Improving Madvr HDR to SDR mapping for projector - Page 157 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 4754Likes
Reply
 
Thread Tools
post #4681 of 6031 Old 12-29-2018, 07:01 AM
Senior Member
 
steevo123's Avatar
 
Join Date: Oct 2004
Location: NYC
Posts: 363
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 170 Post(s)
Liked: 48
Quote:
Originally Posted by Manni01 View Post
@madshi

I've posted the measurements for my HDR example titles in my algo post (link in sig).

I've left the previous measurements, so we can compare the measurements taken with black bars (older builds) and without (latest build).

As expected, MaxFALL, AvgFALL and AvgFMLL values go up as the aspect ratio gets wider.

It looks like you might have slightly changed the calculations for MaxCLL, as it usually goes up but occasionally goes down.

I haven't spotted anything weird, but I thought the data might be useful if you (and others) wanted to check for these 25 titles or so.

I'm hoping to be able to do some watching over the next few days...

Question - if Ive used your algo to create a profile, is there a setting in madvr to load that profile automatically ? (without needing a keyboard shortcut) Thanks!

Some guy
steevo123 is offline  
Sponsored Links
Advertisement
 
post #4682 of 6031 Old 12-29-2018, 07:50 AM
Member
 
the_jaguar's Avatar
 
Join Date: Dec 2008
Posts: 197
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 155 Post(s)
Liked: 50
Quote:
Originally Posted by pandm1967 View Post
Yes,This is what I did. I use "Run Measure" of madMeasurementMakeHardLink to measure the main stream of UHDs, because Kodi DSPlayer use index.bdmv and MPC use ?????.mpls as main media file, so I use "Make Hard Link" to create hard links of these measurement files.
That's awesome! Let me try out the tool and report back here.
the_jaguar is offline  
post #4683 of 6031 Old 12-29-2018, 08:49 AM
Member
 
Join Date: Nov 2016
Posts: 114
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 85 Post(s)
Liked: 54
I just did a fresh install of my windows including a fresh install of madvr, lavfilter and mpc-be. First test with "Mission Impoosible Fallout" shows a strange effect. The light/shine in this scene in the beginning of the movie is kinda spreaded over the screen. Same for other light sources in this dark scene like the car lights. Is this correct and within the movie or am I missing some setting?

https://ibb.co/rFWTm87


https://ibb.co/VCXB3wY

I have basically eveything on standard settings since it's a fresh install including madvr. Within the trade performance section everything is unchecked.

NoTechi
NoTechi is online now  
Sponsored Links
Advertisement
 
post #4684 of 6031 Old 12-29-2018, 08:50 AM
Member
 
BerndFfm's Avatar
 
Join Date: Jan 2018
Location: Frankfurt / Main, Germany
Posts: 99
Mentioned: 19 Post(s)
Tagged: 0 Thread(s)
Quoted: 94 Post(s)
Liked: 104
Peter : Update for You !

The Redirecting of the Output of Meadmeasure.exe to show the progress ist very critical. It is possible that my tool hangs up while reading the output.

I removed it, now you have no progress. I don't need it, I don't watch the pc while measuring movies.

Maybe it is working better now.

I just saw that my Task Manager is not working well (Windows 10). I have to close it and to open it again to see the madmeasurehdr.exe wörking.

NoTechi : yes, that in the movie.

Bernd

Home Cinema with : Onkyo 1009, Nubert, JVC RS-500, madVR

Last edited by BerndFfm; 12-29-2018 at 09:59 AM.
BerndFfm is offline  
post #4685 of 6031 Old 12-29-2018, 09:58 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 8,988
Mentioned: 307 Post(s)
Tagged: 0 Thread(s)
Quoted: 5343 Post(s)
Liked: 5375
Quote:
Originally Posted by NoTechi View Post
Is this correct and within the movie or am I missing some setting?
It's in the movie. When you have a doubt, switch to passthrough, then you'll see what's in the content and what's done by MadVR tonemapping.

@madshi : in that scene from 05:57, the highlights in the headlights are badly destroyed (posterization/banding) by MadVR's tonemapping when the DI is enabled in my RS500, even in auto2 which limits gamma processing. If I switch the DI off, or if I use passthrough, the artifacts go away.

It's hard to see which MadVR option does this as the options are limited in the latest build. It's not "restore highlights", as I switched that to off and it didn't make a difference. I'm using a 425nits target with that film.

I attach a screenshot to show you were though it doesn't show the issue as it's really a weird interference with the gamma processing of the JVC DI.

But on screen it's so bad that I had to disable the DI for that scene/film, way too distracting. I don't know if you have a DI on your JVC but you might want to check that scene, it's very ugly.
Attached Thumbnails
Click image for larger version

Name:	Mission- Impossible Fallout.mkv_snapshot_00.05.58_[2018.12.29_17.50.07].png
Views:	43
Size:	6.01 MB
ID:	2502354  
NoTechi likes this.

JVC Autocal Software V11 Calibration for 2019 Models
Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #4686 of 6031 Old 12-29-2018, 11:05 AM
Member
 
danbez's Avatar
 
Join Date: Feb 2004
Location: Seattle
Posts: 165
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 56 Post(s)
Liked: 23
Quote:
Originally Posted by Manni01 View Post
I've updated my algo(link in my sig) following the new measurements. Here is the explanation:

[EDIT 29-12-18: from test build 38, the black bars are not taken into account anymore. I have updated my algo and for now I've decided to experiment with something very close to @Dexter Kane 's values, apart from a few tuned thresholds. The profiles are often higher than with my older algo as his curve for AVGFMLL is less linear than mine, but it might be a good thing so I'm going to try and see how it is. As I'm going to get close to 140nits in low lamp with my new PJ, I might be happy with this. I'm going to set my lamp to high with the rs500 to get 145nits and see what I like and don't like. Unless I'm very unhappy with a few selections, I'm unlikely to finetune this until I get my new PJ anyway. For now I've taken out all the exceptions/special cases. I haven't had the time to update the list of films attached below, I'll do this when I find the time].

Dexter, thanks for the starting point. I think we had arrived at something very similar, the main difference being the black bars. I thought it would make sense to start with your values now that we have the same measurements as the black bars are not taken into account anymore. I slightly changed a few thresholds to catch a few exceptions, but it's mostly your last values, as linked in the post describing my algo. Let us know if you've changed it since. I plan to watch my usual test files over the next few days (I haven't done any visual testing yet). Can you remind me what your actual peak brightness is?
Manni, I got confused - is your new algo assuming 140nits or still the same 100nits target but with black bars removed from the equation? Since I have a similar environment (100 nits, dark room), I managed to get a free pass using your algo so far :-)
danbez is offline  
post #4687 of 6031 Old 12-29-2018, 11:18 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 8,988
Mentioned: 307 Post(s)
Tagged: 0 Thread(s)
Quoted: 5343 Post(s)
Liked: 5375
Quote:
Originally Posted by danbez View Post
Manni, I got confused - is your new algo assuming 140nits or still the same 100nits target but with black bars removed from the equation? Since I have a similar environment (100 nits, dark room), I managed to get a free pass using your algo so far :-)
I'm doing tests at both 100nits (low lamp) and 140nits (high lamp) at the moment and in most cases I prefer 100nits.

I'm also trying Dexter Kane's values because I agree that The Meg needs more than 600nits. I tried 800nits and it works surprisingly well, even at 100nits actual peak.

I suggest you try my new algo, and if you find it a bit too dim, try Dexter Kane's. I might end up with his values bar a few tweaks, they seem to work well.

I'm reluctant to spend too much time finetuning these as hopefully we won't need profiles anymore for HDR very soon, that's why I embraced Dexter's values to hold me over.

Waiting for Soulnight's delivery now to see if any more work is necessary or if we can finally ditch profiles...

JVC Autocal Software V11 Calibration for 2019 Models
Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #4688 of 6031 Old 12-29-2018, 12:02 PM
AVS Forum Club Gold
 
Ian_Currie's Avatar
 
Join Date: Nov 2001
Location: Holliston, MA
Posts: 1,702
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 293 Post(s)
Liked: 88
Quote:
Originally Posted by Manni01 View Post

I'm reluctant to spend too much time finetuning these as hopefully we won't need profiles anymore for HDR very soon.
Manni, can you expound on this?

JVC-RS4500 4k projector, Lumagen Pro, 138" 2.35:1 ST130
Panasonic DMP-UB820 & i7 PC w/1080ti running Win10/Kodi DSPlayer/madVR
Classe SSP-800, Bryston amplification, Wilson Audio speakers (7.2) + Buttkickers.
Ian_Currie is offline  
post #4689 of 6031 Old 12-29-2018, 12:49 PM
AVS Forum Special Member
 
Sittler27's Avatar
 
Join Date: Aug 2012
Posts: 1,461
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1127 Post(s)
Liked: 86
I recently began using Madvr and it's tone mapping for HDR.
Before that, I was using Manni's custom curves loaded into Custom 1-3 on my JVC x970R.

I noticed that when using madvr tone mapping, I still use high lamp, iris at 0, BT.2020 on the JVC like I would before madvr for HDR material, but the Gamma has to be set to to Normal (not one of the custom curve slots I put Manni's curves into).

Is this correct?

Are there other settings I should be setting on the JVC to get it to work with madvr properly?
Sittler27 is offline  
post #4690 of 6031 Old 12-29-2018, 01:18 PM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 8,988
Mentioned: 307 Post(s)
Tagged: 0 Thread(s)
Quoted: 5343 Post(s)
Liked: 5375
Quote:
Originally Posted by Ian_Currie View Post
Manni, can you expound on this?
For those of us using measurements files, Madshi has added a new target nits field in each file so we can specify a target nits for each title. Provided Soulnight's utility allows to populate this new field automatically using a selected formula, we won't need profiles anymore for HDR, each title will have its own target automatically used by MadVR when populated.

Those who don't use measurements files will still need profiles.

JVC Autocal Software V11 Calibration for 2019 Models
Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #4691 of 6031 Old 12-29-2018, 01:24 PM
AVS Forum Club Gold
 
Ian_Currie's Avatar
 
Join Date: Nov 2001
Location: Holliston, MA
Posts: 1,702
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 293 Post(s)
Liked: 88
Quote:
Originally Posted by Manni01 View Post
For those of us using measurements files, Madshi has added a new target nits field in each file so we can specify a target nits for each title. Provided Soulnight's utility allows to populate this new field automatically using a selected formula, we won't need profiles anymore for HDR, each title will have its own target automatically used by MadVR when populated.

Those who don't use measurements files will still need profiles.
Thank you. I am changing the target nits manually for titles which is a bit of a pain. I'm looking forward to learning how to make HDR profiles (I use profiles for resolutions in many categories but can't figure out how to make them for HDR). Either I figure it out or learn how to use Soulnight's utility.

Either way it seems life will get more convenient.

JVC-RS4500 4k projector, Lumagen Pro, 138" 2.35:1 ST130
Panasonic DMP-UB820 & i7 PC w/1080ti running Win10/Kodi DSPlayer/madVR
Classe SSP-800, Bryston amplification, Wilson Audio speakers (7.2) + Buttkickers.
Ian_Currie is offline  
post #4692 of 6031 Old 12-29-2018, 01:27 PM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,273
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 934 Post(s)
Liked: 1368
madmeasuredynamicclipping Tool V2.5

Hi guys

So, after Madshi Christmas gift, here is ours. (or the begining of it)

Edit: changed to V2.6 in the attachement and fixed a bug

New update of our madMeasure dynamic clipping tool in V2.5.
Attached 2 zip file: one with an *exe only, the other one with an installer.

NEW:

1) compatible with madVRmeasureHDR V38 ( and hopefully upwards ;-) )
2) clearly separated calculated values before and after clipping in the GUI
3) Button to delete all measurement files
4) Removed self calculated "avgFALL & maxFALL" without black-bars since Madshi now uses directly the correct definition.
5) Export madMeasure output directly to log files (no % update anymore )

6) New section to populate the "Target Nits" available in the header (it works, thanks @madshi )
Available right now as "Target Nits" :
a) Manual
b) Neo-XP: avgFALL + avg avg highlights
c) Manni01 algo of today, scaled to the entered peak nits of your display.
@Manni01 has 100nits, so if you specify 100nits, you will get exactly his values.
If you put: real display target nits 50nits, then all the target nits you will get will be exactly half of the one used by Manni

We plan to (try to) add to the Target Nits selection to possibility to add yourself your own script or formula as "user defined".
But this is already a good starting point right now.
We will further experiment ourselves but ideas for a global formula are welcomed.

Warning:
When a Target Nits is saved in the measurement, it overules whatever "Target Nits" profile you have set-up directly in madVR.

Enjoy,

Anna&Flo

ps: it looks like this now:



Edit: changed to V2.6. Please re-download. V2.5 had a bug.



OUTDATED: please follow link to latest version in signature
Attached Thumbnails
Click image for larger version

Name:	madmeasure dynamicclipping tool V2.5.PNG
Views:	332
Size:	221.9 KB
ID:	2502476  
Attached Files
File Type: zip madMeasureDynamicClipping V2.6 exe.zip (314.1 KB, 36 views)
File Type: zip madMeasureDynamicClipping Installer V2.6.zip (710.3 KB, 28 views)
SamuriHL, stef2, Manni01 and 11 others like this.

Last edited by Soulnight; 12-31-2018 at 02:41 AM.
Soulnight is offline  
post #4693 of 6031 Old 12-29-2018, 01:47 PM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 8,988
Mentioned: 307 Post(s)
Tagged: 0 Thread(s)
Quoted: 5343 Post(s)
Liked: 5375
Quote:
Originally Posted by Soulnight View Post
Hi guys

So, after Madshi Christmas gift, here is ours. (or the begining of it)

New update of our madMeasure dynamic clipping tool in V2.5.
Attached 2 zip file: one with an *exe only, the other one with an installer.

NEW:

1) compatible with madVRmeasureHDR V38 ( and hopefully upwards ;-) )
2) clearly separated calculated values before and after clipping in the GUI
3) Button to delete all measurement files
4) Removed self calculated "avgFALL & maxFALL" without black-bars since Madshi now uses directly the correct definition.
5) Export madMeasure output directly to log files (no % update anymore )

6) New section to populate the "Target Nits" available in the header (it works, thanks @madshi )
Available right now as "Target Nits" :
a) Manual
b) Neo-XP: avgFALL + avg avg highlights
c) Manni01 algo of today, scaled to the entered peak nits of your display.
@Manni01 has 100nits, so if you specify 100nits, you will get exactly his values.
If you put: real display target nits 50nits, then all the target nits you will get will be exactly half of the one used by Manni

We plan to (try to) add to the Target Nits selection to possibility to add yourself your own script or formula as "user defined".
But this is already a good starting point right now.
We will further experiment ourselves but ideas for a global formula are welcomed.

Warning:
When a Target Nits is saved in the measurement, it overules whatever "Target Nits" profile you have set-up directly in madVR.

Enjoy,

Anna&Flo
Xmas again!

Thank you both, sounds great.

Looking forward to testing and saying bye-bye to profiles!
Soulnight likes this.

JVC Autocal Software V11 Calibration for 2019 Models
Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #4694 of 6031 Old 12-29-2018, 02:20 PM
AVS Forum Special Member
 
stef2's Avatar
 
Join Date: Mar 2005
Location: Canada
Posts: 1,271
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 482 Post(s)
Liked: 282
Red face

Quote:
Originally Posted by Soulnight View Post
Hi guys

So, after Madshi Christmas gift, here is ours. (or the begining of it)

New update of our madMeasure dynamic clipping tool in V2.5.
Attached 2 zip file: one with an *exe only, the other one with an installer.

NEW:

1) compatible with madVRmeasureHDR V38 ( and hopefully upwards ;-) )
2) clearly separated calculated values before and after clipping in the GUI
3) Button to delete all measurement files
4) Removed self calculated "avgFALL & maxFALL" without black-bars since Madshi now uses directly the correct definition.
5) Export madMeasure output directly to log files (no % update anymore )

6) New section to populate the "Target Nits" available in the header (it works, thanks @madshi )
Available right now as "Target Nits" :
a) Manual
b) Neo-XP: avgFALL + avg avg highlights
c) Manni01 algo of today, scaled to the entered peak nits of your display.
@Manni01 has 100nits, so if you specify 100nits, you will get exactly his values.
If you put: real display target nits 50nits, then all the target nits you will get will be exactly half of the one used by Manni

Enjoy,

Anna&Flo
Thanks a million!
Soulnight likes this.
stef2 is offline  
post #4695 of 6031 Old 12-29-2018, 02:46 PM
AVS Forum Club Gold
 
Ian_Currie's Avatar
 
Join Date: Nov 2001
Location: Holliston, MA
Posts: 1,702
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 293 Post(s)
Liked: 88
Quote:
Originally Posted by Soulnight View Post
Hi guys

So, after Madshi Christmas gift, here is ours. (or the begining of it)
Any chance of some documentation/guide for how to use this?

Does it need to run on a PC with madVR & LAV installed?

JVC-RS4500 4k projector, Lumagen Pro, 138" 2.35:1 ST130
Panasonic DMP-UB820 & i7 PC w/1080ti running Win10/Kodi DSPlayer/madVR
Classe SSP-800, Bryston amplification, Wilson Audio speakers (7.2) + Buttkickers.
Ian_Currie is offline  
post #4696 of 6031 Old 12-29-2018, 02:47 PM
Senior Member
 
chros73's Avatar
 
Join Date: Jan 2015
Posts: 400
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 227 Post(s)
Liked: 100
Quote:
Originally Posted by Soulnight View Post
New update of our madMeasure dynamic clipping tool in V2.5.
Thank You for the update!

Unfortunately something fishy is going on:
- measure a sample with madmeasure.exe
- open it up in @padnm 's tool: main values look correct (maxFALL, avgFALL, avg avg/median highlights, etc.)
- run through your tool: apply dynamic clipping
- open the adjusted measurement file in @padnm 's tool again:
-- the mentioned values are absurd: +1000 , +2000

v2.4 worked fine in this regard.

Ryzen 5 2600,Asus Prime b450-Plus,16GB,MSI GTX 1060 Gaming X 6GB(v385.28),Win10 LTSB 1607,MPC-BEx64+LAV+MadVR,Yamaha RX-A870,LG OLED65B8(04.10.25+PC4:4:[email protected]/24/25/29/30/50/59/60Hz)
chros73 is offline  
post #4697 of 6031 Old 12-29-2018, 03:18 PM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 8,988
Mentioned: 307 Post(s)
Tagged: 0 Thread(s)
Quoted: 5343 Post(s)
Liked: 5375
@Soulnight

So I did a quick test with my usual example files. All the targets are identical, except Blade Runner 2049 (225nits instead of 200nits), which is fine.

Well done, this works great!

And thanks again @madshi for implementing this target nits field.

Here is my simplified algo now:

if (hdr) and (srcWidth <= 3840) and (srcHeight <= 2160) and (fps > 31) "BT2390 Mode 10 (Fast)"
else if (hdr) "HDR BT2390"
else "Send Metadata SDR"



[EDIT: I have a default target of 401nits in my single HDR BT2390 profile, to make sure that a different target displayed in the OSD comes from the measurements file.]

A few comments:

1) It looks like the utility still doesn't skip a file when an up-to-date measurement already exists. Did you run out of time to implement this?
2) Performance is still poor here, both from the MBP and the HTPC. It takes around 4'10'' for 22 files on the MBP, 6'10'' on the HTPC, so 12-18s per file.
3) Please could you use 107nits instead of 100nits as the baseline for my algo? That what I've used most and my actual nits target if I have more brightness on tap. As you know, that's the Dolby Cinema target for diffuse white and a realistic target for many users. It's great that we can adjust the algo so easily by indicating the actual peak brightness in the utility. Also you might want to specify in your notes that my algo is designed for 107nits peak brightness in a dedicated room (bat loft, 0nits ambient light), as the environment influences the values. Someone with 107nits actual peak but some ambient light will probably want to "lie" and specify a lower actual nits value. I'll try to add something about this in the post related to my algo. I'll add a link to your latest utility as well.
4) It might be worth adding Dexter's latest algo posted above. It's quite similar to mine but brighter for dimmer titles, and a bit more contrasty for The Meg. So it might suit some with a different environment and more brightness to start with. Its actual peak brightness is 212nits. I attach the measurements comparing my algo, Dexter's algo and the implementation in your utility (Manni01-S column) for my usual examples if anyone wants to take a look.

Keep up the good work! I'll try to do more visual tests with the new thresholds using 107nits actual peak brightness as a reference and I'll let you know if there are any adjustments to the algo.

Very curious to see which algo of your own you're going to come up with...

@madshi , do you have a new build coming up, and are there any changes to be expected for MadMeasureHDR? If there are no changes to MadMeasureHDR, I'll run the measurements for the rest of my titles over the next few days, otherwise I'll wait for the new build.
Attached Thumbnails
Click image for larger version

Name:	HDR Examples (with DK and Soulnight).JPG
Views:	45
Size:	273.1 KB
ID:	2502510  
CoryW and Soulnight like this.

JVC Autocal Software V11 Calibration for 2019 Models
Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders

Last edited by Manni01; 12-30-2018 at 10:47 PM.
Manni01 is online now  
post #4698 of 6031 Old 12-29-2018, 03:38 PM
Member
 
Join Date: Nov 2018
Posts: 172
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by Manni01 View Post
@Soulnight

So I did a quick test with my usual example files. All the targets are identical, except Blade Runner 2049 (225nits instead of 200nits), which is fine.

Well done, this works great!

And thanks again @madshi for implementing this target nits field.

Here is my simplified algo now:

if (hdr) and (srcWidth <= 3840) and (srcHeight <= 2160) and (fps > 31) "BT2390 Mode 10 (Fast)"
else if (hdr) "HDR BT2390"
else "Send Metadata SDR"



I have zero as a target in my single HDR BT2390 profile, to make sure that whatever non-zero target displayed in the OSD comes from the measurements file.

A few comments:

1) It looks like the utility still doesn't skip a file when an up-to-date measurement already exists. Did you run out of time to implement this?
2) Performance is still poor here, both from the MBP and the HTPC. It takes around 4'10'' for 22 files on the MBP, 6'10'' on the HTPC, so 12-18s per file.
3) Please could you use 107nits instead of 100nits as the baseline for my algo? That what I've used most and my actual nits target if I have more brightness on tap. As you know, that's the Dolby Cinema target for diffuse white and a realistic target for many users. It's great that we can adjust the algo so easily by indicating the actual peak brightness in the utility. Also you might want to specify in your notes that my algo is designed for 107nits peak brightness in a dedicated room (bat loft, 0nits ambient light), as the environment influences the values. Someone with 107nits actual peak but some ambient light will probably want to "lie" and specify a lower actual nits value. I'll try to add something about this in the post related to my algo. I'll add a link to your latest utility as well.
4) It might be worth adding Dexter's latest algo posted above. It's quite similar to mine but brighter for dimmer titles, and a bit more contrasty for The Meg. So it might suit some with a different environment and more brightness to start with. Its actual peak brightness is 212nits. I attach the measurements comparing my algo, Dexter's algo and the implementation in your utility (Manni01-S column) for my usual examples if anyone wants to take a look.

Keep up the good work! I'll try to do more visual tests with the new thresholds using 107nits actual peak brightness as a reference and I'll let you know if there are any adjustments to the algo.

Very curious to see which algo of your own you're going to come up with...

@madshi , do you have a new build coming up, and are there any changes to be expected for MadMeasureHDR? If there are no changes to MadMeasureHDR, I'll run the measurements for the rest of my titles over the next few days, otherwise I'll wait for the new build.
@Soulnight

If you do add my algo to your tool it's only the avgFALL values that would need to be adjusted to suit a given measured brightness, target nits should equal avgFMLL unless it hits the avgFALL threshold first. I'm not sure what the best way to adjust it to different measured brightness would be but I think it would work best to add or subtract the same amount from each avgFALL values.
Dexter Kane is offline  
post #4699 of 6031 Old 12-29-2018, 03:52 PM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,273
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 934 Post(s)
Liked: 1368
madmeasuredynamicclipping Tool V2.6

@chros73

ooops

So here attached a new version V2.6 of our madMeasure dynamic clipping tool.
It should be fixed.

@Manni01 : also changed your algo to consider 107nits to be the reference and changed associated text.





OUTDATED: please follow link to latest version in signature
Attached Files
File Type: zip madMeasureDynamicClipping V2.6 exe.zip (314.1 KB, 59 views)
File Type: zip madMeasureDynamicClipping Installer V2.6.zip (710.3 KB, 44 views)
CoryW, SamuriHL and Manni01 like this.

Last edited by Soulnight; 12-31-2018 at 02:42 AM.
Soulnight is offline  
post #4700 of 6031 Old 12-29-2018, 04:08 PM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,273
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 934 Post(s)
Liked: 1368
Quote:
Originally Posted by Manni01 View Post
@Soulnight

So I did a quick test with my usual example files. All the targets are identical, except Blade Runner 2049 (225nits instead of 200nits), which is fine.

Well done, this works great!

And thanks again @madshi for implementing this target nits field.

Here is my simplified algo now:

if (hdr) and (srcWidth <= 3840) and (srcHeight <= 2160) and (fps > 31) "BT2390 Mode 10 (Fast)"
else if (hdr) "HDR BT2390"
else "Send Metadata SDR"



I have zero as a target in my single HDR BT2390 profile, to make sure that whatever non-zero target displayed in the OSD comes from the measurements file.
Glad you like it.
We have looked at the reason for the slight target nits different for Blade Runner 2049 and we believe this is due to rounding of avgfmll. We use our own calculated avgfmll (before clipping) and we don't round it.
Since avgfmll is right on the edge for the title, the rounding makes the difference.

Quote:
Originally Posted by Manni01 View Post
@Soulnight
A few comments:

1) It looks like the utility still doesn't skip a file when an up-to-date measurement already exists. Did you run out of time to implement this?
2) Performance is still poor here, both from the MBP and the HTPC. It takes around 4'10'' for 22 files on the MBP, 6'10'' on the HTPC, so 12-18s per file.
3) Please could you use 107nits instead of 100nits as the baseline for my algo? That what I've used most and my actual nits target if I have more brightness on tap. As you know, that's the Dolby Cinema target for diffuse white and a realistic target for many users. It's great that we can adjust the algo so easily by indicating the actual peak brightness in the utility. Also you might want to specify in your notes that my algo is designed for 107nits peak brightness in a dedicated room (bat loft, 0nits ambient light), as the environment influences the values. Someone with 107nits actual peak but some ambient light will probably want to "lie" and specify a lower actual nits value. I'll try to add something about this in the post related to my algo. I'll add a link to your latest utility as well.
4) It might be worth adding Dexter's latest algo posted above. It's quite similar to mine but brighter for dimmer titles, and a bit more contrasty for The Meg. So it might suit some with a different environment and more brightness to start with. Its actual peak brightness is 212nits. I attach the measurements comparing my algo, Dexter's algo and the implementation in your utility (Manni01-S column) for my usual examples if anyone wants to take a look.

Keep up the good work! I'll try to do more visual tests with the new thresholds using 107nits actual peak brightness as a reference and I'll let you know if there are any adjustments to the algo.
1) No skip yet. It's on our to do list. And we have a big to do list. :-)
It will come, but we used up all our time available for things which had higher priority for us.

2) Ok.

3) Done in V2.6

4) Sure. It was on our to do list already

Quote:
Originally Posted by Manni01 View Post
@Soulnight
Very curious to see which algo of your own you're going to come up with...
So are we!
Time to sleep now.
Soulnight is offline  
post #4701 of 6031 Old 12-29-2018, 04:13 PM
Member
 
Join Date: Nov 2018
Posts: 172
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
Quote:
Originally Posted by chros73 View Post
Thank You for the update!

Unfortunately something fishy is going on:
- measure a sample with madmeasure.exe
- open it up in @padnm 's tool: main values look correct (maxFALL, avgFALL, avg avg/median highlights, etc.)
- run through your tool: apply dynamic clipping
- open the adjusted measurement file in @padnm 's tool again:
-- the mentioned values are absurd: +1000 , +2000

v2.4 worked fine in this regard.
I'm seeing something similar. Also the calculated avgFMLL after clipping that the tool displays is not what the measurement file ends up with. For example with Fury Road the original avgFMLL vlaue is 1601, the tool says that after clipping it is 723 but the measurement file has 1595
Dexter Kane is offline  
post #4702 of 6031 Old 12-29-2018, 04:17 PM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,273
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 934 Post(s)
Liked: 1368
Quote:
Originally Posted by Dexter Kane View Post
I'm seeing something similar. Also the calculated avgFMLL after clipping that the tool displays is not what the measurement file ends up with. For example with Fury Road the original avgFMLL vlaue is 1601, the tool says that after clipping it is 723 but the measurement file has 1595
As stated above. Should be fixed in V2.6.
Dexter Kane likes this.
Soulnight is offline  
post #4703 of 6031 Old 12-29-2018, 05:00 PM
Advanced Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 835
Mentioned: 134 Post(s)
Tagged: 0 Thread(s)
Quoted: 572 Post(s)
Liked: 685
@Soulnight Thank you for the update!

I realized I was aiming too low on the target for bright titles with my previous formula, for no particular reason.

So here is the new one I use:

Target = "avg avg of highlights" + "avg avg of highlights" * avgFALL / 100

For instance, for Planet Earth II E01: 179 + 179 * 55 / 100 = 277.45 nits

The Meg: 394 + 394 * 154 / 100 = 1000.76 nits

It does not change much for dark titles, but changes a lot for bright ones.
Neo-XP is online now  
post #4704 of 6031 Old 12-29-2018, 05:10 PM
Member
 
Join Date: Apr 2015
Posts: 95
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 82 Post(s)
Liked: 5
Hi guys ,
A strange situation here....
I've installed two times madvrhdrmeasure 38 suite just to be sure.... but still have problems.

1) look at the attached picture. Like you can see the highlights recovery functions, advanced blur settings, ecc. Are vanished!
Is it normal??

2) if I play a Mkv file with the measurement file in the same folder (file created with Soulnight tool v
2.4) madvr ignore it.

With old madvrhdrmeasure suite all was OK.

Any idea?

Thank you
FidelioX is offline  
post #4705 of 6031 Old 12-29-2018, 08:38 PM
Member
 
Join Date: Nov 2018
Posts: 172
Mentioned: 45 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 98
@Soulnight Have you changed the dynamic clipping at all since 2.4? It seems to be less aggressive than it was previously.

Here's the measurement files if you want to have a look: https://owncloud.aquinasnet.com/inde...XjUzIegyaGTzdg

Last edited by Dexter Kane; 12-30-2018 at 01:59 AM.
Dexter Kane is offline  
post #4706 of 6031 Old 12-29-2018, 08:52 PM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 7,837
Mentioned: 475 Post(s)
Tagged: 0 Thread(s)
Quoted: 6779 Post(s)
Liked: 6395
Quote:
Originally Posted by Manni01 View Post
It's in the movie. When you have a doubt, switch to passthrough, then you'll see what's in the content and what's done by MadVR tonemapping.

@madshi : in that scene from 05:57, the highlights in the headlights are badly destroyed (posterization/banding) by MadVR's tonemapping when the DI is enabled in my RS500, even in auto2 which limits gamma processing. If I switch the DI off, or if I use passthrough, the artifacts go away.

It's hard to see which MadVR option does this as the options are limited in the latest build. It's not "restore highlights", as I switched that to off and it didn't make a difference. I'm using a 425nits target with that film.

I attach a screenshot to show you were though it doesn't show the issue as it's really a weird interference with the gamma processing of the JVC DI.

But on screen it's so bad that I had to disable the DI for that scene/film, way too distracting. I don't know if you have a DI on your JVC but you might want to check that scene, it's very ugly.
You are describing exactly the issue I had with Batman vs Superman with the DI in use.

Ive seen it a number of times now with various content, once you know its going to be there on scenes like this, your eyes will be searching for it.

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 #4707 of 6031 Old 12-29-2018, 09:26 PM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 7,837
Mentioned: 475 Post(s)
Tagged: 0 Thread(s)
Quoted: 6779 Post(s)
Liked: 6395
Quote:
Originally Posted by Manni01 View Post
Here is my simplified algo now:

if (hdr) and (srcWidth <= 3840) and (srcHeight <= 2160) and (fps > 31) "BT2390 Mode 10 (Fast)"
else if (hdr) "HDR BT2390"
else "Send Metadata SDR"



I have zero as a target in my single HDR BT2390 profile, to make sure that whatever non-zero target displayed in the OSD comes from the measurements file.
Wow, been away from this thread a while, but looks a perfect time to jump back in

Manni, you say have zero entered in your lone BT 2390 profile Target Nits setting, but MadVR is not letting my type '0'... Are you definitely entering zero there? If so, how is it allowing you to apply the changes, just comes up red for me and wont let me save it.

if I typed 350 in there, will this use 350 for any file which does not have a measurement file, and if there is a measurement file, will it overrule the entry?

Wasnt sure if you just put 0 as a matter of testing, what happens if you play a file which is not measured?


EDIT - I think this quote from Madshi answers my question...

Quote:
c) The new header field is "TargetPeakNits" and is always set to 0 by madMeasureHDR. If you set this value to a specific number, madVR will use that number as the "target peak nits" setting for this one movie, instead of the "target peak nits" setting from your HDR settings page.
So yeah, seems having 350 as the target peak will apply to any film or file not measured... Easy enough. If I care about the file, I will measure it.

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; 12-29-2018 at 09:34 PM.
Javs is offline  
post #4708 of 6031 Old 12-29-2018, 10:47 PM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 7,837
Mentioned: 475 Post(s)
Tagged: 0 Thread(s)
Quoted: 6779 Post(s)
Liked: 6395
Quote:
Originally Posted by BerndFfm View Post
Peter : Update for You !

The Redirecting of the Output of Meadmeasure.exe to show the progress ist very critical. It is possible that my tool hangs up while reading the output.

I removed it, now you have no progress. I don't need it, I don't watch the pc while measuring movies.

Maybe it is working better now.

I just saw that my Task Manager is not working well (Windows 10). I have to close it and to open it again to see the madmeasurehdr.exe wörking.

NoTechi : yes, that in the movie.

Bernd
Hey Bernd,

Your tool is unfortunately hanging still for me after completing one file only.

Also, another bug, when I have the "Only Scan Incomplete" ticked:

Your tool has a bug where if I use custom measurement directory for eg C:/MadVR Measurements/ It will delete any file in that directory for eg E.!UHD Movies!3 10 to Yuma [2007].mkv.measurements and remeasure from the start, even though I have specified correct custom directory in your settings.

If I paste the measurement file in the same directory as the video file, then your software will correctly scan the next film and skip the file where a measurement file exists. but when its complete, it will still place measurement file in the same directory as the video file not the custom directory in the settings. Could you please test this and fix?

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; 12-29-2018 at 11:48 PM.
Javs is offline  
post #4709 of 6031 Old 12-30-2018, 12:12 AM
Member
 
Join Date: Mar 2018
Posts: 43
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 29 Post(s)
Liked: 13
HI!

Many thanks for all the recent updates and tools.

I solved all the problems with incomplete files and madvr is using them without problems. Wonderfull picture quality we have now!

I like the current situation a lot with my profiles rules and expect Soulnight arrives with an updated tool to define our own script/formula to assign TN to each file.

Currently I have 133 nits brightness and I try to use profiles using steps of 25 nits from 200 to 525 target nits and using avgfmll as profile selector.
I think with arves you could set a brightness factor taking in account your real brightness, so I thought of applying a factor of 4 here with the Target Nits, but not sure if this works the same way.

@Manni01 Manni What gamma value do you use in the calibration tab in madvr? 2.4 or 2.2 by default?

Regards!

Last edited by Axelpowa; 12-30-2018 at 02:02 AM.
Axelpowa is online now  
post #4710 of 6031 Old 12-30-2018, 02:12 AM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,273
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 934 Post(s)
Liked: 1368
Quote:
Originally Posted by FidelioX View Post
Hi guys ,
A strange situation here....
I've installed two times madvrhdrmeasure 38 suite just to be sure.... but still have problems.

1) look at the attached picture. Like you can see the highlights recovery functions, advanced blur settings, ecc. Are vanished!
Is it normal??

2) if I play a Mkv file with the measurement file in the same folder (file created with Soulnight tool v
2.4) madvr ignore it.

With old madvrhdrmeasure suite all was OK.

Any idea?

Thank you
Well madvrhdrmeasure 38 is a new "cleaned up" version. That's why you see less options. Madshi explained it very well a few post ago.
This thread needs to be followed with a lot of attention since this is not a support thread and things keeps changing quickly due to the beta nature of the project.

And as Madshi also explained, the V38 has a new measurement file format. This means it's not compatible with the old format.
You need to remeasure with the new version V38 and use V2.6 if you want dynamic clipping.
AmigoHD, Manni01, neo_2009 and 3 others like this.
Soulnight is offline  
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