click image for larger version

Name:	The-secret-life-of-pets_00.02.19_dumbV2.jpg
Views:	351
Size:	3.33 MB
ID:	2386848  
stevenjw likes this.

Last edited by Soulnight; 04-08-2018 at 02:27 AM.
Soulnight is online now  
post #427 of 7035 Old 04-08-2018, 02:21 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,042
Mentioned: 491 Post(s)
Tagged: 0 Thread(s)
Quoted: 6927 Post(s)
Liked: 6637
Quote:
Originally Posted by madshi View Post
But you do understand that the scene it not supposed to be yellow? It doesn't bother you at all that the scene is rendered in a way that doesn't match the colorist's intent? We've recently learned that studio mastering monitors usually clip. Which means they don't use a hue shifting tone mapping algo like the JVC or Samsung TVs/projectors seem to use. Clipping will introduce some hue shifts, too, but only if the content goes above the clipping point, and it's *much* less strong than the hue shift produced by naive tone mapping, especially because the studio monitors have relatively high peak luminance capability, compared to what we're testing with (projectors).
Yes I would definitely prefer to see as close as possible to what the colonist saw, absolutely.

I am not sure dynamic is the answer though would you agree? Are we not trading one evil for another in a sense?

I think I would rather side with what most of the world seems to be observing (hue shift tone mapping) than add white to the colours in question which while may not be a hue shift, can become a drastic saturation shift, its pick your poison it seems. I find white highlights in the content we have been testing more attention grabbing that 'something is off' rather than a bit of a hue shift I have to say.

If I cant see exactly what the colorist saw, then I suppose I would side with the next best IMO and that is 'generally' what we are seeing in the SDR Bluray, of Superman and Mad Max, of which both seem to very closely resembling the dumber hue shift tone mapping.






JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #428 of 7035 Old 04-08-2018, 02:29 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Quote:
Originally Posted by Soulnight View Post
Some examples where dynamic is clearly better than JVC's dumb approach
Some great screenshots there, thanks! I agree that in all those images, I prefer "dynamic" over "dumb". The colors pop more, skies are more blue instead of mushy etc.

FWIW, I've found it much easier to see the difference by copying the image URLs to separate tabs and switching between them. I don't suppose it's possible to replace the full sized images from your post with URLs, to make the tab comparison faster/easier, is it? I don't know exactly how the AVSForum image attachment logic works.

P.S: Which settings did you use for the 3 different color space options? xyY? ICtCp 0%, 50%, 100%? Did you check if changing those options makes any difference?
madshi is online now  
post #429 of 7035 Old 04-08-2018, 02:38 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Quote:
Originally Posted by madshi View Post
Some great screenshots there, thanks! I agree that in all those images, I prefer "dynamic" over "dumb". The colors pop more, skies are more blue instead of mushy etc.

FWIW, I've found it much easier to see the difference by copying the image URLs to separate tabs and switching between them. I don't suppose it's possible to replace the full sized images from your post with URLs, to make the tab comparison faster/easier, is it? I don't know exactly how the AVSForum image attachment logic works.

P.S: Which settings did you use for the 3 different color space options? xyY? ICtCp 0%, 50%, 100%? Did you check if changing those options makes any difference?
Yes, those examples are really impressive if doing back and forth.
Just download the attachements below, and in the picture viewer you can do very easily back and forth because we used the same name for a pair of pictures.
You can also directly click on the attachements and it will open them in a new tab so you can compare 2 tabs back abd forth like you wanted ;-)

Those were wit ICtCp 100% and yes I did check some days ago what happends with 0% instead and it looked identical to me.

Looking forward for the next step.

Last edited by Soulnight; 04-08-2018 at 02:43 AM.
Soulnight is online now  
post #430 of 7035 Old 04-08-2018, 02:42 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,042
Mentioned: 491 Post(s)
Tagged: 0 Thread(s)
Quoted: 6927 Post(s)
Liked: 6637
Quote:
Originally Posted by Soulnight View Post
Some examples where dynamic is clearly better than JVC's dumb approach:

The whole demo The World in 4K HDR looks better with dynamic:

Also I noticed that in all these examples dynamic, 50-50 and even 0-100 all looked identical or had very few differences. For me it closes this topic, because such examples like the spear or the swordsmith are very rare and dynamic does a good job on them.
Nice shots,

What settings are you using for the dumber hue shift tone mapping? Did you create an Arve curve?

Hmm, so I just downloaded that clip. It did look really good on the dynamic option, and I could observe the same colour differences as you. I do agree they appeared to be more appealing,

But this just makes me confused now.

Why is dynamic mode affecting the entire colour spectrum including near black content when the option implies its there to deal with highlights and 'too bright pixels?'

I cant see how/why this is affecting literally the entire overall colour of the images, it should not be such a difference. It looks like its jumping between Rec709 and BT2020 in some of those shots when I cycle between the dynamic mode and the 'none' mode, and this is not even on particularly bright shots at all, perhaps a couple hundred nits.

Interesting indeed.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #431 of 7035 Old 04-08-2018, 02:44 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,042
Mentioned: 491 Post(s)
Tagged: 0 Thread(s)
Quoted: 6927 Post(s)
Liked: 6637
Quote:
Originally Posted by madshi View Post
Some great screenshots there, thanks! I agree that in all those images, I prefer "dynamic" over "dumb". The colors pop more, skies are more blue instead of mushy etc.

FWIW, I've found it much easier to see the difference by copying the image URLs to separate tabs and switching between them. I don't suppose it's possible to replace the full sized images from your post with URLs, to make the tab comparison faster/easier, is it? I don't know exactly how the AVSForum image attachment logic works.

P.S: Which settings did you use for the 3 different color space options? xyY? ICtCp 0%, 50%, 100%? Did you check if changing those options makes any difference?
Madshi I would really prefer images be left in line in the posts rather than URL's to click.

I have a big enough screen its easy to fit multiple images on the page when they are posted in line. You can just drag the window to the left to make it 'thinner' and it will dynamically scale the images, that way if you really wanted you could look at 4 or 5 at the same time.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #432 of 7035 Old 04-08-2018, 02:49 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Smile

Quote:
Originally Posted by Javs View Post
Nice shots,

What settings are you using for the dumber hue shift tone mapping? Did you create an Arve curve?

Hmm, so I just downloaded that clip. It did look really good on the dynamic option, and I could observe the same colour differences as you. I do agree they appeared to be more appealing,

But this just makes me confused now.

Why is dynamic mode affecting the entire colour spectrum including near black content when the option implies its there to deal with highlights and 'too bright pixels?'

I cant see how/why this is affecting literally the entire overall colour of the images, it should not be such a difference. It looks like its jumping between Rec709 and BT2020 in some of those shots when I cycle between the dynamic mode and the 'none' mode, and this is not even on particularly bright shots at all, perhaps a couple hundred nits.

Interesting indeed.
The difference does not come from dynamic, but from the simplified "Dumb V2" ton mapping with hue shift and so for anything ABOVE your chosen "display target nits". In my case=300nits with a white reference of 100nits.

Madshi can probably better explain why the differences affects for than the highlights.

I am using the same BT2390 hdr curve for all those shots, no Arve curve.
Soulnight is online now  
post #433 of 7035 Old 04-08-2018, 03:09 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,042
Mentioned: 491 Post(s)
Tagged: 0 Thread(s)
Quoted: 6927 Post(s)
Liked: 6637
Quote:
Originally Posted by Soulnight View Post
The difference does not come from dynamic, but from the simplified "Dumb V2" ton mapping with hue shift and so for anything ABOVE your chosen "display target nits". In my case=300nits with a white reference of 100nits.

Madshi can probably better explain why the differences affects for than the highlights.

I am using the same BT2390 hdr curve for all those shots, no Arve curve.
Hmm, that's not quite right, since I can see it affecting colours and shadow detail way way below anything like that. I am talking 20 nits.

Can we not combine something like what this is doing to content clearly under 100 nits anyway, with the dumber option for a look? I want what dynamic is doing for content up to 2000 nits, then I want what the dumber curves are doing for content above that

Check out this waveform,



Dynamic is boosting saturation on content as low as 5-10 nits it seems.

I lifted the next couple screenshots in photoshop to show what its doing otherwise its hard to see on a desktop unless looking in a light controlled enviro.

'Dynamic'



'None'


JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves

Last edited by Javs; 04-08-2018 at 03:16 AM.
Javs is offline  
post #434 of 7035 Old 04-08-2018, 03:18 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Quote:
Originally Posted by Javs View Post
Why is dynamic mode affecting the entire colour spectrum including near black content when the option implies its there to deal with highlights and 'too bright pixels?'
It's the other way round:

It's not scientific tone mapping (or dynamic mode) which changes the entire colour spectrum. It's dumb mode which does that. Scientific tone mapping, especially when combined with arve custom curves, is designed to not touch pixels below diffuse white *at all* (except a carefully applied brightness boost to make the pixels measure with the correct brightness on your dark projector). So the color differences you're seeing are not produced by scientific tone mapping (or dynamic mode) "improving" the image. Instead what you're seeing is dumb mode harming the color balance, and not only in HDR highlight areas, but over the whole colour and brightness spectrum.

Dumb mode looks at each color component (red, green, blue) separately, totally ignoring what the other color components do. Which means if you have e.g. a dark pixel with no green, a little bit of red and a lot of blue, dumb mode will compress blue, but leave red and green untouched. Which means the hue of this pixel will change, even though it's overall a rather dark pixel. Added to that, the brightness boost needed to make the video measure with the correct brightness on your dark projector is done in RGB, and RGB is neither hue linear nor perceptionally uniform, so that further harms the color balance.

All that said, I'm not sure right now if the order in which I do tone mapping and gamut conversion in dumb mode is optimal. I'll try if swapping the order improves things. But that will only have an effect if you set "this display is already calibrated" to something different than BT.2020, and I'm not sure if it will make a visible difference at all.

Quote:
Originally Posted by Javs View Post
Yes I would definitely prefer to see as close as possible to what the colonist saw, absolutely.

I think I would rather side with what most of the world seems to be observing [...].
According to the information provided by Kris Deering, studio mastering monitors usually clip. Which means that "what most of the world seems to be observing" is *NOT* what the colorist saw on his mastering monitor at all. Clipping introduces some hue shifts, as well, but only if the content goes above the clipping point. I've tried simulating the hue shifts produced by a mastering monitor clipping at 4000nits with Mad Max and Batman vs Superman, and in my tests these hue shifts were almost invisible. I'm fairly certain that when working on a Pulsar display, the colorist saw red explosions (not orange/yellow) in Mad Max, and he saw a purely green (not yellow) spear in Batman vs Superman.

So you can't have both: Either you try to reproduce what the colorist saw (red explosions, green spear), or you try to simulate what most of the world seems to be observing (orange/yellow explosions, yellow-ish spear).

It definitely adds some salt to the discussion that the spear is yellow-ish on the SD Blu-Ray. But IMHO that just shows that the studio mastering software uses subpar algorithms, nothing else. The key thing is what the colorist saw on his Pulsar display when mastering the UHD HDR Blu-Rays (isn't it?), and I'm pretty sure that was not a yellow spear in Batman vs Superman and not orange/yellow explosions in Mad Max.

Quote:
Originally Posted by Javs View Post
I think I would rather side with what most of the world seems to be observing (hue shift tone mapping) than add white to the colours in question which while may not be a hue shift, can become a drastic saturation shift, its pick your poison it seems. I find white highlights in the content we have been testing more attention grabbing that 'something is off' rather than a bit of a hue shift I have to say.
Some pages ago I quoted a lot of different scientific papers. They *ALL* said that maintaining hue is critically important. I've not found even one single scientific paper which said otherwise. And it's not "a bit of a hue shift", it's pretty dramatic:



All that said, I will keep the "dumb" tone mapping algorithm. I might even make it the default one, because it's the fastest. But I certainly recommend to use the scientific approach, if your GPU can handle the additional load.
stevenjw and Javs like this.
madshi is online now  
post #435 of 7035 Old 04-08-2018, 04:48 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Quote:
Originally Posted by madshi View Post
It's the other way round:

It's not scientific tone mapping (or dynamic mode) which changes the entire colour spectrum. It's dumb mode which does that.

All that said, I will keep the "dumb" tone mapping algorithm. I might even make it the default one, because it's the fastest. But I certainly recommend to use the scientific approach, if your GPU can handle the additional load.
Very nice explanation. Makes sense why dumb v2 corrupts colors regardless of their brightness.


Scientific should stay default!
If the user does not have the "horse power ", then he can use the wrong one.

I have a german HTPC&MADVR facebook group with many happy user currently using scientific hdr to sdr dynamic ton mapping with their projector. (Many jvc btw., but also some Sony 4k projector, epson etc...).

It may be even the reason why jvc and some tvs are doing it wrong: save power. Or maybe it is just that they don't fully understand hdr ton mapping.
mkohman likes this.
Soulnight is online now  
post #436 of 7035 Old 04-08-2018, 05:00 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Quote:
Originally Posted by Soulnight View Post
Scientific should stay default!
If the user does not have the "horse power ", then he can use the wrong one.
Ok, I'll use the slightly lower quality xyY scientific variant as default then, though, because it's about twice as fast as the ICtCp version (and about twice as slow as the dumb mode).

Quote:
Originally Posted by Soulnight View Post
I have a german HTPC&MADVR facebook group with many happy user currently using scientific hdr to sdr dynamic ton mapping with their projector. (Many jvc btw., but also some Sony 4k projector, epson etc...).
Sounds great!

One important remaining question is if any of the manual x%/x% options is overall better (for all variety of scenes) than the new "dynamic" mode? IMHO dynamic is better overall, so I'd like to get rid of all the manual options. Of course that doesn't mean that "dynamic" can't be further improved. Although right now I don't know how to.
madshi is online now  
post #437 of 7035 Old 04-08-2018, 05:31 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Quote:
Originally Posted by madshi View Post
Ok, I'll use the slightly lower quality xyY scientific variant as default then, though, because it's about twice as fast as the ICtCp version (and about twice as slow as the dumb mode).
Could you explain the differences behind itcp 0,50,100 or xyY?
:-)
What did use per default before: itcp 100?



Quote:
Originally Posted by madshi View Post
One important remaining question is if any of the manual x%/x% options is overall better (for all variety of scenes) than the new "dynamic" mode? IMHO dynamic is better overall, so I'd like to get rid of all the manual options. Of course that doesn't mean that "dynamic" can't be further improved. Although right now I don't know how to.
Dynamic is more than fine. Just drop the rest. I have difficulties anyway to find examples where it even matters. ;-)
Soulnight is online now  
post #438 of 7035 Old 04-08-2018, 05:40 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Quote:
Originally Posted by Soulnight View Post
Could you explain the differences behind itcp 0,50,100 or xyY?
:-)
What did use per default before: itcp 100?
xyY and ICtCp are different color spaces. xyY is mathematically easier to deal with, but it's not hue linear, nor perceptionally uniform. 0,50,100 are different ways to change brightness in ICtCp. Dolby suggests to use 100, which I had used before, but it does reduce saturation ever so slightly, when reducing brightness. So I added 0 and 50 as a test, because you guys complained about too high desaturation. But really, the difference between 0,50,100 is pretty small. So I think I'm going back to Dolby's recommendation of 100.

Quote:
Originally Posted by Soulnight View Post
Dynamic is more than fine. Just drop the rest. I have difficulties anyway to find examples where it even matters. ;-)
Ok, I'll probably keep the manual options for one more official version, to give users a last chance to test and complain, before I remove it. After all, dynamic isn't even available yet in any official version.
madshi is online now  
post #439 of 7035 Old 04-08-2018, 05:45 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Quote:
Originally Posted by madshi View Post
xyY and ICtCp are different color spaces. xyY is mathematically easier to deal with, but it's not hue linear, nor perceptionally uniform. 0,50,100 are different ways to change brightness in ICtCp. Dolby suggests to use 100, which I had used before, but it does reduce saturation ever so slightly, when reducing brightness. So I added 0 and 50 as a test, because you guys complained about too high desaturation. But really, the difference between 0,50,100 is pretty small. So I think I'm going back to Dolby's recommendation of 100.
Ok. I will check again to be sure I don't see any benefit from itcp 0% or any downside from xyY.

What's your next step? :-)
Soulnight is online now  
post #440 of 7035 Old 04-08-2018, 06:03 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
There are many steps to go. The very next step is to wait for Manni to double check "dumb" tone mapping to make sure it serves as a full replacement for JVC HDR passthrough with arve custom curves. Once that's confirmed, Manni and Javs can stop comparing with JVC passthrough. Instead they can simply switch between dumb vs scientific mode in madVR which will make testing for them much easier and faster. Once we're there, we can take further steps, with faster and easier testing.

xyY does have a downside. It's faster, but introduces small hue shifts. Using it only makes sense to save performance. So I'll definitely offer both xyY and ICtCp as an option. It's what was formerly available as "low" vs "high" quality hue preserving.
madshi is online now  
post #441 of 7035 Old 04-08-2018, 06:18 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,042
Mentioned: 491 Post(s)
Tagged: 0 Thread(s)
Quoted: 6927 Post(s)
Liked: 6637
Quote:
Originally Posted by madshi View Post
There are many steps to go. The very next step is to wait for Manni to double check "dumb" tone mapping to make sure it serves as a full replacement for JVC HDR passthrough with arve custom curves. Once that's confirmed, Manni and Javs can stop comparing with JVC passthrough. Instead they can simply switch between dumb vs scientific mode in madVR which will make testing for them much easier and faster. Once we're there, we can take further steps, with faster and easier testing.

xyY does have a downside. It's faster, but introduces small hue shifts. Using it only makes sense to save performance. So I'll definitely offer both xyY and ICtCp as an option. It's what was formerly available as "low" vs "high" quality hue preserving.
I definitely found ictcp to render fire better in mad Max, at least to my liking, xyy went too white.

I also still saw pretty hard edged and stepped posterization inside highlights in mad Max which completely went away using the none option.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #442 of 7035 Old 04-08-2018, 06:28 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
xyY should not desaturate more than ICtCp, but xyY makes explosions in Mad Max look a bit more purple than red. ICtCp is clearly better than xyY, but also slower. As for "hard edged and stepped posterization", that's something we should look at, but not right now. I'd really like Manni to first (hopefully) confirm that dumb tone mapping can fully replace JVC passthrough. Because we'll be able to work together best, if we're all on the same page and if we're all working on the same thing.
madshi is online now  
post #443 of 7035 Old 04-08-2018, 07:25 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,172
Mentioned: 334 Post(s)
Tagged: 0 Thread(s)
Quoted: 5490 Post(s)
Liked: 5708
I’m back briefly today, leaving again tomorrow. I’ll try to steal a few hours if I can today, otherwise it will have to wait until I’m back on thursday. I’ll do my best.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #444 of 7035 Old 04-08-2018, 07:56 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Please don't feel time pressured too much. I won't have too much time to work on this in the next week, anyway. Of course I appreciate if you can squeeze the tests in, but it won't harm much if you wait until you're "fully" back.
madshi is online now  
post #445 of 7035 Old 04-08-2018, 08:52 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,172
Mentioned: 334 Post(s)
Tagged: 0 Thread(s)
Quoted: 5490 Post(s)
Liked: 5708
Quote:
Originally Posted by madshi View Post
Please don't feel time pressured too much. I won't have too much time to work on this in the next week, anyway. Of course I appreciate if you can squeeze the tests in, but it won't harm much if you wait until you're "fully" back.
Just checking levels and arve curve parameters and I'm finding a few weird things.

1) Levels seem to be fixed to "superwhite" in the custom arve tool. Is that by design? I don't use superwhite, I usually use standard or enhanced.
2) I'm getting weird levels outputs using the latest nVidia driver. I always select 0-255 (enhanced) in the GPU, but I have to select 0-255 in MadVR and Standard in the JVC to get decent levels. That's weird. I would expect 16-235 in MadVR and standard in the JVC or 0-255 in MadVR and Enhanced in the JVC to produce the correct levels (at least for video).

I might revert to an earlier version of the drivers just to rule that out.

If others in this thread are using nVidia, I'd be curious to know their settings in nVidia CP )driver version), JVC and MadVR to make sure I'm comparing apples to apples.

This is with Windows 10 latest 1709 Pro x64.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #446 of 7035 Old 04-08-2018, 08:55 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Quote:
Originally Posted by Manni01 View Post
1) Levels seem to be fixed to "superwhite" in the custom arve tool. Is that by design? I don't use superwhite, I usually use standard or enhanced.
I've hard coded the tool to always do video levels. Because all 3 options do the same thing, the tool detects that as "superwhite", for some reason. But don't worry about it, it should still work correctly.

Quote:
Originally Posted by Manni01 View Post
2) I'm getting weird levels outputs using the latest nVidia driver. I always select 0-255 (enhanced) in the GPU, but I have to select 0-255 in MadVR and Standard in the JVC to get decent levels. That's weird. I would expect 16-235 in MadVR and standard in the JVC or 0-255 in MadVR and Enhanced in the JVC to produce the correct levels (at least for video).
Just to be safe: This also occurs with BT.2390 tone mapping, right?

If all else fails, try using the madLevelsTweaker tool. It usually does the trick, if the GPU driver refuses to listen to your preferences. I've set the GPU driver to RGB Full Range 12bit.
madshi is online now  
post #447 of 7035 Old 04-08-2018, 11:40 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,172
Mentioned: 334 Post(s)
Tagged: 0 Thread(s)
Quoted: 5490 Post(s)
Liked: 5708
Quote:
Originally Posted by madshi View Post
I've hard coded the tool to always do video levels. Because all 3 options do the same thing, the tool detects that as "superwhite", for some reason. But don't worry about it, it should still work correctly.
Thanks, I think I remember you saying this when you first released the tool, sorry I forgot.

Quote:
Originally Posted by madshi View Post
Just to be safe: This also occurs with BT.2390 tone mapping, right?

If all else fails, try using the madLevelsTweaker tool. It usually does the trick, if the GPU driver refuses to listen to your preferences. I've set the GPU driver to RGB Full Range 12bit.
This is even in SDR. I always set levels in SDR first, then check with HDR. So not MadVR related at all.

Anyway, I reverted to 385.28 which is my last driver where everything works. The levels are behaving as expected. I don't know when exactly nVidia broke the levels, but I'll try to find out when I'm back.

With this levels issue fixed (nothing to do with MadVR) and after making sure that the parameters in Arve's curve (passthrough) and custom tool (pixel shader) are exactly the same, I can confirm after a quick check of a few clips that pixel shader with arve custom curve and passthrough seem virtually identical, so don't wait for me to move to the next step. I'm happy to leave passthrough behind at this stage and only compare naive and scientifically correct going forward.

I decided to use GPU 0-255, MadVR 16-235 and PJ standard as this gave the best/most consistent results in both SDR and HDR.

I need to recalibrate the PJ to get my color clipping 100% right but I don't have the time to do it now.

I only have a bump at 1000nits in scientifically correct with BT2390 or with my Arve curve. In that case, I resolve up to 10,000nits.

If I select my arve curve clipping at 4000nits and the naive mode, I clip white around 4000nits but I clip colors lower (same as passthrough though). The bump around 1000nits goes away.

So it looks like I can get no clipping with desaturation (scientifically correct) and a bump around 1000nits, or more clipping and less desaturation (naive mode), without a bump around 1000nits.

As far as I can see, scientifically correct doesn't take the hard clip point in the custom curve into account, which might be some of the problem.

Question: would it be possible to get keyboard shortcuts (if they don't exist already) to switch manually between passthrough / pixel shader on the fly? Any chance to have an indication of the current mode on the OSD? This would make testing/reporting easier and faster.

I don't have the time to take pictures but I will when I'm back next week if that helps.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders

Last edited by Manni01; 04-08-2018 at 11:43 AM.
Manni01 is online now  
post #448 of 7035 Old 04-08-2018, 11:50 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,418
Mentioned: 217 Post(s)
Tagged: 0 Thread(s)
Quoted: 1052 Post(s)
Liked: 1653
Quote:
Originally Posted by Manni01 View Post
I can confirm after a quick check of a few clips that pixel shader with arve custom curve and passthrough seem virtually identical, so don't wait for me to move to the next step. I'm happy to leave passthrough behind at this stage and only compare naive and scientifically correct going forward.

Question: would it be possible to get keyboard shortcuts (if they don't exist already) to switch manually between passthrough / pixel shader on the fly? Any chance to have an indication of the current mode on the OSD? This would make testing/reporting easier and faster.
Very good news! :-)

For the shortcuts, just create some new sub profiles in madvr associated with the Shortcuts you want and you will be to compare anything on the fly in madvr, including all hdr setup possibilities. :-)
Manni01 likes this.
Soulnight is online now  
post #449 of 7035 Old 04-08-2018, 12:21 PM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,775
Mentioned: 535 Post(s)
Tagged: 0 Thread(s)
Quoted: 2641 Post(s)
Liked: 3593
Quote:
Originally Posted by Manni01 View Post
With this levels issue fixed (nothing to do with MadVR) and after making sure that the parameters in Arve's curve (passthrough) and custom tool (pixel shader) are exactly the same, I can confirm after a quick check of a few clips that pixel shader with arve custom curve and passthrough seem virtually identical, so don't wait for me to move to the next step. I'm happy to leave passthrough behind at this stage and only compare naive and scientifically correct going forward.
Awesome!! Thanks for the quick test.

Quote:
Originally Posted by Manni01 View Post
I decided to use GPU 0-255, MadVR 16-235 and PJ standard as this gave the best/most consistent results in both SDR and HDR.
Yes, that's one of the 2 recommended configurations. The other one would be to set all three to 0-255. Both are fine. Your current config has the advantage of passing BTB & WTW, but the disadvantage that the Windows desktop, GUI and games will have incorrect levels.

Quote:
Originally Posted by Manni01 View Post
So it looks like I can get no clipping with desaturation (scientifically correct) and a bump around 1000nits, or more clipping and less desaturation (naive mode), without a bump around 1000nits.

As far as I can see, scientifically correct doesn't take the hard clip point in the custom curve into account, which might be some of the problem.
Ok, that's one thing we should check out in more detail when you're back.

Quote:
Originally Posted by Manni01 View Post
Question: would it be possible to get keyboard shortcuts (if they don't exist already) to switch manually between passthrough / pixel shader on the fly? Any chance to have an indication of the current mode on the OSD? This would make testing/reporting easier and faster.
For keyboard shortcuts see Soulnight's reply. I guess I could make a test build which shows the gamut & tone mapping mode in the OSD, that's really easy to do.

-------

If I may suggest, the next step for all of us would be to once more double check "dynamic" vs the manual x%/x% settings (in scientific mode) to confirm that if we are forced to pick one setting, "dynamic" is really the least bad one. If we can confirm that, we would only be left with 2 settings to compare: naive vs scientific, with no further sub-options. That would make testing so much easier and faster!!!

For this dynamic vs manual comparison, please make sure that you test with a variety of scenes. E.g. green spear in Batman vs Superman, explosions in Mad Max, the orange car lamp in the Sony Campfire demo, etc. The one setting you choose as your most favorite setting must have the least amount of catastrophic failures in all those scenes. Please note: You are not allowed to pick different manual settings for different scenes (e.g. 50%/50% for explosions, and 15%/85% for the green spear). You need to nominate one single setting that you prefer, considering all the various scenes... I hope that we can all agree on "dynamic" (but I'm willing to listen, if at least 2 of you guys disagree). Once we agree on one setting, we can try to further improve it. Not sure how much improvement is possible, but we can try...
Manni01 likes this.
madshi is online now  
post #450 of 7035 Old 04-08-2018, 01:03 PM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,172
Mentioned: 334 Post(s)
Tagged: 0 Thread(s)
Quoted: 5490 Post(s)
Liked: 5708
Quote:
Originally Posted by madshi View Post

For keyboard shortcuts see Soulnight's reply. I guess I could make a test build which shows the gamut & tone mapping mode in the OSD, that's really easy to do.
If it's easy, that would be great and very useful.

Quote:
Originally Posted by madshi View Post

If I may suggest, the next step for all of us would be to once more double check "dynamic" vs the manual x%/x% settings (in scientific mode) to confirm that if we are forced to pick one setting, "dynamic" is really the least bad one. If we can confirm that, we would only be left with 2 settings to compare: naive vs scientific, with no further sub-options. That would make testing so much easier and faster!!!

For this dynamic vs manual comparison, please make sure that you test with a variety of scenes. E.g. green spear in Batman vs Superman, explosions in Mad Max, the orange car lamp in the Sony Campfire demo, etc. The one setting you choose as your most favorite setting must have the least amount of catastrophic failures in all those scenes. Please note: You are not allowed to pick different manual settings for different scenes (e.g. 50%/50% for explosions, and 15%/85% for the green spear). You need to nominate one single setting that you prefer, considering all the various scenes... I hope that we can all agree on "dynamic" (but I'm willing to listen, if at least 2 of you guys disagree). Once we agree on one setting, we can try to further improve it. Not sure how much improvement is possible, but we can try...
I had already done some tests with these and as far as I could see dynamic is the best compromise. I'm also happy to only have the ICtCp 100% saturation option (no 0%, no 50%, and maybe keep the xyY option for performance) for tonemapping and luminance reduction colorspace. Couldn't see much of a difference between these options and I settled on the ICtCp 100%. I'll do more thorough tests when I'm back if it's still up for debate.

I agree that reducing the options would make testing/comparing naive vs scientific much easier and faster.

I did some quick tests today and I liked scientific a lot in many instances, but there are still some situations where naive produces a more convincing result (usually with fire/explosions). Whether it is subjective/habit or whether it's because it's closer to what the colorist might have seen is difficult to say. I'd really like to find out, because honestly I don't care at all about preference. The fact that there is no reference makes it difficult to improve things with certainty, but if there was any way to know that one option gives use more often something closer to what was seen during the grading, that's the option I would use, even if I liked the result less (possibly because that's not what I'm used to see).

When looking at the Masciola patterns in various modes, I really wish there was a way to strike a compromise between perfect hue / significant desaturation / no clipping of scientific and variable hue / consistant saturation but some additional clipping (at least here, with my current SDR BT2020 calibration) of naive.

If the hard clip point in the custom curve isn't taken into account in scientific as it seems to be the case, maybe resolving this would allow to strike this compromise, or something closer to it?

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
Sponsored Links
Advertisement
 
Reply Digital Hi-End Projectors - $3,000+ USD MSRP

Tags
dynamic tone mapping , hdr , madvr , sdr , ton mapping

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off