New Lumagen Radiance Pro Series - Page 99 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 3119Likes
Reply
 
Thread Tools
post #2941 of 4877 Old 04-15-2018, 12:06 PM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,573
Mentioned: 458 Post(s)
Tagged: 0 Thread(s)
Quoted: 2338 Post(s)
Liked: 3040
@KarlKlammer , @Ian_Currie ,

if I may chime in here, although not currently being a Lumagen user:

My personal opinion matches Jim's (no surprise there). Basically I consider what the Arve custom curves do (and also the official JVC projector HDR implementation) really bad, from a color science point of view. I mean Arve did the best that was technically possible, but scientifically speaking, it's a bad solution. Here's a quick demonstration why:



These screenshots were made with a special test pattern I created which asks for colors which after YCbCr -> RGB conversion "overflow". Such colors should probably not be used in UHD Blu-Rays, but in reality they sometimes are. And what is worse: Tone mapping often turns perfectly valid colors into such "overflowing" colors. So although my test pattern is pretty mean, it's still valid to test for extreme situations.

What you can see in the test pattern above is how the test pattern should ideally be rendered (left side): All color bars should maintain their hue, just get brighter from left to right. On the right side you can see what the Arve type of color handling does to that test pattern: It totally destroys the hue! Basically red first turns into orange and then into yellow. Blue turns pink. And (you can't see that in this test pattern) green turns yellow, too.

Here's another image for you, this time real world material:



From left to right:

1) Arve/JVC color handling, tone mapped for a 4,000nits display.
2) Arve/JVC color handling, tone mapped for a 200nits display.
3) Proper color handling, tone mapped for a 4,000nits display.
4) Proper color handling, tone mapped for a 200nits display.

As you can see, Arve/JVC color handling renders a deep red for a 4,000nits display, but switches to bright yellow for a 200nits display. That obviously doesn't make any sense. Arve/JVC color handling isn't even consistent with itself. Depending on how bright your display is (LCD flat panel vs projector), it will either produce deep red or bright yellow. Proper color handling doesn't do that. Instead it renders the same hue, regardless of the display's peak luminance capabilities.

I have to admit, though, that the hue shifts produced by Arve/JVC color handling do make Mad Max explosions look more like what we're used to from other movies. But that is actually a lucky accident, nothing more! In this specific case the orange/yellow hue shift looks subjectively natural to our eyes. But in other cases such hue shifts can be very detrimental to image quality.

But instead of making Arve/JVC color handling look bad, we can also go the other way and showcase how much nicer proper tone mapping can look with real content. @Soulnight was kind enough to make some nice comparison images here, using the The World in HDR demo:

other AVSForum thread post with many comparison screenshots

The images titled "dynamic" were made with madVR's best quality tone mapping, which I expect should look similar to what the Lumagen does (?). The so-called "dumb" images are made with Arve/JVC type color handling. For comparison purposes I'd suggest that you copy the URLs of those images from the other thread and paste them into different browser tabs, to make quick back & forth switching possible, which will allow you to see the difference much better.

Contrast is a mostly separate issue to color handling, btw. Contrast will mostly depend on which tone mapping curve is used. Color handling instead decides whether there are any hue shifts, and how much bright + highly saturated colors are desaturated.
madshi is online now  
Sponsored Links
Advertisement
 
post #2942 of 4877 Old 04-16-2018, 01:11 AM
Senior Member
 
KarlKlammer's Avatar
 
Join Date: Apr 2010
Location: Germany
Posts: 299
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 250 Post(s)
Liked: 243
Thank you @madshi for your insights.

I wrote earlier, it is not my intention to endlessly complain about the Radiance Tone Mapping. And now that there is an explanation for the dumb enduser why colors are handled the way they are, I was already making my peace with it.
As you wrote, it is also a problem what we expect some things to look like based on past experience. And when I look at those correctly reproduced scenes in Mad Max and compare them to what is promised by the words "High" and "Dynamic" in the term "HDR" then it is not entirely satisfactory.

I spent some time again measuring the 1D LUT. And I found the error in my settings as why the image subjectively lacked contrast. Jim was right. The nit-for-nit-area of the curve is extremely important.
My settings for nit-for-nit were right at the beginning. That was until someone convinced me, that my settings must be wrong the way they were. I will never listen to people in forums again, when my measurements are OK.

So with my current settings I am able to follow the curve of my LightSpace reference up to 65%. That goes for 1000 and 4000 Nits. Some compromises had to be made for clipping. And I still wish there would be a little more control for the upper end.
Attached Thumbnails
Click image for larger version

Name:	IM_4000_20180413.JPG
Views:	65
Size:	60.2 KB
ID:	2391536   Click image for larger version

Name:	IM_1000_20180413.JPG
Views:	65
Size:	60.9 KB
ID:	2391538  
alex_t likes this.

Projection: JVC DLA-NX9
VP/Calibration: Lumagen Radiance Pro, LightSpace CMS, x-rite i1 Pro 2, x-rite i1 Display 3

Last edited by KarlKlammer; 04-16-2018 at 04:38 AM.
KarlKlammer is offline  
post #2943 of 4877 Old 04-16-2018, 03:49 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,573
Mentioned: 458 Post(s)
Tagged: 0 Thread(s)
Quoted: 2338 Post(s)
Liked: 3040
Quote:
Originally Posted by KarlKlammer View Post
I wrote earlier, it is not my intention to endlessly complain about the Radiance Tone Mapping. And now that there is an explenation for the dumb enduser why colors are handled the way they are, I was already making my peace with it.
Well, I'm often slightly annoyed when users complain about my algorithms/implementation, but in the end, it often does help improving things. So as long as there's a good reason for (constructive) complaining, it's a good thing to do. I'm sure my own tone mapping can be improved further, maybe Lumagen's can be improved, too, I really can't say. Probably the best approach is to compare various tone mapping approaches for difficult movie scenes. That might show if and where there's room for improvement.
mikela, Mike_WI and KarlKlammer like this.
madshi is online now  
Sponsored Links
Advertisement
 
post #2944 of 4877 Old 04-16-2018, 04:25 PM
jrp
AVS Forum Special Member
 
jrp's Avatar
 
Join Date: Jan 2003
Location: Portland, Oregon, USA
Posts: 1,462
Mentioned: 57 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 938
Quote:
Originally Posted by madshi View Post
@KarlKlammer , @Ian_Currie ,

if I may chime in here, although not currently being a Lumagen user:

My personal opinion matches Jim's (no surprise there). Basically I consider what the Arve custom curves do (and also the official JVC projector HDR implementation) really bad, from a color science point of view. I mean Arve did the best that was technically possible, but scientifically speaking, it's a bad solution. ...
madshi:

Thanks for posting the test pattern to show the color issue with using a 1D LUT for tone mapping. It really shows the issue you and I have been trying to get across and I appreciate you posting it here.

===

All:

One other thought I had on the Mad Max Fury Road sand storm scene (and similar scenes in other movies), and that madshi commented on by saying "what we are used to," is that I initially considered that the Pro HDR IM might have an issue with the Fury Road truck explosion since when Kris and I looked at the SDR version the explosion had more yellow.

However, after looking at the HDR data, I concluded that the explosion is more yellow in the SDR version because the colorist wanted more intensity (achieved by increasing green which made it both brighter and more yellow), and so gave on getting the correct color. However, this compromise is not needed for HDR since the explosion is well within the maximum nit range possible with HDR. So the intended color in the HDR version are much more burnt orange and less yellow.
alex_t likes this.

Jim Peterson
Lumagen
jrp is offline  
post #2945 of 4877 Old 04-16-2018, 05:13 PM
jrp
AVS Forum Special Member
 
jrp's Avatar
 
Join Date: Jan 2003
Location: Portland, Oregon, USA
Posts: 1,462
Mentioned: 57 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 938
Quote:
Originally Posted by KarlKlammer View Post
Thank you @madshi for your insights.

I wrote earlier, it is not my intention to endlessly complain about the Radiance Tone Mapping. And now that there is an explanation for the dumb enduser why colors are handled the way they are, I was already making my peace with it.
As you wrote, it is also a problem what we expect some things to look like based on past experience. And when I look at those correctly reproduced scenes in Mad Max and compare them to what is promised by the words "High" and "Dynamic" in the term "HDR" then it is not entirely satisfactory.

I spent some time again measuring the 1D LUT. And I found the error in my settings as why the image subjectively lacked contrast. Jim was right. The nit-for-nit-area of the curve is extremely important.
My settings for nit-for-nit were right at the beginning. That was until someone convinced me, that my settings must be wrong the way they were. I will never listen to people in forums again, when my measurements are OK.

So with my current settings I am able to follow the curve of my LightSpace reference up to 65%. That goes for 1000 and 4000 Nits. Some compromises had to be made for clipping. And I still wish there would be a little more control for the upper end.
Thanks for the update. I am glad you have been able to find better HDR IM settings.

The range for the upper end is controlled by the Tran (transition) parameter and the Shape parameter. On my LED TV I can get the sand storm explosion to look very good (or "flat" if I give too much to the nit-for-bit range). With a projector you have less room for the upper end and if you give enough nit-for-nit you will lose detail in the upper end. So with a projector if you increase the Tran parameter you will limited the upper end and you will lose details in the upper end. There is no way to "improve the upper end" when there is no upper end range left due to the setting of Tran, Shape and Display Max Light.

There is something you can do to compromise favoring the upper end more. If you select a larger Display Max Light value, the nit-for-nit range becomes a "linear-fraction" of the nit-for nit range. So you will still have linear response like the nit-for-nit is intended to have in the lower values (but they will be somewhat darker). This way you will have more room left at the upper end detail. Then adjust the HDR parameters with the new Display Max value to again optimize the low, mid and upper range.

There has to be a trade-off and when you give most of the range to nit-for-nit, you cannot have as much range for the upper end and you will have a more "flat" response in the upper end. This is simply how the math works. However, the above may help you get a trade-off that is more along the lines of your goal of more definition at the upper end while preserving the desired linear response in the low end.
JFR0317, alex_t, Wookii and 2 others like this.

Jim Peterson
Lumagen
jrp is offline  
post #2946 of 4877 Old 04-16-2018, 10:09 PM
Senior Member
 
KarlKlammer's Avatar
 
Join Date: Apr 2010
Location: Germany
Posts: 299
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 250 Post(s)
Liked: 243
Quote:
Originally Posted by jrp View Post
Thanks for the update. I am glad you have been able to find better HDR IM settings.

The range for the upper end is controlled by the Tran (transition) parameter and the Shape parameter. On my LED TV I can get the sand storm explosion to look very good (or "flat" if I give too much to the nit-for-bit range). With a projector you have less room for the upper end and if you give enough nit-for-nit you will lose detail in the upper end. So with a projector if you increase the Tran parameter you will limited the upper end and you will lose details in the upper end. There is no way to "improve the upper end" when there is no upper end range left due to the setting of Tran, Shape and Display Max Light.

There is something you can do to compromise favoring the upper end more. If you select a larger Display Max Light value, the nit-for-nit range becomes a "linear-fraction" of the nit-for nit range. So you will still have linear response like the nit-for-nit is intended to have in the lower values (but they will be somewhat darker). This way you will have more room left at the upper end detail. Then adjust the HDR parameters with the new Display Max value to again optimize the low, mid and upper range.

There has to be a trade-off and when you give most of the range to nit-for-nit, you cannot have as much range for the upper end and you will have a more "flat" response in the upper end. This is simply how the math works. However, the above may help you get a trade-off that is more along the lines of your goal of more definition at the upper end while preserving the desired linear response in the low end.
For the 1D chart to look like above, I increased Max Display Light from 500 to 700. Transition is at 15 and Shape at 3/4. With Max Display Light at 500, Tranistion was at 12/13 and Shape the same as now. So it seems setting Max Display Light higher requires Transition to be higher. And with Transition set to 15, I have reached the end.

Just as an information
madshi has written this about some screenshots I posted in the other thread
Quote:
Originally Posted by madshi View Post
Thank you for the screenshots! Ok, so we can conclude that:

1) Lumagen's IM does not desaturate at all. It's the same approach which madVR uses if you select the option "100% luminance reduction, 0% saturation reduction". IMHO that's not a good choice, to be honest. As you can see, it produces highlight detail loss in some scenes. Ok, these scenes are extreme cases, but they do occur.

2) LightSpace seems to use "dumb/naive" color handling, which is surprising to me.

Projection: JVC DLA-NX9
VP/Calibration: Lumagen Radiance Pro, LightSpace CMS, x-rite i1 Pro 2, x-rite i1 Display 3

Last edited by KarlKlammer; 04-16-2018 at 10:40 PM.
KarlKlammer is offline  
post #2947 of 4877 Old 04-17-2018, 01:21 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,573
Mentioned: 458 Post(s)
Tagged: 0 Thread(s)
Quoted: 2338 Post(s)
Liked: 3040
Quote:
Originally Posted by jrp View Post
However, after looking at the HDR data, I concluded that the explosion is more yellow in the SDR version because the colorist wanted more intensity (achieved by increasing green which made it both brighter and more yellow), and so gave on getting the correct color.
It's possible that he did that, or alternatively I could also imagine that the software the colorist used for tone mapping HDR -> SDR used the same naive arve/JVC color handling approach.

(I'll send some more comments to you via PM.)
madshi is online now  
post #2948 of 4877 Old 04-17-2018, 04:10 AM
Advanced Member
 
Roland Janus's Avatar
 
Join Date: Mar 2005
Location: Switzerland
Posts: 535
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 131 Post(s)
Liked: 33
Quote:
Originally Posted by Roland Janus View Post
anything planned on the auto-aspect feature?

I usually use HDMI+image but the credits often trigger the wrong aspect.

Also any plans of switching much faster, it would be very useful in particular for Nolan movies (I hate that).

How about a customizable value in seconds on how fast it would detect changes?

-roland
Jim, any feedback on that?
Dirk44 likes this.
Roland Janus is online now  
post #2949 of 4877 Old 04-17-2018, 10:59 AM
jrp
AVS Forum Special Member
 
jrp's Avatar
 
Join Date: Jan 2003
Location: Portland, Oregon, USA
Posts: 1,462
Mentioned: 57 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 938
We will be changing the HDR IM parameter defaults in the next release. These changes will affect only the defaults when the parameters are reset (in the HDR IM parameters menu, or a factory reset, plus a Save). We need to thank Kris Deering for his work testing the various permutations of Display Max Light, Tran and Shape to discover the optimal settings.

Currently the Display Max Light default is 500, and for the MaxCLL > 2500 parameter set, Shape = 3 and Tran = 10. Kris discovered, and we have confirmed, for a display that has excellent HDR quality using these defaults, that if one reduces Display Max Light to 400, and increases Shape to 5 and Tran to 11, the results are even better. Note, you may, or may not, want to reduce the Display Max Light. As always try different settings to determine the best one for the > 2500 nit source parameter set, than adjust Ratio as needed to modify the effective Display Max Light for the <= 2500 nit source parameter set.

Hint: If you want to change Display Max Light in finer than 100 nit steps, use the digit keys to enter the value you want in multiples of 10 nits.

The new defaults for HDR IM parameters for MaxCLL > 2500 nits are Tran = 11 and Shape = 5, and for MaxCLL <= 2500 nits are Tran = 7 and Shape = 3.
Mike_WI likes this.

Jim Peterson
Lumagen
jrp is offline  
post #2950 of 4877 Old 04-17-2018, 12:19 PM
AVS Forum Addicted Member
 
Kris Deering's Avatar
 
Join Date: Oct 2002
Location: The Pacific Northwet
Posts: 10,604
Mentioned: 186 Post(s)
Tagged: 0 Thread(s)
Quoted: 3750 Post(s)
Liked: 6400
Quote:
Originally Posted by jrp View Post
We will be changing the HDR IM parameter defaults in the next release. These changes will affect only the defaults when the parameters are reset (in the HDR IM parameters menu, or a factory reset, plus a Save). We need to thank Kris Deering for his work testing the various permutations of Display Max Light, Tran and Shape to discover the optimal settings.

Currently the Display Max Light default is 500, and for the MaxCLL > 2500 parameter set, Shape = 3 and Tran = 10. Kris discovered, and we have confirmed, for a display that has excellent HDR quality using these defaults, that if one reduces Display Max Light to 400, and increases Shape to 5 and Tran to 11, the results are even better. Note, you may, or may not, want to reduce the Display Max Light. As always try different settings to determine the best one for the > 2500 nit source parameter set, than adjust Ratio as needed to modify the effective Display Max Light for the <= 2500 nit source parameter set.

Hint: If you want to change Display Max Light in finer than 100 nit steps, use the digit keys to enter the value you want in multiples of 10 nits.

The new defaults for HDR IM parameters for MaxCLL > 2500 nits are Tran = 11 and Shape = 5, and for MaxCLL <= 2500 nits are Tran = 7 and Shape = 3.
Glad I could help with this. I have been doing continuous testing with as many "difficult" scenes as I can to generate numbers. You have to be very careful with the T/S settings as they change quite a bit depending on display max. Otherwise you get an overly processed look.


For example, yesterday I played around with setting display max to my actual display light value. This looks great with A LOT of content but it caused issues with others. For example, it created banding artifacts in Mad Max (or emphasized existing banding that is normally not seen) and it also made the "Sand storm" sequence look garish and overly red. I found that using any setting of display max less than about 300 resulted in visible banding with Mad Max (though other titles looked fantastic). So right now I am using 300, which required me to tweak my other numbers. I will post them tonight so folks that want to try them can.


The numbers Jim has above are when I was setting display max at 500. I also set contrast to 2, but I could easily live with 0 as well for that setting.
LJG and Mike_WI like this.

My Home Theater UPDATED DEC 2017
Technical Editor/Writer Sound and Vision Magazine
Deep Dive AV - Calibration, Consulting and Education
Kris Deering is offline  
post #2951 of 4877 Old 04-17-2018, 05:04 PM
jrp
AVS Forum Special Member
 
jrp's Avatar
 
Join Date: Jan 2003
Location: Portland, Oregon, USA
Posts: 1,462
Mentioned: 57 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 938
Quote:
Originally Posted by Roland Janus View Post
anything planned on the auto-aspect feature?

I usually use HDMI+image but the credits often trigger the wrong aspect.

Also any plans of switching much faster, it would be very useful in particular for Nolan movies (I hate that).

How about a customizable value in seconds on how fast it would detect changes?

-roland
Currently image-based auto-aspect work is not on the to-do list. Does not mean it cannot rise enough in priority to get on the list, but we would need feedback from a number of people on this being more important than other things already on the to-do list.

A programmable time to "make sure" of an aspect ratio change is a reasonable request. I will mention it to Patrick. If it turns out to be simple we might put it in to a release in the near future. If not, probably not getting done anytime soon.

===

If you send an email to us at [email protected] detailing what you mean by "credits often trigger the wrong aspect" we can at least understand what you are seeing and maybe it is something simple to fix. What we need to know is are credits that should be 16:9 detected as 2.35, or are credits that should be 2.35 detected as 16:9, or what is happening. It would be helpful to know what movie(s) you see this on and if possible the timestamp of the credits that can easily demonstrate the issue. If there are additional details you can give it would be helpful.

On this issue it seems likely we can reproduce it without any information but it is always quicker to have as much info to start as possible. And having someone take the time to get the info to us might just bump the priority up.
Roland Janus, Dirk44 and Mike_WI like this.

Jim Peterson
Lumagen
jrp is offline  
post #2952 of 4877 Old 04-18-2018, 02:00 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 Kris Deering View Post

The numbers Jim has above are when I was setting display max at 500. I also set contrast to 2, but I could easily live with 0 as well for that setting.
Did you mean to say Gamma?

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 #2953 of 4877 Old 04-18-2018, 03:14 PM
AVS Forum Addicted Member
 
Kris Deering's Avatar
 
Join Date: Oct 2002
Location: The Pacific Northwet
Posts: 10,604
Mentioned: 186 Post(s)
Tagged: 0 Thread(s)
Quoted: 3750 Post(s)
Liked: 6400
Quote:
Originally Posted by Ian_Currie View Post
Did you mean to say Gamma?
I did, sorry!

My Home Theater UPDATED DEC 2017
Technical Editor/Writer Sound and Vision Magazine
Deep Dive AV - Calibration, Consulting and Education
Kris Deering is offline  
post #2954 of 4877 Old 04-18-2018, 10:57 PM
Member
 
wfdTamar's Avatar
 
Join Date: Jun 2012
Location: Tasmania, Australia
Posts: 23
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 14
After installing my Radiance Pro 4242 in my setup, I'm finding on some (or just one) source devices, possibly only sometimes, that I have audio/video sync issues. I find it mostly with my PVR, but it's hard to pin down. I didn't notice a problem before the Radiance.

I think what I need are some video files that are the equivalent of a clapper board so I can see how I need to adjust the delay (although I think you can only go one way on the Radiance). Does anyone know of any, or what I should be googling for?

My chain of devices is PVR (Beyonwiz T4) and other source devices (Bluray, streaming boxes) - Radiance 4242 - video to TV, audio to Bryston SP3 processor.

Panasonic EZ1000 65" OLED TV
Lumagen Radiance Pro 4242 (with X-Rite Eyeone & Chromapure)
Oppo 203, Roku, Beyonwiz T4 PVR
Bryston SP3, Emotiva XPA5, Wharfedale Diamond 9 5.1
wfdTamar is offline  
post #2955 of 4877 Old 04-19-2018, 01:07 AM
jrp
AVS Forum Special Member
 
jrp's Avatar
 
Join Date: Jan 2003
Location: Portland, Oregon, USA
Posts: 1,462
Mentioned: 57 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 938
Quote:
Originally Posted by wfdTamar View Post
After installing my Radiance Pro 4242 in my setup, I'm finding on some (or just one) source devices, possibly only sometimes, that I have audio/video sync issues. I find it mostly with my PVR, but it's hard to pin down. I didn't notice a problem before the Radiance.

I think what I need are some video files that are the equivalent of a clapper board so I can see how I need to adjust the delay (although I think you can only go one way on the Radiance). Does anyone know of any, or what I should be googling for?

My chain of devices is PVR (Beyonwiz T4) and other source devices (Bluray, streaming boxes) - Radiance 4242 - video to TV, audio to Bryston SP3 processor.
Please contact us at [email protected] or call 503-574-2211 Option 2 for support.

Note that drop outs and sync issues generally have to be resolved by changing HDMI cables.

Jim Peterson
Lumagen
jrp is offline  
post #2956 of 4877 Old 04-19-2018, 01:58 AM
Member
 
wfdTamar's Avatar
 
Join Date: Jun 2012
Location: Tasmania, Australia
Posts: 23
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 14
Ok just emailed support. I'm using reasonable new HDMI cables from here:

https://www.monoprice.com/product?p_id=15427

Panasonic EZ1000 65" OLED TV
Lumagen Radiance Pro 4242 (with X-Rite Eyeone & Chromapure)
Oppo 203, Roku, Beyonwiz T4 PVR
Bryston SP3, Emotiva XPA5, Wharfedale Diamond 9 5.1
wfdTamar is offline  
post #2957 of 4877 Old 04-19-2018, 07:15 AM
aka jfinnie
 
Join Date: Aug 2015
Location: Norwich, UK
Posts: 2,871
Mentioned: 44 Post(s)
Tagged: 0 Thread(s)
Quoted: 2328 Post(s)
Liked: 1480
Quote:
Originally Posted by wfdTamar View Post
Ok just emailed support. I'm using reasonable new HDMI cables from here:

https://www.monoprice.com/product?p_id=15427
I guess by your post about a clapper what you actually mean is that you have some delay between video and audio, not that the audio and/or video are breaking up (I can't see how cables would be related to audio delay)
bobof is online now  
post #2958 of 4877 Old 04-19-2018, 08:19 AM
Member
 
Join Date: Jan 2016
Location: UK
Posts: 21
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 15 Post(s)
Liked: 7
Hi,
Just seen this quote below on UK forums, is this remotely accurate, or just a case of the OP having a not-so hidden agenda (they seem to be a big Madvr fan).....


I have no agenda just pointing out the limitations like others do on this forum.
Mad Max Fury Road is 1 of many that will clip if using Lumagens Pro solution.

A quick search finds the below films exceed Lumagens Pro Intensity Mapping - Max Content Light Level (MaxCLL) threshold so scenes will clip like the above sample pic, released almost a year ago for HDR content Intensity Mapping and only now getting an update.......soon............

Inferno
The Da Vinci Code
Ghostbusters 2016
Ghostbusters
Ghostbusters II
Life
The Magnificent Seven
Crouching Tiger Hidden Dragon
Mad Max Fury Road
Starship Troopers
Chappie
Kong Skull Island
Interstella
Underworld
Underworld Blood Wars
Dunkirk
Blade Runner 2049
mickb1965 is offline  
post #2959 of 4877 Old 04-19-2018, 09:54 AM
AVS Forum Addicted Member
 
Kris Deering's Avatar
 
Join Date: Oct 2002
Location: The Pacific Northwet
Posts: 10,604
Mentioned: 186 Post(s)
Tagged: 0 Thread(s)
Quoted: 3750 Post(s)
Liked: 6400
Quote:
Originally Posted by mickb1965 View Post
Hi,
Just seen this quote below on UK forums, is this remotely accurate, or just a case of the OP having a not-so hidden agenda (they seem to be a big Madvr fan).....


I have no agenda just pointing out the limitations like others do on this forum.
Mad Max Fury Road is 1 of many that will clip if using Lumagens Pro solution.

A quick search finds the below films exceed Lumagens Pro Intensity Mapping - Max Content Light Level (MaxCLL) threshold so scenes will clip like the above sample pic, released almost a year ago for HDR content Intensity Mapping and only now getting an update.......soon............

Inferno
The Da Vinci Code
Ghostbusters 2016
Ghostbusters
Ghostbusters II
Life
The Magnificent Seven
Crouching Tiger Hidden Dragon
Mad Max Fury Road
Starship Troopers
Chappie
Kong Skull Island
Interstella
Underworld
Underworld Blood Wars
Dunkirk
Blade Runner 2049

There is no clipping on the Pro. It is actually the other solution (Arve) that IS clipping, the Radiance is actually displaying the colors based on the pixel information correctly, it just doesn't look as good subjectively. This is just another case of people chiming in on something they don't understand.

My Home Theater UPDATED DEC 2017
Technical Editor/Writer Sound and Vision Magazine
Deep Dive AV - Calibration, Consulting and Education
Kris Deering is offline  
post #2960 of 4877 Old 04-19-2018, 10:17 AM
jrp
AVS Forum Special Member
 
jrp's Avatar
 
Join Date: Jan 2003
Location: Portland, Oregon, USA
Posts: 1,462
Mentioned: 57 Post(s)
Tagged: 0 Thread(s)
Quoted: 374 Post(s)
Liked: 938
To add to Kris' comment, here is a portion of information I sent to Gordon in the UK to address the inaccurate posts going around:

The scenes are NOT clipping in the Pro. I checked. The color rendered by the Radiance Pro is the color on the disc. In fact the Pro looks as it does because it does NOT clip the colors and other Tone Mapping apparently is clipping or at least desaturating the colors.

The Radiance Pro's response is mathematically accurate. However, some experienced people are suggesting the Pro desaturate the colors for these extreme bright saturated colors. So, we are looking at desaturating the brighter colors some since some people seem to prefer this even though this will actually make the color less accurate as compared to what is on the disc.

I would also like to note these scenes are one, or at most a few, frames. For example the Mad Max Fury Road Lightening scene (28:53), the bright lightning involved in this discussion is a single frame (plus one or two frames with significantly reduced brightness not affected by this discussion). You would have to pause on the scene to notice the differences if we choose to desaturate the color. The flames on the car are shown for a few flames but again the Pro is showing the colors as they are on the disc.

It is also commonly discussed that early HDR titles like Mad Max Fury Road actually have inappropriate choices for these brightly saturated colors. And that as the colorists have more experience editing HDR the colors on the disc will be desaturated in editing so that they do not have to be desaturated by Tone Mapping. This is subject to debate, but the big issue with having over saturated color on the disc is that if Tone Mapping is changed to render the over saturated colors as some people want, rather than what is on the disc, when a colorist truly wants to exploit the 2020 color Gamut and produce highly saturated colors it will not be possible because the Tone Mapping algorithms will desaturate the colors. So this discussion is really about do we change the Tone Mapping algorithm to NOT be accurate for these early titles at the expense of future titles that truly need the highly saturated colors?

We will be looking at ways to make the Radiance Pro Tone Mapping even better and we may well choose to put in some desaturation. However, it is my strongly held opinion that Tone Mapping should accurately portray the color on the disc/video-stream (as I believe the Pro does) and that it is up to the colorist to desaturate colors if they would otherwise be over saturated from what a viewer would like to see. The colorist has knowledge of what he/she wants to see, we have only the data on the disc and math.
Mark_H, SJHT, Kris Deering and 3 others like this.

Jim Peterson
Lumagen
jrp is offline  
post #2961 of 4877 Old 04-19-2018, 10:53 AM
AVS Forum Addicted Member
 
Kris Deering's Avatar
 
Join Date: Oct 2002
Location: The Pacific Northwet
Posts: 10,604
Mentioned: 186 Post(s)
Tagged: 0 Thread(s)
Quoted: 3750 Post(s)
Liked: 6400
Quote:
Originally Posted by jrp View Post
We will be looking at ways to make the Radiance Pro Tone Mapping even better and we may well choose to put in some desaturation. However, it is my strongly held opinion that Tone Mapping should accurately portray the color on the disc/video-stream (as I believe the Pro does) and that it is up to the colorist to desaturate colors if they would otherwise be over saturated from what a viewer would like to see. The colorist has knowledge of what he/she wants to see, we have only the data on the disc and math.
To say that Jim and I have had long conversations about all this is putting it lightly. This is really difficult to tackle from any perspective because the information you're dealing with may be faulty in the first place.


For example, Mad Max has LOTS of content that exceeds the display max. So with that, you have to try and figure out what happened with the mastering. According to the people I've talked to about it, the display would just clip any information above its max (in this case the Pulsar, which has no tone mapping). So by that rationale, would it be better to just clip anything that exceeds the display max in the metadata because that is what the mastering monitor did? But if it truly does have content that is above the mastering monitor, why do some other HDR encodes of the SAME movie NOT exceed the display max? I have a hard time believing that they did multiple HDR grades of the same film at this point in the game. So was it a bug in the metadata or the encode side? I'd love to say that the studios would be studious in this but we've seen exactly the opposite. The metadata for most titles is either not there or completely out to lunch, and that isn't even addressing that the metadata that they chose to provide was a poor choice to begin with in its implementation.


For the dimming, do you make this kind of choice for what amounts to a few seconds of issues at the risk of compromising other titles without issues? Like Jim said, just about every example that is being scrutinized here are literally a couple frames AT BEST and would generally be unnoticeable unless you paused on that exact frame and had something to compare it to that simply looks different.


I do not envy Jim, Patrick or Madshi on this stuff at all.
Mike_WI likes this.

My Home Theater UPDATED DEC 2017
Technical Editor/Writer Sound and Vision Magazine
Deep Dive AV - Calibration, Consulting and Education
Kris Deering is offline  
post #2962 of 4877 Old 04-19-2018, 12:20 PM
LJG
AVS Forum Special Member
 
LJG's Avatar
 
Join Date: Jun 2000
Location: Brookville, NY
Posts: 4,958
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 701 Post(s)
Liked: 410
Quote:
Originally Posted by Kris Deering View Post
Glad I could help with this. I have been doing continuous testing with as many "difficult" scenes as I can to generate numbers. You have to be very careful with the T/S settings as they change quite a bit depending on display max. Otherwise you get an overly processed look.


For example, yesterday I played around with setting display max to my actual display light value. This looks great with A LOT of content but it caused issues with others. For example, it created banding artifacts in Mad Max (or emphasized existing banding that is normally not seen) and it also made the "Sand storm" sequence look garish and overly red. I found that using any setting of display max less than about 300 resulted in visible banding with Mad Max (though other titles looked fantastic). So right now I am using 300, which required me to tweak my other numbers. I will post them tonight so folks that want to try them can.


The numbers Jim has above are when I was setting display max at 500. I also set contrast to 2, but I could easily live with 0 as well for that setting.
Kris:

Id like to try your numbers for 300 also when you can post. The 500 numbers looked really good for Hell or High Water 4K UHD disc.
LJG is offline  
post #2963 of 4877 Old 04-19-2018, 12:23 PM
AVS Forum Addicted Member
 
Kris Deering's Avatar
 
Join Date: Oct 2002
Location: The Pacific Northwet
Posts: 10,604
Mentioned: 186 Post(s)
Tagged: 0 Thread(s)
Quoted: 3750 Post(s)
Liked: 6400
Quote:
Originally Posted by LJG View Post
Kris:

Id like to try your numbers for 300 also when you can post. The 500 numbers looked really good for Hell or High Water 4K UHD disc.
I believe they are similar TBH, but I can check when I get home and post.
LJG likes this.

My Home Theater UPDATED DEC 2017
Technical Editor/Writer Sound and Vision Magazine
Deep Dive AV - Calibration, Consulting and Education
Kris Deering is offline  
post #2964 of 4877 Old 04-19-2018, 01:19 PM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,573
Mentioned: 458 Post(s)
Tagged: 0 Thread(s)
Quoted: 2338 Post(s)
Liked: 3040
Quote:
Originally Posted by jrp View Post
The Radiance Pro's response is mathematically accurate. However, some experienced people are suggesting the Pro desaturate the colors for these extreme bright saturated colors. So, we are looking at desaturating the brighter colors some since some people seem to prefer this even though this will actually make the color less accurate as compared to what is on the disc.

It is also commonly discussed that early HDR titles like Mad Max Fury Road actually have inappropriate choices for these brightly saturated colors. And that as the colorists have more experience editing HDR the colors on the disc will be desaturated in editing so that they do not have to be desaturated by Tone Mapping. This is subject to debate, but the big issue with having over saturated color on the disc is that if Tone Mapping is changed to render the over saturated colors as some people want, rather than what is on the disc, when a colorist truly wants to exploit the 2020 color Gamut and produce highly saturated colors it will not be possible because the Tone Mapping algorithms will desaturate the colors. So this discussion is really about do we change the Tone Mapping algorithm to NOT be accurate for these early titles at the expense of future titles that truly need the highly saturated colors?

We will be looking at ways to make the Radiance Pro Tone Mapping even better and we may well choose to put in some desaturation. However, it is my strongly held opinion that Tone Mapping should accurately portray the color on the disc/video-stream (as I believe the Pro does) and that it is up to the colorist to desaturate colors if they would otherwise be over saturated from what a viewer would like to see. The colorist has knowledge of what he/she wants to see, we have only the data on the disc and math.
I fully agree with your sentiment that we should try our best to reproduce the content as near to the encoding as possible - warts and all! However, please double check that you honor the Y channel just as much as the CbCr channels. I did not suggest desaturation to hide Mad Max encoding bugs. The one and only reason why I've suggested that is because it's the only possible way to reproduce the Y channel correctly in extreme (and rare) cases like some of the Mad Max frames.

(Did you receive my email? I've gone into a lot more detail in the email. If you haven't received it, please check your spam folder, thanks.)
madshi is online now  
post #2965 of 4877 Old 04-19-2018, 02:26 PM
Member
 
wfdTamar's Avatar
 
Join Date: Jun 2012
Location: Tasmania, Australia
Posts: 23
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 14
Quote:
Originally Posted by bobof View Post
I guess by your post about a clapper what you actually mean is that you have some delay between video and audio, not that the audio and/or video are breaking up (I can't see how cables would be related to audio delay)
Yes, hence the phrase 'audio/video sync'. I've seen on device that showed a clip an animation of a ball slowly bouncing on a floor. It makes a sound when it makes contact with the floor. I need something like that to play on the PVR so I can work out if adding delay to the Radiance input works. It's too hard just using a regular clip of people talking.

Panasonic EZ1000 65" OLED TV
Lumagen Radiance Pro 4242 (with X-Rite Eyeone & Chromapure)
Oppo 203, Roku, Beyonwiz T4 PVR
Bryston SP3, Emotiva XPA5, Wharfedale Diamond 9 5.1
wfdTamar is offline  
post #2966 of 4877 Old 04-19-2018, 02:56 PM
Senior Member
 
KarlKlammer's Avatar
 
Join Date: Apr 2010
Location: Germany
Posts: 299
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 250 Post(s)
Liked: 243
A quick search via google found this
https://archive.org/details/BBCHDAudioSyncTest
I don't know if this clip is OK. But you could give it a try.

Projection: JVC DLA-NX9
VP/Calibration: Lumagen Radiance Pro, LightSpace CMS, x-rite i1 Pro 2, x-rite i1 Display 3
KarlKlammer is offline  
post #2967 of 4877 Old 04-19-2018, 03:18 PM
AVS Forum Club Gold
 
appelz's Avatar
 
Join Date: May 2004
Location: A Hilton property near you!
Posts: 767
Mentioned: 26 Post(s)
Tagged: 0 Thread(s)
Quoted: 563 Post(s)
Liked: 562
Quote:
Originally Posted by KarlKlammer View Post
A quick search via google found this
https://archive.org/details/BBCHDAudioSyncTest
I don't know if this clip is OK. But you could give it a try.
YouTube has quite a few that you can stream from an AppleTV or Roku.

"Audio Video Lip Sync Test" turns up quite a few.

Adam Pelz ,Acoustic Mafia - Hear No Evil
JBL Master ARCOS Calibrator, CEDIA Designer, Home Acoustics Alliance Instructor LIII, THX HT1+ HT2+ Video, Level III Trinnov Altitude Calibrator
Mercenary Calibrator for Manufacturers, Integrators and System Owners
appelz is online now  
post #2968 of 4877 Old 04-19-2018, 03:25 PM
Member
 
wfdTamar's Avatar
 
Join Date: Jun 2012
Location: Tasmania, Australia
Posts: 23
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 14
Thanks. I tried googling, but my search terms were off and I didn't find anything. :-)

Yes that BBC one is good, just a bit fast.

Panasonic EZ1000 65" OLED TV
Lumagen Radiance Pro 4242 (with X-Rite Eyeone & Chromapure)
Oppo 203, Roku, Beyonwiz T4 PVR
Bryston SP3, Emotiva XPA5, Wharfedale Diamond 9 5.1
wfdTamar is offline  
post #2969 of 4877 Old 04-19-2018, 05:51 PM
AVS Forum Addicted Member
 
Kris Deering's Avatar
 
Join Date: Oct 2002
Location: The Pacific Northwet
Posts: 10,604
Mentioned: 186 Post(s)
Tagged: 0 Thread(s)
Quoted: 3750 Post(s)
Liked: 6400
Quote:
Originally Posted by LJG View Post
Kris:



Id like to try your numbers for 300 also when you can post. The 500 numbers looked really good for Hell or High Water 4K UHD disc.


> 2000 shape 4 transition 11 gamma 2
< 2000 shape 3 transition 7 gamma 2


Sent from my iPhone using Tapatalk Pro

My Home Theater UPDATED DEC 2017
Technical Editor/Writer Sound and Vision Magazine
Deep Dive AV - Calibration, Consulting and Education
Kris Deering is offline  
post #2970 of 4877 Old 04-19-2018, 07:51 PM
LJG
AVS Forum Special Member
 
LJG's Avatar
 
Join Date: Jun 2000
Location: Brookville, NY
Posts: 4,958
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 701 Post(s)
Liked: 410
Quote:
Originally Posted by Kris Deering View Post
> 2000 shape 4 transition 11 gamma 2
< 2000 shape 3 transition 7 gamma 2


Sent from my iPhone using Tapatalk Pro
TY Kris

Last edited by LJG; 04-20-2018 at 07:04 AM.
LJG is offline  
Sponsored Links
Advertisement
 
Reply Video Processors

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