madVR Tool: MadmeasureHDR Optimizer (Measurements, Dynamic Clipping & Target Nits) - Page 2 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 1014Likes
Reply
 
Thread Tools
post #31 of 1208 Old 01-04-2019, 04:25 AM
Member
 
Join Date: Nov 2017
Location: Stockholm, Sweden
Posts: 95
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 71 Post(s)
Liked: 23
Are the earlier created profile groups (Mannis profiles) of any use or when using Flo dynamic target nits?
Can they be safety removed or do i have to create a different set of groups, ie. a SDR group and a HDR group?
Attached Thumbnails
Click image for larger version

Name:	Profile group.PNG
Views:	59
Size:	196.2 KB
ID:	2505442  
arcspin is online now  
Sponsored Links
Advertisement
 
post #32 of 1208 Old 01-04-2019, 04:30 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Dexter Kane View Post
I've tried it with my actual measured screen nits (212) and while the selected target nits values were very high, 2 or 3 times what I would normally use, the results when playing are very good. I'll try a few more titles but one thing is that the brightness change can be quite noticeable, I'll try with a longer rolling average value. I've also noticed that often the dynamic target nits value is quite a lot higher than the tone mapping value, is your dynamic target nits algorithm being applied to the original measurements or the clipped measurements? It looks good I'm just wondering if it's intended or not.

EDIT: Actually, yeah I see what you mean, I will try it with a lower display nits value.
YEs.
Probaly you should use 212 as the minimum Target Nits, and "misuse" the "real peak nits" to something lower like 107 or lower.

Rolling avg value can be adjusted. That's the goal.
We have all the tastes in the world there.
I was just surprises myself NOT to see any brightness change to my eyes because centered rolling avg of 10s is already pretty long. And the bonus is it can reacts faster.

But I after look at the graphs of the target nits during the movie, I may prefer 1min smoothing: 1440 frames for 24p movies.

Quote:
Originally Posted by Manni01 View Post
I've done the measurements on my usual example titles, I like your targets better (assuming the dynamic targets work as intended, I haven't been able to do any visual testing).

Most of the time, we are very close, but on the titles where I need to compromise due to some dark scenes, your higher targets should provide better contrast and saturation on bright scenes, without losing shadow details on dark scenes. I attach a comparison, in bold the titles with the most differences.

If we could get more than five paths that would be useful.

Good night, will test tomorrow visually, it's too late to start the PJ now...

Looks very promising, well done
Glad you like it. The static target nits formula could be tweaked maybe a bit more but it works fairly well already.
The dynamic one is different even If I am using similar thought process for it.

More than 5 paths is easy to do. But like everything time consuming.
We'll put it on the to do list.

Quote:
Originally Posted by Dexter Kane View Post
It's interesting that without dynamic target nits when playing a video with a measurement file madvr doesn't seem to do any real time measuring and instead just reads values from the measurement file, which give a slight performance boost. But with dynamic target nits enabled it measures in real time again, and also there is less info about the measurements in the OSD (only frame/scene).
Well this is the bug I mentioned earlier to @madshi .
If dynamic target nits is selected, for some reason, the rest of the measurement file is ignored.
I am sure he will fix it in the next release.

Quote:
Originally Posted by Dexter Kane View Post
Okay I've tried out a few scenes using 107 for the display nits and for the most part it looks very good. The brightness change isn't distracting me anymore so that might only be an issue with using higher display nits, which doesn't seem to scale well as you said.

The only problem I'm seeing is that in some movies, like John Wick 2 and Deadpool for example there are scenes which are very dark but have some small but very bright lights in them which cause the target nits to be set much higher than it should (The bath scene in John Wick 2 at around 48 minutes). I'm not sure how your dynamic target nits works but I think it should look at the distribution of the highlights and not just the peak.

But it completely fixes movies which have very right and very dark scenes.
Good to hear :-)
I suppose you kept your real peak as "Minimum Target Nits"?

It's already kind of looking at the distribution...but it could look in more detail if needed.

Quote:
Originally Posted by toby5 View Post
I was messing around with this and I would say try and change your maximum target nit to a lower value for more brightness that's natural. Also, I am getting an error on my end. After scanning the measurement files it does it right. If I don't like it and delete the original folder it gives me an error saying a file should have 290 lines and mine has 291 afterward and won't scan again. This is the first time it's done this.
Yes "real target nits" might be a bad name since people taste are different, plus the scaling I am using may not be perfect. But even if it was perfect, I am sure that some poeple would like it differently with the same real nits on display.
A sliding bar would probably be better.

If you deleted the original folder, you deleted your original measurement with it! Never do that! And this means you try to reanalysed the already tweaked measurment which of course does not work.

Quote:
Originally Posted by Dexter Kane View Post

@Soulnight

I'm currently using a value of 75 for the display nits with a measured value of 212, so perhaps this should be renamed something else? Maybe treat it as a percentage value to adjust the overall target nits dynamic range? I think using the measured display nits as the minimal target nits is fairly objective and the display nits value can be used to adjust preference.

Overall from what I've looked at it's a huge improvement over static target nits.
Agreed. Slider would be better.

And yes, I also believe it is a huge improvement. Glad you also see it that way.

Quote:
Originally Posted by pandm1967 View Post
For someone who might be interested in the dynamic target nits, ability to read @Soulnight dynamic target nits data.
Awesome! Thanks!


Quote:
Originally Posted by Manni01 View Post
If you use any value below 200nits for the minimum target you will most likely cause visible fluctuation below diffuse white in the content (0-100nits will be tonemapped, which you should avoid with your 200+nits actual brightness). The measured display nits shouldn't be used as a minimal target UNLESS you have more than 200nits.

I personally would like to keep actual (measured) display nits as a parameter and change the scaling if it doesn't work, because that's what we want to end up with.

Did you try using 107nits (or 100nits) as the actual display brightness and 200nits as the minimum target? I plan to do tests this morning but haven't been able to do so yet.

My current settings are 100nits actual display nits, 200-4000nits min/max, flo algo. With these settings the static targets seem very good. I'll tell you about the dynamic target within a couple of hours hopefully.
It does make a lot of sense to have the minimum allowed target nits to be equal or higher than the REAL PEAK NITS that you have measured on your screen.

Having a target Nits lower than what you real peak is means that you showing the content brighter than it was intended to.

Having the target Nits EQUAL to your real peak measured on your screen, means that you watching the content with the same brightness it was intended for.

Having a target Nits higher than your real peak nits, means that you are scaling down the brightness compared to what it was intended for.

Having said that, the field "real peak in the software" does not need to be the real peak at all. As Dexter said. It's more a taste thing. And with the minimum target target settings, you ensure it does not go lower than your real peak display, so that the content never goes brighter than it was intended for.

So what you propose:
field real peak: 100nits
minimum target: 200 nits
should work fine for you.

The min and max target are not used in the static target algo. It's smart enough not to need it because it's using global avg value over the whole movie.

However for the dynamic target, they are used. I believe 1000 maximum target is really a good setting.
Going above that in combination with a projector should result into too dark a picture, and more brightness change.

Btw. @Manni01 , madVR tone mapping works like this :
if "frame peak nits" is lower than current "Target nits", then no tone mapping applied--> Render untouched (clipped at target nits).
This means that, if you have a target nits of 100nits, and your current frame peak is lower than 100nits, then everything below the diffuse white will stay untouched and nothing will get compressed at all.

Actually, if you have a display with 100 real nits, and you set your target nits to 100 nits, and the whole movie has peaks below 100nits (Mountain between us, Blade Runner 2049), you are effectively watching the movie exactly as it was intended for. No tone mapping applied and intended brightness respected.

But the same is even true for a dynamic target nits of 50 with a 50nits display. If the frame peak are lower than 50, then you effectively watching the content exactly as it was intended for with the exact correct brightness. No tone maping applied.

I understad that you do want to respect the old SDR rule on projector with a 2:1 ratio to diffuse white at all time.
But that's a philosophical decision.
I personnally prefer to watch the content exactly as bright as it was intended for whenever possible.
But good news: our tool allow you to choose your philosophy ;-)

Last edited by Soulnight; 01-04-2019 at 04:33 AM.
Soulnight is online now  
post #33 of 1208 Old 01-04-2019, 04:43 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Looking at the dynamic target nits profile generated with our tool and plotted with @pandm1967 tool, **maybe** 1 min centered rolling avg (1440 frames for 24p content) would be better.

Maybe not and it will be too slow.
Thing is, I don't see "visible" brightness variation even with 10s and then it reacts faster than it would with 1min.

Some testing is needed.

The meg:
- 50 real nits
- 240frames (10s) rolling avg
- min target nits=50
- max target nits=10000



The meg:
- 50 real nits
- 1440frames (1min) rolling avg
- min target nits=50
- max target nits=10000



The shallows:
- 50 real nits
- 240 frames (10s) rolling avg
- min target nits=50
- max target nits=10000


The shallows:
- 50 real nits
- 1440frames (1min) rolling avg
- min target nits=50
- max target nits=10000
Attached Thumbnails
Click image for larger version

Name:	TheMeg-10sRollAVG-DynTargetNits.PNG
Views:	552
Size:	421.4 KB
ID:	2505444   Click image for larger version

Name:	TheMeg-1minRollAVG-DynTargetNits.PNG
Views:	510
Size:	214.9 KB
ID:	2505446   Click image for larger version

Name:	TheShallows-10sRollAVG-DynTargetNits.PNG
Views:	491
Size:	307.5 KB
ID:	2505448   Click image for larger version

Name:	TheShallows-1minRollAVG-DynTargetNits.PNG
Views:	494
Size:	184.0 KB
ID:	2505450  
Manni01 likes this.

Last edited by Soulnight; 01-04-2019 at 05:41 AM.
Soulnight is online now  
Sponsored Links
Advertisement
 
post #34 of 1208 Old 01-04-2019, 05:11 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Lost_62 View Post
What it consists of deselect the setting (Apply dynamic Clipping) keeping only enabled (Apply target nits selection) that what changes for the purposes of measurement final?
Well, if you uncheck dynamic clipping, then you are choosing to loose a bit of contrast and punch.
And if you have only target nits selection checked, then you will only changed those value static or dynamic depending what you have chosen in the measurement file.

Quote:
Originally Posted by arcspin View Post
Are the earlier created profile groups (Mannis profiles) of any use or when using Flo dynamic target nits?
Can they be safety removed or do i have to create a different set of groups, ie. a SDR group and a HDR group?
You can also safely keep them.
They just get bypassed if a static target nits or a dynamic target nits has been specified in the measurement file.

Your old profiles will only be used if you did not overwrite the target nits in the measurement file.
Or if have chosen in the measurement file a manual target nits of 0 ... which is the start value.
arcspin likes this.
Soulnight is online now  
post #35 of 1208 Old 01-04-2019, 05:23 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,360
Mentioned: 361 Post(s)
Tagged: 0 Thread(s)
Quoted: 5648 Post(s)
Liked: 5978
@Soulnight

I've done some testing with my torture tests for brightness variation within the same scene, especially the wood attack scene in The Revenant and the opening of John Wick 2, and your dynamic algo passed it brilliantly with my settings (100nits actual display nits, flo algo, dynamic, 200nits min, [EDIT: the default 1000nits max is a better value for projectors than my initial 4000nits]. There was no visible fluctuation. The dynamic targets were spot on, even when the target fluctuated significantly in the same scene.

I also tested the darkest scene in The Revenant (camp fire scene) and it was perfect to see it at 200nits target, which is what this scene needs to resolve shadow details and provide enough saturation in such low light, while in the attack scene earlier the target would go up to 350-400 nits, which is what these scenes need to maximise contrast/highlights/saturation.

I also tested the transition at the end of chapter 3 / beginning of Chapter 4 in Pacific Rim, when the dark-ish scene at the end of chapter 3 needs around 500nits, while the highlights in the sky at the beg of chapter 4 need around 1300nits to resolve the clouds properly. that worked fine too, without any visible transition.

The Meg is much improved compared to my algo. My target was already too low for the Meg at 600nits for the bright scenes (800nits worked better for these), but that's because some dark scenes need down to 200nits. So in a title like this where targets go from 200nits to more than 1000nits, your dynamic algo does wonder.

Of course I had disabled my DI to make sure that any unwanted variations would come from your dynamic algo/MadVR and not the DI in the JVC. Since the issues I've seen with the DI in MI:Fallout from 05:57, I'm switching the DI off most of the time anyway.

I need to do more tests, but from what I've seen, your dynamic algo is much better than any static one, including mine.

Fantastic work, thank you!

Anyone with a dedicated room and around 100nits of actual peak brightness, I recommend trying the above settings.

@Dexter Kane : please could you provide title/timecode where you see instability with the settings above? I really can't see any reason to lower measured peak brightness to 75nits as you did, and you have almost twice as much brightness I have.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro

Last edited by Manni01; 01-04-2019 at 07:31 AM.
Manni01 is online now  
post #36 of 1208 Old 01-04-2019, 05:38 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Ian_Currie View Post
Hi Soulnight,

Thanks for starting this thread.

I've been attempting to use this tool, but could use a little info about how to use it properly. For example:

- at first glance it looks like you can either measure just one file or an entire folder but when you choose the 'Choose File' button it seems to be looking for a measurement file and not a .MKV file. Is the purpose of the CHOOSE FILE button to load a measurement file and modify it?
Hi!
Yes you can either measure a single file or a folder.

If you choose single file, you can measure a *.mkv *.ts and *.mp4 or a *.measurement file. You just have to pick the format in the bottom right corner ;-)

If you choose path instead, you will create measurement for ovie which have not been measured yet AND you will optinize the directly afterwards.
If measurement files already exist, the tool will check if they have been already optimized with the same settings. If yes, they are skipped, if not they are optimized.

Quote:
Originally Posted by Ian_Currie View Post
- When I used the CHOOSE PATH route it started measuring my MKV files. I noticed two things:
1) it reports that it might take up to 20 minutes when in fact it seems to be an hour plus for my files (using a 1080ti with files on a separate NAS). Looking at task manager performance I see the GPU is at 100% so I don't think it's network related?
EDIT: I missed this crucial tip: If you change LAV video filter to D3D11 (native) it works much faster.

2) The app seems to go into 'non-responding' mode (cursor is just a rotating circle instead of arrow) and I can't seem to click the STOP button.
EDIT: Tried this a second time and while the app still went white and seemed to not respond, it DID stop after the current movie and displayed a log file.
1) yes the lav filter setting are very important for correct speed. Nothing to do with our tool.

2) The non-responding issue is because we did not yet split the programm on different thread. This is on the do to list. It should makes things much more responsive and I expect also faster optimize process (currently about 5s per movie).


Quote:
Originally Posted by Ian_Currie View Post
- I see that it creates 3 files: one in an ORIGINAL folder and then a measurement file + a 'details.txt' file. Having read various posts about whether to store the measurement files in the same folder or in a dedicated one, it seems that at this point it is safer to keep them all together? I ask because I'd prefer them to be separate but don't want to make your app unhappy.
Let's say it's much simpler if the mkv are in the same folder than the measurement file.
So yes, if you can live with that: keep it that way.

Otherwise, and that's for @Javs , you need to go in madVR and select a "master folder for measurement file".
What will happen if you do that is that madVR madmeasureHDR will ITSELF save those measurement files in the said folder.
For this it does not matter what tool you use to generate those measurement file (Manni01 batch, Benrd tool, or ours or even directly the exe from madshi.

As Javs pointed out (I did not know before), the measurements file have been then renamed directly by madVR to include the path to the movie. If you move your movie, it will not work anymore. Same for the measurement files.

Now say you want to optimize your "Measurement folder":
Step 1: you select in madVR your measurement master folder
Step 2: you use one of the tools available and it will create the renamed measurements files in that said folder
Step 3: To optimize them, you need to select in our tool 2 times the path to this folder. Bottom left " measurement file folder" and at the top "paths".
If you keep "same as movie file it will not work because then it looks for mkv.
--> @Javs : you actually did the correct things the first time. I believe you only had access problem because it was on the root. Try again and move the folder somewhere else and select it twice in our programm.

Quote:
Originally Posted by Ian_Currie View Post
- Once measurement files exist, in what way should we be modifying them to match the peak nits of our display?
EDIT: I now see we can click on either the Manni or Neo-XP button as an 'adjustment' and then when re-processing the folder it will revise the measurement files to have the corresponding target nits. Still whether manual means 'use profile' or one's own desired value...?

Thanks so much for all your (and Anna's) work on this!
If you generate the measurement file with our tool. It wil do both directly: madmeasureHDR + optimize directly after.
If you have generated measurement files with an other tool, they will be only optimized.

Right now, just make sure you have dynamic clpping and target nits selected.
Also dynamic "check box" selelect.

Choose "Flo".

Enter a real peak nits between 50 and 107.
Go the minimum target and enter here your real peak.

Press start, you're done.

iF YOU CHOOSE manual: you can manual set a STATIC target nits to whatever value you like. Using this with "dynamic is not advised". And we have not implemented the possibility for the user to input his own rule. That's retty comlicated and the user would have to know how to code in Vb.ent.

Quote:
Originally Posted by NoTechi View Post
As of today I would run Bernds tool to measure since it is more reliable/responsive for me while measuring. Run MadmeasureHDR Optimizer to optimize.
NoTechi
This should get much better and responsive once we split the Main form thread from the calculation thread. On the to do list.
But right now, priority if to get the best picture quality first, so more work on the optinize algo kind of thing.
Manni01 likes this.
Soulnight is online now  
post #37 of 1208 Old 01-04-2019, 05:48 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Manni01 View Post
@Soulnight

I've done some testing with my torture tests for brightness variation within the same scene, especially the wood attack scene in The Revenant and the opening of John Wick 2, and your dynamic algo passed it brilliantly with my settings (100nits actual display nits, flo algo, dynamic, 200nits min, 4000nits max). There was no visible fluctuation. The dynamic targets were spot on, even when the target fluctuated significantly in the same scene.

I also tested the darkest scene in The Revenant (camp fire scene) and it was perfect to see it at 200nits target, which is what this scene needs to resolve shadow details and provide enough saturation in such low light, while in the attack scene earlier the target would go up to 350-400 nits, which is what these scenes need to maximise contrast/highlights/saturation.

I also tested the transition at the end of chapter 3 / beginning of Chapter 4 in Pacific Rim, when the dark-ish scene at the end of chapter 3 needs around 500nits, while the highlights in the sky at the beg of chapter 4 need around 1300nits to resolve the clouds properly. that worked fine too, without any visible transition.

The Meg is much improved compared to my algo. My target was already too low for the Meg at 600nits for the bright scenes (800nits worked better for these), but that's because some dark scenes need down to 200nits. So in a title like this where targets go from 200nits to more than 1000nits, your dynamic algo does wonder.

Of course I had disabled my DI to make sure that any unwanted variations would come from your dynamic algo/MadVR and not the DI in the JVC. Since the issues I've seen with the DI in MI:Fallout from 05:57, I'm switching the DI off most of the time anyway.

I need to do more tests, but from what I've seen, your dynamic algo is much better than any static one, including mine.

Fantastic work, thank you!

Anyone with a dedicated room and around 100nits of actual peak brightness, I recommend trying the above settings.

@Dexter Kane : please could you provide title/timecode where you see instability with the settings above? I really can't see any reason to lower measured peak brightness to 75nits as you did, and you have almost twice as much brightness I have.
Awesome! Thank you for the detailed feedback!

Yes, it's really a game changer compared to a static target nits (even if it's a good static one adapted to the said movie).

I opened the madVR tone mapping improvement thread with dynamic clipping and dynamic tone target nits in mind.
We ended up improving a lot madVR tone mapping not even touching those. :-)

And now they also both exist and not even one year after opening the original thread. So Happy.
Of course, we can probably improve further but for a first shot, I also think it works very well.

Cheers,

Flo
Manni01, neo_2009 and Dexter Kane like this.
Soulnight is online now  
post #38 of 1208 Old 01-04-2019, 05:50 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Manni01 View Post
@Soulnight

I've done some testing with my torture tests for brightness variation within the same scene, especially the wood attack scene in The Revenant and the opening of John Wick 2, and your dynamic algo passed it brilliantly with my settings (100nits actual display nits, flo algo, dynamic, 200nits min, 4000nits max). There was no visible fluctuation. The dynamic targets were spot on, even when the target fluctuated significantly in the same scene.

I also tested the darkest scene in The Revenant (camp fire scene) and it was perfect to see it at 200nits target, which is what this scene needs to resolve shadow details and provide enough saturation in such low light, while in the attack scene earlier the target would go up to 350-400 nits, which is what these scenes need to maximise contrast/highlights/saturation.

I also tested the transition at the end of chapter 3 / beginning of Chapter 4 in Pacific Rim, when the dark-ish scene at the end of chapter 3 needs around 500nits, while the highlights in the sky at the beg of chapter 4 need around 1300nits to resolve the clouds properly. that worked fine too, without any visible transition.

The Meg is much improved compared to my algo. My target was already too low for the Meg at 600nits for the bright scenes (800nits worked better for these), but that's because some dark scenes need down to 200nits. So in a title like this where targets go from 200nits to more than 1000nits, your dynamic algo does wonder.

Of course I had disabled my DI to make sure that any unwanted variations would come from your dynamic algo/MadVR and not the DI in the JVC. Since the issues I've seen with the DI in MI:Fallout from 05:57, I'm switching the DI off most of the time anyway.

I need to do more tests, but from what I've seen, your dynamic algo is much better than any static one, including mine.

Fantastic work, thank you!

Anyone with a dedicated room and around 100nits of actual peak brightness, I recommend trying the above settings.

@Dexter Kane : please could you provide title/timecode where you see instability with the settings above? I really can't see any reason to lower measured peak brightness to 75nits as you did, and you have almost twice as much brightness I have.
I only saw the fluctuations when I used 212 as the display peak, I'm using 75 because I found 107/100 to be too dark in some scenes. If you can, have a look at the bath scene in John Wick 2 at around 48 minutes, tell me what you think of how the dynamic tone mapping treats it. For me I think the target should be much lower.
Manni01 likes this.
Dexter Kane is offline  
post #39 of 1208 Old 01-04-2019, 05:55 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,360
Mentioned: 361 Post(s)
Tagged: 0 Thread(s)
Quoted: 5648 Post(s)
Liked: 5978
Quote:
Originally Posted by Soulnight View Post

It does make a lot of sense to have the minimum allowed target nits to be equal or higher than the REAL PEAK NITS that you have measured on your screen.
Agreed, this is what I said

Quote:
Originally Posted by Soulnight View Post
Having a target Nits lower than what you real peak is means that you showing the content brighter than it was intended to.
Agreed.

Quote:
Originally Posted by Soulnight View Post
Having the target Nits EQUAL to your real peak measured on your screen, means that you watching the content with the same brightness it was intended for.
Agreed, provided you are in a non-dedicated room with some ambient light, which is how consumer HDR content is mastered.


Quote:
Originally Posted by Soulnight View Post
Having a target Nits higher than your real peak nits, means that you are scaling down the brightness compared to what it was intended for.
Agreed, provided you are in a non-dedicated room with some ambient light, which is how consumer HDR content is mastered.

In a dedicated room with projectors, we have always mapped 0-100nits to 0-50nits, to emulate the 0-48nits of SDR cinema.

With HDR consumer content, we are trying to do exactly the same, map 0-100nits to 0-50nits, so that we have some room for headlights from 50nits up.

Otherwise, there is no room for highlights or you are tonemapping 0-50nits more than necessary.

Quote:
Originally Posted by Soulnight View Post
Having said that, the field "real peak in the software" does not need to be the real peak at all. As Dexter said. It's more a taste thing. And with the minimum target target settings, you ensure it does not go lower than your real peak display, so that the content never goes brighter than it was intended for.
In a non-dedicated room with ambient light, for displays with at least 600nits peak brightness. That's not what we have with projectors (usually a dedicated room, and rarely more than 250-300nits).

Quote:
Originally Posted by Soulnight View Post
So what you propose:
field real peak: 100nits
minimum target: 200 nits
should work fine for you.
It does

Quote:
Originally Posted by Soulnight View Post

The min and max target are not used in the static target algo. It's smart enough not to need it because it's using global avg value over the whole movie.

However for the dynamic target, they are used. I believe 1000 maximum target is really a good setting.
Going above that in combination with a projector should result into too dark a picture, and more brightness change.
I agree, I was considering lowering my 4000nits to 1000nits for this reason, though I have seen the peak nits going up to 1500nits and it didn't look too dark here (bat loft, no ambient light, 100nits actual peak brightness, YMMV).

Quote:
Originally Posted by Soulnight View Post

Btw. @Manni01 , madVR tone mapping works like this :
if "frame peak nits" is lower than current "Target nits", then no tone mapping applied--> Render untouched (clipped at target nits).
This means that, if you have a target nits of 100nits, and your current frame peak is lower than 100nits, then everything below the diffuse white will stay untouched and nothing will get compressed at all.

Actually, if you have a display with 100 real nits, and you set your target nits to 100 nits, and the whole movie has peaks below 100nits (Mountain between us, Blade Runner 2049), you are effectively watching the movie exactly as it was intended for. No tone mapping applied and intended brightness respected.
If you have a non-dedicated room with ambient light, yes.

Quote:
Originally Posted by Soulnight View Post

But the same is even true for a dynamic target nits of 50 with a 50nits display. If the frame peak are lower than 50, then you effectively watching the content exactly as it was intended for with the exact correct brightness. No tone maping applied.

I understad that you do want to respect the old SDR rule on projector with a 2:1 ratio to diffuse white at all time.
But that's a philosophical decision.
It's not a philosophical decision at all. It's about trying to emulate HDR in cinema (48nits for diffuse white, 49-107nits for highlights), which is designed for projectors (peak bnrightness of 107nits) in a dedicated room with no ambient light, rather than HDR in the consumer living room with ambient light and displays able to go up to 600-1200nits and more (100nits diffuse white, 101-10000nits for highlights).

I disagree that showing 0-100nits in the content as 0-100nits in a dedicated room with no ambient light is the way the content was intended to be shown. Otherwise we'd be watching SDR at 100nits.

For me Blade Runner 2047 HDR is shown properly in my dedicated room if it's mapped to 0-50nits, not if it's mapped to 0-100nits. That's not "as intended".

Quote:
Originally Posted by Soulnight View Post
I personnally prefer to watch the content exactly as bright as it was intended for whenever possible.
But good news: our tool allow you to choose your philosophy ;-)
Again, not as intended for projectors in a dedicated room as far as I'm concerned, but let's agree to disagree!

It's true that we can adjust settings as we want to, and I am grateful for this, so thanks again.

One last thing: beyond the brightness factor issue, it is my understanding that when the target is below 200nits, there is some tonemapping applied to 0-100nits. That's one of the reasons why I prefer to keep my targets at a 200nits min. Has this changed in any way?

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro
Manni01 is online now  
post #40 of 1208 Old 01-04-2019, 06:02 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,360
Mentioned: 361 Post(s)
Tagged: 0 Thread(s)
Quoted: 5648 Post(s)
Liked: 5978
Quote:
Originally Posted by Dexter Kane View Post
I only saw the fluctuations when I used 212 as the display peak, I'm using 75 because I found 107/100 to be too dark in some scenes. If you can, have a look at the bath scene in John Wick 2 at around 48 minutes, tell me what you think of how the dynamic tone mapping treats it. For me I think the target should be much lower.
Thanks I'll test this and will report back.

[EDIT: I did and did not see anything wrong. Maybe you are clipping shadow details on your display? The target is high but the scene is meant to be dark, and I couldn't see any unintentional black crush].

Maybe provide a more exact time code and your targets, as well as a screenshot, and we can compare.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro

Last edited by Manni01; 01-04-2019 at 06:07 AM.
Manni01 is online now  
post #41 of 1208 Old 01-04-2019, 06:07 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Manni01 View Post
Again, not as intended for projectors in a dedicated room as far as I'm concerned, but let's agree to disagree!

It's true that we can adjust settings as we want to, and I am grateful for this, so thanks again.
So yes, I agree to disagree

Quote:
Originally Posted by Manni01 View Post

One last thing: beyond the brightness factor issue, it is my understanding that when the target is below 200nits, there is some tonemapping applied to 0-100nits. That's one of the reasons why I prefer to keep my targets at a 200nits min. Has this changed in any way?
No. It does not work like that.

Remember the 480 Target Nits limit we discussed a while ago? Well if you choose that or above, you will NEVER compress anything below 0 to 100nits. No matter what the current frame peak is.

Now let's say you choose 200 Target nits.
If your frame peak is lower than the 200nits, then 0-200 will be rendered without tone mapping. MadVR swicth to the non tone-mapped clipping at 200nits algo.
So of course in the case 0-100 is also untouched.

Now, if for the same 200 target nits, your frame peak goes to 10000nits let's say, then the 0-100 will touched (but only the upper portion of it).

So for any target nits below 480 nits, it depends on your the coupled "Target nits ; current frame peak) if you are touching the 0-100nits or not.
If peak lower than target, all untouch below target.

There was a nice figure in the ITU document showing this.
I'll try to post it there.

Edit: there it is:
https://www.itu.int/dms_pub/itu-r/op...2017-PDF-E.pdf

Here you can see that the 500 nits curve (~480nits) stops compressing at the 100nits threshold.
(and this considering tone mapping to 10000nits so maximum compression needed)






Now one small complication to that is that there is a smooth transiention between "clipping/non-tone mapped" if peak lower than target and tone mapping when peak higher than target.
So if you have a target of 200, and your peak is 210, then probably 0-100 is still untouched os very very very slightly compressed on the upper hand.
That was the formula given by @fhoech @madshi .
Attached Thumbnails
Click image for larger version

Name:	ITU-R BT-2390-3.PNG
Views:	423
Size:	428.1 KB
ID:	2505484  
Manni01 likes this.

Last edited by Soulnight; 01-04-2019 at 06:19 AM.
Soulnight is online now  
post #42 of 1208 Old 01-04-2019, 06:13 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Manni01 View Post
Thanks I'll test this and will report back.

[EDIT: I did and did not see anything wrong. Maybe you are clipping shadow details on your display? The target is high but the scene is meant to be dark, and I couldn't see any unintentional black crush].

Maybe provide a more exact time code and your targets, as well as a screenshot, and we can compare.
I'm not seeing any black crush, it's just unnecessarily dark. On a scene like that I'd rather preserve the 0-100 part, I don't think anything would be lost in the highlights at a lower target nits.
Dexter Kane is offline  
post #43 of 1208 Old 01-04-2019, 06:18 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Dexter Kane View Post
I'm not seeing any black crush, it's just unnecessarily dark. On a scene like that I'd rather preserve the 0-100 part, I don't think anything would be lost in the highlights at a lower target nits.
Again this is really a taste thing. I think we all need to agree to disagree on that.
Everybody can choose what pleases him most: visually, theoretically etc.

ps: that being said.
I am really against having the minimum target nits set lower than what you really have measured on your screen.
Because this means it can results in scenes being even brighter than it was meant to be watched on a TV then.

- Same is good: if you want to have the intended TV brightness
- Twice is also good: if you want to have dolby theater like brightness.
RioBar4U and Dexter Kane like this.

Last edited by Soulnight; 01-04-2019 at 06:24 AM.
Soulnight is online now  
post #44 of 1208 Old 01-04-2019, 06:18 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,360
Mentioned: 361 Post(s)
Tagged: 0 Thread(s)
Quoted: 5648 Post(s)
Liked: 5978
Quote:
Originally Posted by Dexter Kane View Post
I'm not seeing any black crush, it's just unnecessarily dark. On a scene like that I'd rather preserve the 0-100 part, I don't think anything would be lost in the highlights at a lower target nits.
I think the darkness is intentional, it's in a crypt!

I need 200nits for the Revenant Campfire scene (6nits peak brightness) but that scene's peak brightness is much higher, so I don't find it too dark even with a target of 800nits or so. Please provide the exact time code, so I'm sure we're talking about the same shot, and a screenshot of the OSD so we can compare values/targets, and I'll take another look.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro
Manni01 is online now  
post #45 of 1208 Old 01-04-2019, 06:48 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Manni01 View Post
I think the darkness is intentional, it's in a crypt!

I need 200nits for the Revenant Campfire scene (6nits peak brightness) but that scene's peak brightness is much higher, so I don't find it too dark even with a target of 800nits or so. Please provide the exact time code, so I'm sure we're talking about the same shot, and a screenshot of the OSD so we can compare values/targets, and I'll take another look.
Sorry, I know these aren't the same frames, I had to change the measurement file for one and I messed up another, but it's close enough to show what I'm talking about.

Dynamic:



350 nits:



200 nits:




I don't think the image is improved at all by the higher target nits, it's just darker and harder to make out detail. Out of these three I prefer 350. I understand that there's a degree of personal preference involved but isn't the point of the dynamic target nits to preserve dark scenes like this while also increasing contrast in brighter scenes that would be over compressed?
Manni01 likes this.
Dexter Kane is offline  
post #46 of 1208 Old 01-04-2019, 07:07 AM
Member
 
Lost_62's Avatar
 
Join Date: Apr 2018
Posts: 102
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 65 Post(s)
Liked: 5
Quote:
Originally Posted by Soulnight View Post
Well, if you uncheck dynamic clipping, then you are choosing to loose a bit of contrast and punch.
And if you have only target nits selection checked, then you will only changed those value static or dynamic depending what you have chosen in the measurement file.

OK. I think that I have understand , question in general because I want to understand better ...all these optimizations ,Manual,Neo-xp,Manni01 and Flo in (Target nits selection) at the end , they are optimizations for those who own a projector ? for those who own an OLED how they should be interpreted ? for example , my OLED it measures 680 nits of peak of brightness , how should I interpret these data in association to your instrument ? thanks and I hope not to have written "hogwash"
Lost_62 is offline  
post #47 of 1208 Old 01-04-2019, 07:08 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,360
Mentioned: 361 Post(s)
Tagged: 0 Thread(s)
Quoted: 5648 Post(s)
Liked: 5978
Quote:
Originally Posted by Dexter Kane View Post

I don't think the image is improved at all by the higher target nits, it's just darker and harder to make out detail. Out of these three I prefer 350. I understand that there's a degree of personal preference involved but isn't the point of the dynamic target nits to preserve dark scenes like this while also increasing contrast in brighter scenes that would be over compressed?
I agree that there is no improvement in this case, but there is nothing objectionable either, at least here.

What you might want to change is the area taken into account before clipping, because in this scenes there are lots of candles which must be taken into account for the tone mapping. that might be better than having too high targets for most scenes, which is what you achieve with a 75nits peak brightness.

I'm happy to compromise slightly in scenes like this as long as I get dark scenes and bright ones much better, such as in The Revenant and The Meg.

In order to deal with a scene like this, you would need my suggested algo, which never, ever touches 0-100 whatever the tonemapping (and ideally maps it 2:1 for projectors if wished), but that means two different brightness factors, and I'm not even sure it wouldn't look weird in the transition area between the two brightness factors. Maybe Madshi will implement it (or a version of it) in the next stage.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro
Manni01 is online now  
post #48 of 1208 Old 01-04-2019, 07:17 AM
Advanced Member
 
grendelrt's Avatar
 
Join Date: Jun 2004
Location: Richmond, VA
Posts: 962
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 578 Post(s)
Liked: 310
Quote:
Originally Posted by Dexter Kane View Post
Sorry, I know these aren't the same frames, I had to change the measurement file for one and I messed up another, but it's close enough to show what I'm talking about.

Dynamic:



350 nits:




I don't think the image is improved at all by the higher target nits, it's just darker and harder to make out detail. Out of these three I prefer 350. I understand that there's a degree of personal preference involved but isn't the point of the dynamic target nits to preserve dark scenes like this while also increasing contrast in brighter scenes that would be over compressed?
What does the 1080p Blu Ray version of that screen look like? Both look ok, I wouldn't immediately think there is anything wrong with either, but the 200 example I can tell its too bright off the bat. The 350 is getting close the way she kinda pops in the darkness, but I probably wouldn't notice that without the comparison. Would be curious how she look in the blu ray since that would probably remove the candles as being a high value in a dark scene. As Manni said, if its an outlier and you arent noticing it in a lot of content, its probably just a trade off made for 95% of content to look better and live with the 5% that doesn't.
grendelrt is offline  
post #49 of 1208 Old 01-04-2019, 07:17 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Manni01 View Post
I agree that there is no improvement in this case, but there is nothing objectionable either, at least here.

What you might want to change is the area taken into account before clipping, because in this scenes there are lots of candles which must be taken into account for the tone mapping. that might be better than having too high targets for most scenes, which is what you achieve with a 75nits peak brightness.

I'm happy to compromise slightly in scenes like this as long as I get dark scenes and bright ones much better, such as in The Revenant and The Meg.

In order to deal with a scene like this, you would need my suggested algo, which never, ever touches 0-100 whatever the tonemapping (and ideally maps it 2:1 for projectors if wished), but that means two different brightness factors, and I'm not even sure it wouldn't look weird in the transition area between the two brightness factors. Maybe Madshi will implement it (or a version of it) in the next stage.
It would be interesting to see how something like that would work.

I don't think we'd have to give anything up to improve scenes like this, I can only speculate on how the algorithm works but changing how it looks at distribution so scenes like this get a lower target shouldn't affect scenes which already have a low target or scenes that have a lot of highlights. But if making this scene look better made those other scenes look worse then I agree that it's a worthwhile compromise.
Dexter Kane is offline  
post #50 of 1208 Old 01-04-2019, 07:22 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by grendelrt View Post
What does the 1080p Blu Ray version of that screen look like? Both look ok, I wouldn't immediately think there is anything wrong with either, but the 200 example I can tell its too bright off the bat. The 350 is getting close the way she kinda pops in the darkness, but I probably wouldn't notice that without the comparison. Would be curious how she look in the blu ray since that would probably remove the candles as being a high value in a dark scene. As Manni said, if its an outlier and you arent noticing it in a lot of content, its probably just a trade off made for 95% of content to look better and live with the 5% that doesn't.
From the blu-ray:



seems a lot closer to how 200 looks
grendelrt likes this.
Dexter Kane is offline  
post #51 of 1208 Old 01-04-2019, 07:27 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,360
Mentioned: 361 Post(s)
Tagged: 0 Thread(s)
Quoted: 5648 Post(s)
Liked: 5978
Quote:
Originally Posted by Dexter Kane View Post
I don't think we'd have to give anything up to improve scenes like this, I can only speculate on how the algorithm works but changing how it looks at distribution so scenes like this get a lower target shouldn't affect scenes which already have a low target or scenes that have a lot of highlights. But if making this scene look better made those other scenes look worse then I agree that it's a worthwhile compromise.
Yes, this is why I said we should look at the distribution factor / % of pixels taken into account, rather than selecting a too low peak brightness.

I'm sure Soulnight can look at this and see if something in the algo can be improved, but I wouldn't get the wrong targets for most scenes just to handle this one. In that case it's probably better to stay with static.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro
Manni01 is online now  
post #52 of 1208 Old 01-04-2019, 07:30 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
@Dexter Kane :
What target nits do you get for this frame with "real peak" 50nits?

Probably something like 500nits, correct?

Basically I am switching smoothly between 2 algos.

1 fully activated at "50nits" real peak
1 fully activated at "100" nits real peak

At the beginning I only had the 50 one which is less sensitive to highlights having a bright avg highlights nits value IF it corresponds to only a small portion of the picture (basically what you asked for).

I should really rename real peak nits to something else

Last edited by Soulnight; 01-04-2019 at 07:33 AM.
Soulnight is online now  
post #53 of 1208 Old 01-04-2019, 07:33 AM
Advanced Member
 
grendelrt's Avatar
 
Join Date: Jun 2004
Location: Richmond, VA
Posts: 962
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 578 Post(s)
Liked: 310
Quote:
Originally Posted by Dexter Kane View Post
From the blu-ray:



seems a lot closer to how 200 looks
Wow she looks super bright haha, not what I was expecting. Dynamic is way too dark in comparison, hopefully that is something that can be fixed as you said without affecting other content.
grendelrt is offline  
post #54 of 1208 Old 01-04-2019, 07:41 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Soulnight View Post
@Dexter Kane :
What target nits do you get for this frame with "real peak" 50nits?

Probably something like 500nits, correct?

Basically I am switching smoothly between 2 algos.

1 fully activated at "50nits" real peak
1 fully activated at "100" nits real peak

At the beginning I only had the 50 one which is less sensitive to highlights having a bright avg highlights nits value IF it corresponds to only a small portion of the picture (basically what you asked for).

I should really rename real peak nits to something else
yes it's around 485 in that shot. This scene looks much better with 50 nits actually, but I suspect some of the brighter scenes will look better with a higher value.
Dexter Kane is offline  
post #55 of 1208 Old 01-04-2019, 07:43 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Manni01 View Post
Yes, this is why I said we should look at the distribution factor / % of pixels taken into account, rather than selecting a too low peak brightness.

I'm sure Soulnight can look at this and see if something in the algo can be improved, but I wouldn't get the wrong targets for most scenes just to handle this one. In that case it's probably better to stay with static.
Already doing it in a simple way IF you have selected 50nits real nits.
Again no need to be scared by the number. I just switch between 1 algo and another.

But I will try to propose a new algo which always scales with the highlights distribution

It's actually rare than a scene has super bright avg highlights ( avg above 100 nits) and at the same time very dark.

The current algo try to always set the target nits ideally in regards to the "avg highlights ".
Manni01 likes this.
Soulnight is online now  
post #56 of 1208 Old 01-04-2019, 07:44 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by grendelrt View Post
Wow she looks super bright haha, not what I was expecting. Dynamic is way too dark in comparison, hopefully that is something that can be fixed as you said without affecting other content.
Yeah, to be a fair comparison though you'd have to double the brightness of the HDR ones. The 700 nits one is too dark though, it's dark enough that the image becomes flat and detail is hard to distinguish.
Dexter Kane is offline  
post #57 of 1208 Old 01-04-2019, 07:48 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Soulnight View Post
Already doing it in a simple way IF you have selected 50nits real nits.
Again no need to be scared by the number. I just switch between 1 algo and another.

But I will try to propose a new algo which always scales with the highlights distribution

It's actually rare than a scene has super bright avg highlights ( avg above 100 nits) and at the same time very dark.

The current algo try to always set the target nits ideally in regards to the "avg highlights ".
That makes sense, looking forward to trying the new algo.

Other than a handful of scenes like this one it's been working great, I watched all of Close Encounters today and it was perfect.
Manni01 likes this.
Dexter Kane is offline  
post #58 of 1208 Old 01-04-2019, 07:53 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
Quote:
Originally Posted by Dexter Kane View Post
yes it's around 485 in that shot. This scene looks much better with 50 nits actually, but I suspect some of the brighter scenes will look better with a higher value.
Nope. That's what algo 1 (50nits) is designed for.

If global picture brightness is very high (meg outside for example) then you will get the same target nits as with the algo 2 (100nits).


They only differentiate themselves for darker scenes.
Algo 1 "50 nits" is the smarter one in that sense.


Basically if you choose 50nits, the formula becomes:

Avg highlights nits + (avg frame nits / avg highlights nits) × avg highlights nits


With:
avg highlights= avg nits of pixels above 100nits

Avg frame nits: basically Fall for the current frame)

For dark scene, the factor(Fall / avg HL) is close to zero. For bright scene like ourside in the meg, equal to 1.

So the target nits algo varies between avg HL nits and 2 times avg HL nits.


With algo 2: it's always 2 times avg HL

And in between a mix of both.


Of course, it's also looking NOT to get higher than the current frame peak. Which could sometimes happen with 2 times avg HL.
Manni01 and Dexter Kane like this.
Soulnight is online now  
post #59 of 1208 Old 01-04-2019, 07:58 AM
Member
 
Join Date: Nov 2018
Posts: 173
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Soulnight View Post
Nope. That's what algo 1 (50nits) is designed for.

If global picture brightness is very high (meg outside for example) then you will get the same target nits as with the algo 2 (100nits).


They only differentiate themselves for darker scenes.
Algo 1 "50 nits" is the smarter one in that sense.


Basically if you choose 50nits, the formula becomes:

Avg highlights nits + (avg frame nits / avg highlights nits) × avg highlights nits


With:
avg highlights= avg nits of pixels above 100nits

Avg frame nits: basically Fall for the current frame)

For dark scene, the factor(Fall / avg HL) is close to zero. For bright scene like ourside in the meg, equal to 1.

So the target nits algo varies between avg HL nits and 2 times avg HL nits.


With algo 2: it's always 2 times avg HL

And in between a mix of both.


Of course, it's also looking NOT to get higher than the current frame peak. Which could sometimes happen with 2 times avg HL.
Well that explains why 75 looks better than 100, and probably why 212 had all sorts of problems

I'll give 50 a go tomorrow and see if I prefer that overall.
Dexter Kane is offline  
post #60 of 1208 Old 01-04-2019, 08:05 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,472
Mentioned: 224 Post(s)
Tagged: 0 Thread(s)
Quoted: 1121 Post(s)
Liked: 1823
By the way, my thanks to @Neo-XP for the inspiration coming from his static target nits algos. ;-)
Manni01, toby5 and Neo-XP like this.
Soulnight is online now  
Sponsored Links
Advertisement
 
Reply Home Theater Computers

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