HDR Calibration & Discussion - Page 6 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 441Likes
Reply
 
Thread Tools
post #151 of 786 Old 03-02-2016, 01:46 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 Manni01 View Post
If colorists can master in P3/DCI using P3/DCI saturations and send this unaltered in a rec2020 container, then yes it makes more sense to use DCI/P3 saturations to display this, but I didn't think that was possible.
Why would it not be possible? Storing *any* content (with smaller gamut) in a BT.2020 container is a simple linear light matrix multiplication. Restoring the original gamut is again a simple linear light matrix multiplication. The only loss comes from limited bitdepth.

Quote:
Originally Posted by Manni01 View Post
I thought they were using the rec2020 emulation mode (ie rec2020 saturations) even when their display was limited to P3. Hence they would report a P3 limitation in the metadata, yet would still be using rec2020 saturations to master the content.
I would guess that they master to P3 and then simply put it in a 2020 container. But this is only a guess, I don't know for sure.

Quote:
Originally Posted by Manni01 View Post
I know the first models are likely botched anyway, but it looks like when HDR content is detected, they always switch to the same gamut/gamma calibration. So are they using P3 saturations or rec2020 saturations?
An UHD Alliance insider who has tested several HDR displays from different manufacturers told me that some displays react to different metadata and others don't.
nathan_h likes this.
madshi is offline  
Sponsored Links
Advertisement
 
post #152 of 786 Old 03-02-2016, 09:39 AM
AVS Forum Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 7,044
Mentioned: 7 Post(s)
Tagged: 1 Thread(s)
Quoted: 339 Post(s)
Liked: 342
A perfectly capable, perfectly calibrated, rec.2020 display would be able correctly display any UHD content without metadata.

The metadata is only there to support mapping content when the mastering display is more capable than the consumer display.

The reason to use rec.2020 saturation as the targets for gamut calibrations is because they are on the correct hue axis for the primaries and therefore should see less interaction with other controls.

Last edited by sotti; 03-02-2016 at 09:44 AM.
sotti is offline  
post #153 of 786 Old 03-02-2016, 01:05 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 sotti View Post
A perfectly capable, perfectly calibrated, rec.2020 display would be able correctly display any UHD content without metadata.

The metadata is only there to support mapping content when the mastering display is more capable than the consumer display.
True. But let's discuss a perfectly capable DCI-P3 display instead, with a perfectly implemented firmware. I think that's nearer to what we can hope to be able to buy in the near future.

I make the claim that if you calibrate such a perfectly implemented DCI-P3 display by using BT.2020 test patterns in a BT.2020 container with BT.2020 metadata, and if you then play a UHD Blu-Ray with DCI-P3 metadata, you will get inaccurate colors near 100% DCI-P3 saturation. I furthermore make the claim that you would not get this problem if you had calibrated the display with DCI-P3 test patterns in BT.2020 container with DCI-P3 metadata instead.

Do you agree?

To be honest, I'm not sure if you're interested in discussing this in depth. I believe my claims are correct, but getting on the same page on this might take a bit of time, discussing possible display firmware behaviour etc.

Quote:
Originally Posted by sotti View Post
The reason to use rec.2020 saturation as the targets for gamut calibrations is because they are on the correct hue axis for the primaries and therefore should see less interaction with other controls.
Ok, I've not considered this. With "other controls" do you mean saturation & calibration controls in the display setup? Are you saying that those are likely to work in BT.2020 space, and that using those might produce unwanted interaction with Calman calibration if Calman calibrates using DCI-P3 primaries?
madshi is offline  
Sponsored Links
Advertisement
 
post #154 of 786 Old 03-02-2016, 01:24 PM
AVS Forum Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 7,044
Mentioned: 7 Post(s)
Tagged: 1 Thread(s)
Quoted: 339 Post(s)
Liked: 342
Quote:
Originally Posted by madshi View Post
I make the claim that if you calibrate such a perfectly implemented DCI-P3 display by using BT.2020 test patterns in a BT.2020 container with BT.2020 metadata, and if you then play a UHD Blu-Ray with DCI-P3 metadata, you will get inaccurate colors near 100% DCI-P3 saturation. I furthermore make the claim that you would not get this problem if you had calibrated the display with DCI-P3 test patterns in BT.2020 container with DCI-P3 metadata instead.

Do you agree?
A strongly qualified No.

There is no standard as of now for what to do with the metadata, so the difference in the meta could cause a small difference, a large difference, or no difference at all.

A correct calibration between those two should yield the same result. Doing an accurate calibration with metadata means knowing how the metadata will be used and being able to map that correctly to the target values. So the end game of calibration would be that both of those calibrations would produce the results the other generates when flipping the content back and forth.

The question of what metadata to use is complex. I agree, especially till the dust settles, that we should calibrate while using metadata that closely resembles typical mastering metadata. Hopefully we'll have strong guidelines before the mastering displays start to diverge.

Quote:
Originally Posted by madshi View Post
Ok, I've not considered this. With "other controls" do you mean saturation & calibration controls in the display setup? Are you saying that those are likely to work in BT.2020 space, and that using those might produce unwanted interaction with Calman calibration if Calman calibrates using DCI-P3 primaries?
I mean that adjusting CMS yellow, or CMS red may have a non-trival effect on a DCI green because it's hue isn't on the same axis as rec.2020 green.
sotti is offline  
post #155 of 786 Old 03-02-2016, 02:11 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 sotti View Post
A strongly qualified No.

There is no standard as of now for what to do with the metadata, so the difference in the meta could cause a small difference, a large difference, or no difference at all.

A correct calibration between those two should yield the same result. Doing an accurate calibration with metadata means knowing how the metadata will be used and being able to map that correctly to the target values. So the end game of calibration would be that both of those calibrations would produce the results the other generates when flipping the content back and forth.
Well, my claim was about a display with a "perfectly implemented firmware", which means I have a strong opinion about which processing it should apply depending on the metadata. I think we probably have a different opinion about how displays will (or should) behave. If you were in charge of designing a firmware for a display with a native DCI-P3 gamut, how would you make it behave when being fed with:

a) BT.2020 container, metadata says DCI-P3 ?
b) BT.2020 container, metadata says BT.2020 ?

I'm especially interested in whether you would apply some saturation compression in a) or b) to avoid hard clipping?

Quote:
Originally Posted by sotti View Post
The question of what metadata to use is complex. I agree, especially till the dust settles, that we should calibrate while using metadata that closely resembles typical mastering metadata.
So does that mean you recommend using test patterns with DCI-P3 metadata?

Quote:
Originally Posted by sotti View Post
I mean that adjusting CMS yellow, or CMS red may have a non-trival effect on a DCI green because it's hue isn't on the same axis as rec.2020 green.
You mean the display's CMS? When using 3dlut calibration, why would you use the display CMS at all? Wouldn't you normally set it to neutral/native values and let the 3dlut do all the work?
madshi is offline  
post #156 of 786 Old 03-02-2016, 02:44 PM
AVS Forum Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 7,044
Mentioned: 7 Post(s)
Tagged: 1 Thread(s)
Quoted: 339 Post(s)
Liked: 342
Quote:
Originally Posted by madshi View Post
I'm especially interested in whether you would apply some saturation compression in a) or b) to avoid hard clipping?
In an ideal world there would be a published gamut fitting algorithim. If you tell the display that you were mastered on 2020 and the display is less capable there should be some mapping that fits the 2020 gamut into the display gamut preserving as much intent as possible.

But if this is accurately predicted by the calibration software, then it wouldn't mater either way. Both methods would produce the same results. Meaning "Calibration A" was done with "Signal A" it would measure the same as Calibration "B" if it was also fed "Signal B".

Quote:
Originally Posted by madshi View Post
You mean the display's CMS? When using 3dlut calibration, why would you use the display CMS at all? Wouldn't you normally set it to neutral/native values and let the 3dlut do all the work?
I don't believe external LUT boxes will work at all with metadata enabled displays.
thomasfxlt likes this.
sotti is offline  
post #157 of 786 Old 03-03-2016, 12:41 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 sotti View Post
In an ideal world there would be a published gamut fitting algorithim. If you tell the display that you were mastered on 2020 and the display is less capable there should be some mapping that fits the 2020 gamut into the display gamut preserving as much intent as possible.

But if this is accurately predicted by the calibration software, then it wouldn't mater either way. Both methods would produce the same results. Meaning "Calibration A" was done with "Signal A" it would measure the same as Calibration "B" if it was also fed "Signal B".
Ah ok, now I see where you're coming from. If the gamut fitting algorithm were published and actually and properly used by the display, I would agree with you. But that's not the case today. So given that nothing is specified, would you agree that a well designed HDR10 display (with a native DCI-P3 gamut) would behave like the following?

a) For BT.2020 metadata, a reasonable gamma fitting algorithm would be used.
b) For DCI-P3 metadata, no gamma fitting would be done, because the display can handle all colors without needing gamma fitting.

At least that's how I would design a HDR10 display firmware. What do you think?

Quote:
Originally Posted by sotti View Post
I don't believe external LUT boxes will work at all with metadata enabled displays.
You have a point there for HDR content. But with SDR content, shouldn't 3dlut calibration still be possible? At least if you do one separate 3dlut calibration for any common gamut metadata (e.g. one for DCI-P3, one for BT.709 and one for everything wider than DCI-P3)? One problem I see here, though, is that the user might not know which content uses which metadata, so when having one calibration for DCI-P3 and one for BT.2020, ther user might never be sure which one he should activate for any given UHD Blu-Ray title.

And then there's e.g. the Panasonic OLED which reportedly has built in 3dlut functionality. Maybe that 3dlut could be externally "filled" by Calman, if it's processed *after* tone and gamut mapping?

-------

Additional thought: Are you sure that all HDR capable displays out there will do the internal CMS in BT.2020 color space? Isn't it likely that some displays with native DCI-P3 gamut might as a first step convert all input color spaces to DCI-P3 and then do all processing in DCI-P3? It might make sense because if the processing chips have limited bitdepth, the display manufacturer would gain some accuracy this way. If the display does that, it would make more sense to calibrate using DCI-P3 saturations, to avoid negative interaction with the CMS controls, wouldn't it?)
madshi is offline  
post #158 of 786 Old 03-03-2016, 01:05 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,011
Mentioned: 307 Post(s)
Tagged: 0 Thread(s)
Quoted: 5360 Post(s)
Liked: 5412
Quote:
Originally Posted by madshi View Post
One problem I see here, though, is that the user might not know which content uses which metadata, so when having one calibration for DCI-P3 and one for BT.2020, ther user might never be sure which one he should activate for any given UHD Blu-Ray title.
This is exactly the point I was making a few posts up in favor of using rec2020 saturations all the time: there is no second guessing about the content on the end user side. It works for current titles and upcoming titles using a wider than P3 gamut (as in my Inside Out example) and works while both mastering displays and consumer displays progressively evolve towards rec2020, which is the standard for both UHD Bluray and UHDTV (neither of them mentions DCI/P3).

It also works for displays which would be able to reach a wider gamut if not calibrated to P3. For example, when calibrated to rec2020 saturations, a display might reach a wider red than it would be if calibrated to P3, simply because green struggles to get past the P3 green point. Why should we hobble other colors because one is harder to reach? When rec-709 green was undersaturated on my JVC, I wasn't trying to reduce the gamut to SMPTE-C to get the narrowest gamut that would entirely fit. This is the same with P3. Manufacturers are trying to claim full support for a narrower gamut which isn't really relevant anymore, because it looks better on paper. Some reviewers and calibrators feel happy about that because it allows them to show neater graphs at 100% sat, but in reality it's kind of irrelevant.

Also you keep behaving as if all consumer displays were still limited to P3. It's simply not the case. There are more and more consumer displays which support a far wider native gamut than P3, some of them being very close to reaching 100% of rec2020. Why would you want to hobble these, especially as we expect content to be sooner or later mastered wider than P3?

Provided colorists are using the rec2020 emulation mode when grading consumer content on their P3 monitor (which as I understand is what they should be doing), using rec2020 saturations seems to be the only way to go to be sure the content is displayed properly irrespective of what the metadata says.
Gordon Fraser and nathan_h like this.
Manni01 is offline  
post #159 of 786 Old 03-03-2016, 07:34 AM
AVS Forum Special Member
 
sillysally's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 5,145
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 1174 Post(s)
Liked: 1050
Quote:
Originally Posted by sotti View Post

I don't believe external LUT boxes will work at all with metadata enabled displays.
So what you seem to be saying is what Jim posted in this link will not work?

https://www.avsforum.com/forum/37-vid...l#post41661249

ss

"Don't worry be happy"
LG77C8, Klein K10-A, Jeti 1211, CMS Lightspace, Calman home plus licence for K10-A, Murideo 6G.
sillysally is offline  
post #160 of 786 Old 03-03-2016, 07:45 AM
AVS Forum Special Member
 
Join Date: Nov 2014
Location: New York
Posts: 5,241
Mentioned: 181 Post(s)
Tagged: 0 Thread(s)
Quoted: 4283 Post(s)
Liked: 5235
I have an interesting but "low tech" question: Now that some of these new 2016 sets like the LG G6 OLED and others have HDR picture modes, how do they calibrate their sets at the factory? When I looked at some HDR content on the G6, the calibration "looks" pretty good by my eye.

John
Sony 55A1E, A9F / LG 55OLEDC8
Marantz 7012, Ohm Walsh Speakers
Klein K10-A, Jeti 1501, Murideo Six-G Gen2
Calman Ultimate, ISF Level III Certified
jrref is offline  
post #161 of 786 Old 03-03-2016, 07:51 AM
AVS Forum Special Member
 
Light Illusion's Avatar
 
Join Date: Aug 2010
Posts: 1,746
Mentioned: 24 Post(s)
Tagged: 0 Thread(s)
Quoted: 733 Post(s)
Liked: 1062
We are actually working with Lumagen on this - and yes, there should be no problem with LUT boxes working with any display - metadata enabled, or not.

We have been working with a lot of display and other manufacturers on HDR/HLG, and have developed some very interesting (and simple) solutions most of the various issues HDR, HLG and UWG throw-up.

Hopefully the HDR market will settle down, and then life will become a lot easier.
For now, I'm happy to have discussions on an individual level for those interested in HDR/HLG and UWG.

Steve
sillysally and Tomcup like this.

Steve Shaw
LIGHT ILLUSION

Light Illusion is offline  
post #162 of 786 Old 03-03-2016, 07:59 AM
AVS Forum Special Member
 
Light Illusion's Avatar
 
Join Date: Aug 2010
Posts: 1,746
Mentioned: 24 Post(s)
Tagged: 0 Thread(s)
Quoted: 733 Post(s)
Liked: 1062
Quote:
Originally Posted by jrref View Post
I have an interesting but "low tech" question: Now that some of these new 2016 sets like the LG G6 OLED and others have HDR picture modes, how do they calibrate their sets at the factory? When I looked at some HDR content on the G6, the calibration "looks" pretty good by my eye.
Actually, the calibration is very simple.
They calibrate to Rec2020 - to the gamut limit the display is capable of.
This means that up to the gamut limit the colour will be accurate, and beyond that there will basically be no further 'colour' change with input signal change.

And this is exactly how 'professional' grading displays are calibrated to Rec2020.

As most images contain colours that are well within the maximum Rec2020 gamut - actually, not a lot more than Rec709 is capable of - the clipping, or restriction in colours that can be displayed, will be next to invisible for most footage.

It will be 'generated' rather than captured images that will show possible limitation, such as animation, where potentially all the gamut is used during production.

When images are 'captured' via camera, the captured gamut is rarely greater than Rec709.

Steve

Steve Shaw
LIGHT ILLUSION

Light Illusion is offline  
post #163 of 786 Old 03-03-2016, 09:58 AM
AVS Forum Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 7,044
Mentioned: 7 Post(s)
Tagged: 1 Thread(s)
Quoted: 339 Post(s)
Liked: 342
Quote:
Originally Posted by sillysally View Post
So what you seem to be saying is what Jim posted in this link will not work?

https://www.avsforum.com/forum/37-vid...l#post41661249

ss

I'm sure he'll be able to pass on the metadata.

But if you just think about how calibration works, a 3D LUT is based on having it's output triplets having a static meaning. With HDR metadata the output of LUT is going to be re-interpreted by the TV. The TV is going to expect a linear input to do it's mapping, but the manipulation of the 3D LUT will make the signal non-linear. This will lead to the TV's adjustments to be incorrect. Since the amount of adjustment in the TV will vary based on metadata, unless the lumagen has pre-knowledge of how the tv is going to adjust it simply cannot do it accurately in all instances.

You certainly can do it accurately for one set of metadata, but that's probably all an external box will ever be able to do, unless TV's start implementing a "bypass" that puts the TV into HDR/2020 mode, but ignores the metadata. In which case the lumagen it self could apply the EETF/gamut mapping processing before the LUT.
sotti is offline  
post #164 of 786 Old 03-03-2016, 06:54 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
It is early for me to be responding since we have not yet done the implementation for HDR, or Rec 2020 (as far as HDMI EDID and info frames). I am going to try to answer, but also say that once we get the implementation done, my answers may change to some degree.

First a question: Why assume we would pass on the Meta data to the TV/projector since, as someone said, they can do what they want with it?

Historically we have found that the less we let the TV/projector do the better. Their decisions are generally based on marketing and not quality video (IMO). Examples are plenty, but include white points at 30,000 Kelvin, Red push, Green push, Gamma of 1.8 to hide TV noise near black, etc. We (Lumagen) will do our best to prevent the TV/projector from being able to get it wrong. I hope that does not sound too strong a wording, but we feel very strongly that we need to provide the absolute best video, and feel this is not always the case with the TV/Projector manufacturers who must (for good reasons) put their profit margin first. Our business model is to put quality video first and so far it has worked for us - perhaps only because we are a small company serving a niche market of people very concerned about quality.

We will be working to support HDR and Rec 2020 sources on non-HDR and non-Rec2020 devices as well as those that support them. We will also allow user control on how to handle things like "clip to the TVs maximum light output" for HDR source, or "Compress to fit the HDR data within the maximum TV brightness," or some combination.

So, at this time, I believe the Radiance Pro will interpret any Meta Data based on the measured profile of the TV/projector and do appropriate processing in the Radiance Pro based on the measured color/brightness results, and output a video stream without Meta Data that best uses the TV/projector capabilities to provide the best possible image quality. Of course we have yet to do the implementation and my opinion may end up changing at least to some degree as we complete the work.

As Steve of Light Illusion said we are working together to provide the best possible solution for Rec 2020 and HDR, even for TVs/Projectors that do not report these features in their EDID. Light Illusion is further along with these in their calibration software than Lumagen is in our implementation, but we are working to get HDR and Rec 2020 support into the Radiance Pro as soon as we can.

Jim Peterson
Lumagen
jrp is offline  
post #165 of 786 Old 03-04-2016, 06:47 AM
Member
 
cinedream be's Avatar
 
Join Date: Jan 2014
Posts: 59
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 49 Post(s)
Liked: 58
Quote:
I don't believe external LUT boxes will work at all with metadata enabled displays.

Take a look at this then ;-) http://www.murideo.com/prisma-hdr.html
cinedream be is offline  
post #166 of 786 Old 03-04-2016, 07:07 AM
AVS Forum Special Member
 
Gordon Fraser's Avatar
 
Join Date: Dec 1999
Location: U.K.
Posts: 1,912
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 233 Post(s)
Liked: 281
Quote:
Originally Posted by cinedream be View Post
Take a look at this then ;-) http://www.murideo.com/prisma-hdr.html
Maybe i am missing something but this box actually appears to do absolutely nothing for actual HDR content as it is SD and HD input only. They are advertising it's use to take in HD and SD content in and then make the display enable it's HDR preset (if it has one) so as potentially larger colourspace may be deployed....which presumably a 3D LUT would then restrict back to the correct colours for the content you want to watch.....or are they suggesting you watch with incorrect, oversaturated colours?

Convergent-AV
ISF, Home Theater Consultants and Distribution, U.K.
Gordon Fraser is offline  
post #167 of 786 Old 03-04-2016, 07:47 AM
Member
 
cinedream be's Avatar
 
Join Date: Jan 2014
Posts: 59
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 49 Post(s)
Liked: 58
Quote:
Originally Posted by Gordon Fraser View Post
Maybe i am missing something but this box actually appears to do absolutely nothing for actual HDR content as it is SD and HD input only. They are advertising it's use to take in HD and SD content in and then make the display enable it's HDR preset (if it has one) so as potentially larger colourspace may be deployed....which presumably a 3D LUT would then restrict back to the correct colours for the content you want to watch.....or are they suggesting you watch with incorrect, oversaturated colours?


You are correct sir, it is 1080P ;-)
cinedream be is offline  
post #168 of 786 Old 03-04-2016, 01:30 PM
AVS Forum Special Member
 
Chad B's Avatar
 
Join Date: Oct 2002
Location: Bradford, OH
Posts: 2,716
Mentioned: 100 Post(s)
Tagged: 0 Thread(s)
Quoted: 692 Post(s)
Liked: 1575
I did a very basic HDR calibration on a Sony 75X940C, which has a manually selectable HDR mode. I then looked at a few 4K BD movies on the Samsung 4K BD player using this mode.

I switched CalMAN's colorspace target to Rec BT.2020 UHD, and gamma formula to ST 2084 HDR (PQ).

I used the color checker measurement in CalMAN, measuring 100% Rec 2020 primaries and the new Color Checker Video patch set, which appears to be familiar Rec 709 targets in a Rec 2020 container (is this correct?). I used an i1D3 profiled off a Jeti 1211, and I verified the linearity of the profile to make sure the i1D3 wasn't getting oversaturated. I used the L20 size windows with a slightly increased pattern delay. There seemed to be some dynamic limiting or dimming the longer the pattern was displayed.

The results seemed to measure well, and the material looked fantastic, though being critical I would say the gamma looked just a bit low/washed out (not by much).

Here's the report. The "before" measurements are not factory default measurements; I had already done white balance and maybe a couple other things in that mode when the before measurements were taken. Is EOTF the best gamma graph to use?
Attached Thumbnails
Click image for larger version

Name:	Sony 75X940C HDR Video-page-001.jpg
Views:	180
Size:	129.5 KB
ID:	1294001   Click image for larger version

Name:	Sony 75X940C HDR Video-page-002.jpg
Views:	167
Size:	157.0 KB
ID:	1294033   Click image for larger version

Name:	Sony 75X940C HDR Video-page-003.jpg
Views:	207
Size:	122.7 KB
ID:	1294041   Click image for larger version

Name:	Sony 75X940C HDR Video-page-004.jpg
Views:	158
Size:	117.0 KB
ID:	1294049   Click image for larger version

Name:	Sony 75X940C HDR Video-page-005.jpg
Views:	147
Size:	58.5 KB
ID:	1294057  

WiFi-Spy and Tomcup like this.

touring ISF/THX calibrator with Jeti 1211 and Klein K-10
Latest reviews:JVC RS540, JVC RS3000, Sony VPL-VW695ES
Copied settings measured, Review archive

Last edited by Chad B; 03-04-2016 at 01:34 PM.
Chad B is offline  
post #169 of 786 Old 03-08-2016, 09:17 AM - Thread Starter
AVS Forum Special Member
 
ConnecTEDDD's Avatar
 
Join Date: Sep 2010
Location: Athens, Greece
Posts: 7,752
Mentioned: 179 Post(s)
Tagged: 1 Thread(s)
Quoted: 2921 Post(s)
Liked: 3569
HDR/WCG Systems Survey - Technical Paper

The new UHDTV system will benefit greatly from Higher Dynamic Range (HDR) and Wide Color Gamut (WCG). These enhanced features have been received very positively by consumers, and CE manufacturers are moving very quickly to introduce equipment to the market with various levels of UHD capability.

This fast movement in the industry has left many confused about UHDTV capabilities, proposed system features, system compatibility, and standards.

This preliminary draft paper attempts to shed light on HDR and WCG, the different industry proposals, and how we test and calibrate the image accuracy of these new generation TVs. This paper will be repeatedly updated as the industry comes together on system standards.

Download it from here: http://calman.spectracal.com/hdr.html
nathan_h and Manni01 like this.

Ted's LightSpace CMS Calibration Disk Free Version for Free Calibration Software: LightSpace DPS / CalMAN ColorChecker / HCFR
S/W: LightSpace CMS, SpaceMan ICC, SpaceMatch DCM, CalMAN 5, CalMAN RGB, ChromaPure, ControlCAL
V/P: eeColor 3D LUT Box - P/G: DVDO AVLab TPG
Meters: JETI Specbos 1211, Klein K-10A, i1PRO2, i1PRO, SpectraCAL C6, i1D3, C5
ConnecTEDDD is offline  
post #170 of 786 Old 03-08-2016, 12:31 PM
AVS Forum Special Member
 
sillysally's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 5,145
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 1174 Post(s)
Liked: 1050
Quote:
Originally Posted by ConnecTEDDD View Post
HDR/WCG Systems Survey - Technical Paper

The new UHDTV system will benefit greatly from Higher Dynamic Range (HDR) and Wide Color Gamut (WCG). These enhanced features have been received very positively by consumers, and CE manufacturers are moving very quickly to introduce equipment to the market with various levels of UHD capability.

This fast movement in the industry has left many confused about UHDTV capabilities, proposed system features, system compatibility, and standards.

This preliminary draft paper attempts to shed light on HDR and WCG, the different industry proposals, and how we test and calibrate the image accuracy of these new generation TVs. This paper will be repeatedly updated as the industry comes together on system standards.

Download it from here: http://calman.spectracal.com/hdr.html
Interesting.

I just read this part for HDR-10. http://calman.spectracal.com/blog/sp...-format-hdr-10

In that article Derek Smith say “3D LUTs are the new standard for reference display calibration,” noted Smith. “3D LUTs are the best way to improve picture quality, and SpectraCal’s CalMAN is the easiest solution to create and load a 3D LUT"

However according to Joel Barsotti. Calman's 3DLUTs will not work effectively with metadata found in a Rec. 2020 container, no matter what the color space is or the container, when a external 3DLUT processor is used. Even if the new 4K (input, output) Lumagen Pro is used that can process and pass 4K 3DLUT's (still in the beta stage).
https://www.avsforum.com/forum/139-di...l#post42095985

Steve at Lightspace disagrees along with Jim at Lumagen.

Its my understanding that Lightspace is working closely with Lumagen to overcome the metadata issue.
My question, does Calman have any interest in also working with Lumagen to over come the metadata issue?

ss

"Don't worry be happy"
LG77C8, Klein K10-A, Jeti 1211, CMS Lightspace, Calman home plus licence for K10-A, Murideo 6G.
sillysally is offline  
post #171 of 786 Old 03-08-2016, 02:57 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
My best guess is that for proper 3dlut calibration in an external box, that box should first do gamut and tone mapping, and after that run the resulting video through the 3dlut. Furthermore, the external box might then update/modify the metadata to account for the gamut and tone mapping, so that the display still switches into HDR mode (if there is an extra mode for that, with maybe higher brightness), but hopefully skips tone + gamut mapping inside of the display. Alternatively the external box could simply drop all the metadata and make the display believe it's conventional SDR content. By letting the external box do all the gamut + tone mapping, proper 3dlut calibration should be possible. I agree with Joel that 3dlut calibration is going to be problematic if the display applies tone/gamut mapping afterwards.
madshi is offline  
post #172 of 786 Old 03-08-2016, 03:10 PM
Advanced Member
 
<^..^>Smokey Joe's Avatar
 
Join Date: Aug 2006
Location: Hobitville, New Zealand
Posts: 533
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 84 Post(s)
Liked: 109
@sillysally


I believe that currently there is a lot of holes in the whole HDR, calibration arena. End users could lose a lot of money chasing early release hardware for something that is in flux.


My simple solution is do nothing, spend nothing until the dust settles.
JimP and hungro like this.

Masterpiece Calibration Ltd
Lincolnvilleshire
New Zealand
<^..^>Smokey Joe is offline  
post #173 of 786 Old 03-08-2016, 03:56 PM
AVS Forum Special Member
 
gwgill's Avatar
 
Join Date: Jan 2013
Location: Melbourne, Australia
Posts: 1,073
Mentioned: 23 Post(s)
Tagged: 0 Thread(s)
Quoted: 334 Post(s)
Liked: 248
Quote:
Originally Posted by sillysally View Post
However according to Joel Barsotti. Calman's 3DLUTs will not work effectively with metadata found in a Rec. 2020 container, no matter what the color space is or the container, when a external 3DLUT processor is used.
It's not the meta-data as such, it's the whole "who's doing what when" basic logic problem.

TV manufacturers want to offer a plug and play user experience, so they are naturally going to hide the HDR source range to actual display range processing (HDR->ADR) inside the TV.

For calibration/3DLut to work, it has to be after the HDR->ADR that's being done in the TV, and just ahead of the display itself. That's the problem.

One solution is to have some way of telling the TV to not do the HDR->ADR, and to act as just a dumb, HDR display. An external 3DLut box can then do the HDR->ADR and calibration, but it needs to somehow know what the HDR characteristics of the display are, in order to do this. (Maybe the3Dlut contents are enough).

The other alternative is that the TV has to make its own internal calibration processing accessible to external tools, including a "calibration" mode that allows a signal source to exercise the display without the HDR->ADR. Such a "calibration" mode would probably satisfy what's needed for the external 3DLut box solution as well.

It's not clear to me whether the current meta-data standards actually allow for such a thing, or merely distinguish between Standard dynamic range and encoding vs. High dynamic range and encoding.

(It's pretty typical for system developers to overlook the needs of calibration/profiling, as they imagine that it's all going to be done "at the factory".)
gwgill is offline  
post #174 of 786 Old 03-08-2016, 05:15 PM
AVS Forum Special Member
 
sillysally's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 5,145
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 1174 Post(s)
Liked: 1050
Quote:
Originally Posted by gwgill View Post
It's not the meta-data as such, it's the whole "who's doing what when" basic logic problem.

TV manufacturers want to offer a plug and play user experience, so they are naturally going to hide the HDR source range to actual display range processing (HDR->ADR) inside the TV.

For calibration/3DLut to work, it has to be after the HDR->ADR that's being done in the TV, and just ahead of the display itself. That's the problem.
I don't understand why ADR would cause a issue, isn't ADR being used in today's SD/HD movies, TV, video games.?

As far as i can tell when my 4K LG gets a UHD signal, it simply switches over to use preset settings for Gamma, Gamut using wide gamut and increases OLED light and contrast. My guess is that the meta-data is telling the display to switch over to those presets.

As long as the video doesn't have that UHD tag, then why can't you simply use the correct Gamma/Gamut and light output, wide gamut, etc, profile and make a 3DLUT as we do now except for rec2020, HDR-10 or whatever is used.?

I can understand Joel's point about Linear, but I don't think the OLED 4K consumer displays are linear to begin with. I am sure I am over simplifying, but it seems to me with some modifications to codes, 3DLUT's should work in a external processor that can input/output 2.0 HDMI process 4K UHD 3DLUT's.

One thing is for sure, a guy like you will come up with a solution to these issues.

ss

"Don't worry be happy"
LG77C8, Klein K10-A, Jeti 1211, CMS Lightspace, Calman home plus licence for K10-A, Murideo 6G.
sillysally is offline  
post #175 of 786 Old 03-08-2016, 06:01 PM
AVS Forum Special Member
 
gwgill's Avatar
 
Join Date: Jan 2013
Location: Melbourne, Australia
Posts: 1,073
Mentioned: 23 Post(s)
Tagged: 0 Thread(s)
Quoted: 334 Post(s)
Liked: 248
Quote:
Originally Posted by sillysally View Post
I don't understand why ADR would cause a issue, isn't ADR being used in today's SD/HD movies, TV, video games.?
The whole point of an HDR TV is that it has a dynamic range greater than today's Standard Dynamic range displays :-)

So you have a movie encoded in an HDR space (say up to 10000 cd/m^2), and your TV does 1000 cd/m^2. How are the two matched ?
Answer - the TV maker puts a "magic sauce" algorithm in, to convert from one to the other. It could do anything, including having dynamic or spatial dependent behavior. That's not something you can calibrate through, nor is such an algorithm designed to simultaneously cope with a particular instance of a displays characteristic. You modularize it, making it work between standard HDR encoding and a perfectly calibrated display of the type you make for your TV, and you have a separate set of processing to calibrate that particular display instance.
Quote:
One thing is for sure, a guy like you will come up with a solution to these issues.
Sure - except I would need to be in the right place at the right time - which I'm not. It's not like I'm working for a company with the resources or interest to fly me around the world to participate in HDR development, nor pay me to develop standards or solutions. Like most people, I have to be content to hang on the sidelines, and attempt to pick up the pieces on my own, when all the smoke clears.
sillysally likes this.
gwgill is offline  
post #176 of 786 Old 03-08-2016, 07:14 PM
AVS Forum Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 7,044
Mentioned: 7 Post(s)
Tagged: 1 Thread(s)
Quoted: 339 Post(s)
Liked: 342
Quote:
Originally Posted by madshi View Post
My best guess is that for proper 3dlut calibration in an external box, that box should first do gamut and tone mapping, and after that run the resulting video through the 3dlut. Furthermore, the external box might then update/modify the metadata to account for the gamut and tone mapping, so that the display still switches into HDR mode (if there is an extra mode for that, with maybe higher brightness), but hopefully skips tone + gamut mapping inside of the display. Alternatively the external box could simply drop all the metadata and make the display believe it's conventional SDR content. By letting the external box do all the gamut + tone mapping, proper 3dlut calibration should be possible. I agree with Joel that 3dlut calibration is going to be problematic if the display applies tone/gamut mapping afterwards.
This is pretty much it.

Also we've talked with Jim from lumagen and he's on the same page. Remeber the Lumagen is much more capable than a LUT box.

To calibrate HDR with a 3D LUT it needs to be closer to the display in the signal chain than the display management (mapping) module that applies the meta data. Something like a prisma or an EE color will never be HDR capable.
sillysally likes this.
sotti is offline  
post #177 of 786 Old 03-08-2016, 09:43 PM
Advanced Member
 
James Freeman's Avatar
 
Join Date: Sep 2013
Posts: 757
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 517 Post(s)
Liked: 348
I have a few questions.

1. How (if at all) the player actually knows the xyY Chromaticity coordinates of the TV?
2. Who and How does the color conversion from rec.2020 on the UHD-BD disc to the gamut of the TV?
3. What colorimetric intent is used to map the gamut (or is it just clipped)?

Something to think about:
The TV may just signal the Player "I'm a DCI-P3 capable TV" and the player will remap the rec.2020 to DCI-P3, but the TV actually nowhere near DCI-P3 but more like 75% or 90% of it.
Do you think the colors will be off?.. They most certainly will!
I can bet that the player has absolutely no clue of the TVs actual xyY coordinates, just a simple 709/DCI-P3/wide EDID "words" and nothing more.

OR I am completely wrong and the TV actually sends a full "SEI message" with the peak luminance and xyY coordinates and all.

Can anyone provide a good reading about these of the SMPTE standards like 2084, 2086, 2094 etc...

Quote:
HDR Metadata Standards
The following documents describe the HDR extensions for the entire broadcast chain:

SMPTE ST 2084
High Dynamic Range ElectroOptical Transfer Function of Mastering Reference Displays

SMPTE ST 2085
Color Differencing for High Luminance and Wide Color Gamut Images

SMPTE ST 2086
Mastering Display Color Volume Metadata Supporting High Luminance and Wide Color Gamut Images

SMPTE ST 2094
ContentDependent Metadata for Color Volume Transformation of High Luminance and Wide Color Gamut Images

Last edited by James Freeman; 03-08-2016 at 09:54 PM.
James Freeman is offline  
post #178 of 786 Old 03-08-2016, 11:57 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 James Freeman View Post
1. How (if at all) the player actually knows the xyY Chromaticity coordinates of the TV?
2. Who and How does the color conversion from rec.2020 on the UHD-BD disc to the gamut of the TV?
3. What colorimetric intent is used to map the gamut (or is it just clipped)?
My information is that the movie studios can enable or disable HDR -> SDR conversion capability for their UHD Blu-Ray discs, and an UHD Alliance member told me studios were planning to disable that capability and ship an additional 1080p Blu-Ray for downward compatability with old players *and* old displays instead. Which means that the player is not likely to do any complicated processing. The displays are responsible for all the magic.

If we let e.g. the Lumagen do the processing, I'm pretty sure there will be options in the Lumagen setup to allow the user to define the exact xyY coordinates of the TV. Whether or not the Lumagen will have the processing power to do the processing with complicated algorithms (things like colorimetric intent etc), I can't say. It might be simpler algorithms. But I suspect they'd still be higher quality than what any of the displays is likely to do.

BTW, in theory the gamut mapping could actually be baked into the 3dlut. That would allow highest quality gamut mapping, because the calibration software could do the mapping offline in very high precision with slow and complicated algorithms. However, this would only work for a number of predefined sets of metadata (e.g. BT.2020 and DCI-P3). UHD Blu-Ray allows virtually any source gamut to be defined via metadata, though, so I'm not sure if Lumagen + calibration software devs are going to bother offering this option, because in order to cover all the bases, I suppose Lumagen would have to implement gamut mapping in software (instead of relying on the 3dlut), anyway. Unless just using the BT.2020 3dlut for all "unknown" metadata gamuts would be considered better than trying to do real time gamut mapping within the Lumagen firmware / FPGA? I can't really say.
Manni01 likes this.

Last edited by madshi; 03-09-2016 at 12:00 AM.
madshi is offline  
post #179 of 786 Old 03-09-2016, 12:34 AM
Advanced Member
 
James Freeman's Avatar
 
Join Date: Sep 2013
Posts: 757
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 517 Post(s)
Liked: 348
Quote:
Originally Posted by madshi View Post
My information is that the movie studios can enable or disable HDR -> SDR conversion capability for their UHD Blu-Ray discs, and an UHD Alliance member told me studios were planning to disable that capability and ship an additional 1080p Blu-Ray for downward compatability with old players *and* old displays instead. Which means that the player is not likely to do any complicated processing. The displays are responsible for all the magic.
I don't think it is the current situation with the Samsung UHD-BD Player (K8500) .
One of the fellows on the Blu-Ray.com board managed to post some real screenshots straight out of the UHD Player using an HDMI 2.0 Splitter (most probable).
His Player was connected to an HDMI 2.0a TV (HDR capable Samsung 65JS9500) so the Player received a SEI back from the TV with its capabilities.

His UHD-BD captured images: http://forum.blu-ray.com/showpost.ph...&postcount=215
His BD vs UHD-BD captured images: http://forum.blu-ray.com/showpost.ph...&postcount=237
Thread: http://forum.blu-ray.com/showthread....273716&page=11

I can only guess but the captured images using a Splitter and a 4k capturing card (I assume) look like they are in REC.709 & a Gamma curve, that means the Player actually did the conversion before the capturing card.
If the TV did the conversion, the Player should have sent a REC.2020 & ST.2084 EOTF picture and the captured images would have looked desaturated and over-bright just like Life of Pi or Exodus demos we have (not using madVR).
Or simply looking at a REC.2020 & ST.2084 picture on a REC.709 & 2.2 Gamma monitor.

Luckily madVR can read the HEVC metadata and convert Color Spaces, 2084 to Gamma, 2086 to custom brightness range, all these with high precision and quality dithering.

Any thoughts?

Quote:
Originally Posted by madshi View Post
My information is that the movie studios can enable or disable HDR -> SDR conversion capability for their UHD Blu-Ray discs, and an UHD Alliance member told me studios were planning to disable that capability and ship an additional 1080p Blu-Ray for downward compatability with old players *and* old displays instead. Which means that the player is not likely to do any complicated processing. The displays are responsible for all the magic.
That's not what I have read.
To my knowledge the up coming HDMI 2.1 with ST.2094 does just that, it converts HDR to SDR and has Dynamic metadata frame-by-frame if desired.
http://www.ip.philips.com/data/downl...hite_paper.pdf
https://www.smpte.org/sites/default/...2015-FINAL.pdf
That is an advancement not a step back, I don't think anyone stepping back now or anytime soon.

Yes, they already ship the regular 1080p disc with the UHD disc, but I'm pretty sure that the player can scale to SDR the HDR master to an HDMI 2.0 (non-HDR capable like HDMI 2.0a) TV.
What else can do the conversion if the TV can't even understand the 2.0a metadata, or what if it's a 709 HDTV (HDMI 1.4)?

Last edited by James Freeman; 03-09-2016 at 01:18 AM.
James Freeman is offline  
post #180 of 786 Old 03-09-2016, 06:36 AM
AVS Forum Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 7,044
Mentioned: 7 Post(s)
Tagged: 1 Thread(s)
Quoted: 339 Post(s)
Liked: 342
Quote:
Originally Posted by James Freeman View Post
I have a few questions.

1. How (if at all) the player actually knows the xyY Chromaticity coordinates of the TV?
2. Who and How does the color conversion from rec.2020 on the UHD-BD disc to the gamut of the TV?
3. What colorimetric intent is used to map the gamut (or is it just clipped)?.
1) The player does know the EDID info (which usually does have xy values).
2) Gamut mapping happens in the TV.
3) Up to the TV manufacturer.

The samsung player will play the UHD disc on non-HDR devices, but all of them need to be HDMI 2.0, so I haven't seen it work on a 1080p panel.

The remapping based on the metadata is designed to happen in the display itself. Most of the first HDR content was over the top streaming with Amazon and Vudu leading the way.
Manni01 likes this.
sotti is offline  
Sponsored Links
Advertisement
 
Reply Display Calibration

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