HDR Discussion Thread

[Geoff D]

Yes, that's how absolute luminance works at its core: the code values for any given part of any given shot are encoded with a fixed value that ideally needs to be displayed at that value. It's a transfer function that assumes the OOTF (rendering intent) will already be present in the target display whereas relative luminance systems like gamma (SDR) or HLG HDR rely on the end display to build in the OOTF relative to its own luminance, hence the name. This is why you don't require downconversion metadata and/or special sauce Optimiser modes with a relative system, HLG having being selected by many TV broadcasters as the preferred option for HDR for this main reason, though trials have been done with DV and whatnot.

Many LCDs fall woefully short of providing a decent HDR image, it's partly what made this situation such a cluster**** on top of the lack of any mandatory content-led metadata system. People expect to be dazzled by the 4K supermarket special they just bought, but it ain't happening. Such 'dynamic' metadata can't fix a poor LCD backlighting system - while edge lit isn't great it's still way better than all on/all off frame dimming, and even FALD needs the right amount of zones to truly sing - but it can help to ameliorate it by detecting scene changes ahead of time and adjusting the backlighting accordingly so there's less 'pumping', 'blobbing' etc, as well as maintaining the proper APL for low brightness content because if an image was only 100 nits but encoded in a 1000 nit container and the display maps the static 'container' then it would kill the APL. But even with dynamic metadata those edge and frame lit TVs are still nowhere close to doing what an OLED could do with your theoretical content, yet they wouldn't be quite so rubbish as they are now either.

[edit] on the flip side, you've no idea what the tone mapping in the OLED - leaving dynamic metadata aside - is doing to your theoretical image either, those same movies that are almost SDR in an HDR container prompted many complaints from all quarters about looking too dark e.g. Goodfeathers, and the same goes for ASBL and ABL restrictions when it comes to displaying darker/brighter content for sustained periods so you're not getting a truly unfettered HDR image either. Why, it's almost as if there is no perfect TV...

[MindBlank]

Thanks for the answer! Good to know that I've understood correctly.

Furthermore to what I've written, I guess this also means that 10-bit support is really important, to have very fine grain control on per-pixel luminosity. So this also means that if an edge-lit LCD is 10-bit, than there is still more control over per-pixel luminosity than edge-lit and 8-bit, right?

There's also another thing I don't understand and I've seen in TV reviews. And that is testing peak HDR brightness in a predefined scene (frame). Let's say from a TV show or movie. I don't understand how this is a measure of "real-world brightness" and not a measure of EOTF tracking capabilities and tone-mapping approach of the TV/manufacturer.

If the brightest spot (cluster of pixels, not pixel) on that "real-world" frame is, let's say, supposed to be 1200 nits and TV1 displays this at 650 nits and TV2 at 800 nits, this is not necessarily a real-world brightness test to me.

Maybe TV1 can have a maximum (let's say) 10% window output of 1000 nits, and TV2 900 nits. It's just that tone-mapping has different approaches here, that I think border on subjectiveness more so than anything else.

As above, TV1 would have more [compressed] "space" for highlights even higher than that 1200 nits cluster, while TV2 would have even less "space" and even more compressed.

Have I understood this correctly?

Also, maybe this is for another post, but from what I've read (white papers), I gather that color volume is extremely important now with HDR. Like cement is for bricks. Because 1000 nits of washed-out red color is meaningless compared to 1000 nits of properly satured red (I guess this would border on neon aspect look, now that I think about it )

EDIT: Also, in my previous post - it's not to bash on edge-lit LCD's (I've owned one, actually). I think their success is very scene dependent. If we've got a bright lamp on the left side of the screen and the APL for the remainder of that frame is around 150 nits, than edge-lit can do a decent job. Sure, the entire vertical stripe where the lamp is will be brighter, but 85% of the remaining image will still be good to go... I mean, it's still HDR, for sure in this instance.

[Robert Zohn]

Originally Posted by PaulGo

Beauty is in the eye of the beholder. There are many factors that should influence buying an OLED vs an LCD. Does one want a brighter picture? Is the room the TV in dark or bright? The content that will be watched (which could cause burn-in). Price - the cost of a large OLED TV can be prohibitive.

As the quality of LCD TVs improve can LG keep the costs down to stay competitive? To a large degree OLED TVs are a niche product which will never be sold in mass quantities so will enough consumers purchase this for LG to stay profitable and continue its manufacture?

Thansk Paul, I agree 100% with everything you said in the 1st par.

For 55" and 65" sizes OLED prices have moved moderately lower over the the past 6 years and since they mostly compete with premium high-end LCD/LED TVs, the flagship LCD TVs prices have dropped more slowly. When you compare the 75" - 77" class OLED's are signficantly higher than the smaller OLED TVs, and althogh moderately more costly still in line with flagship 75" LCD TVs.

For now, when you get into the 85" 4K HDR screen sizes you only have LCD/LED technology. Later this year we'll have LG's 88" 8K OLED TV and for next year LG Display will have the new OLED module manufacturing facotry and we'll have a significany increase in production in larger sizes and prices will begin to drop more rapidly.

[Geoff D]

Originally Posted by MindBlank

Thanks for the answer! Good to know that I've understood correctly.

Furthermore to what I've written, I guess this also means that 10-bit support is really important, to have very fine grain control on per-pixel luminosity. So this also means that if an edge-lit LCD is 10-bit, than there is still more control over per-pixel luminosity than edge-lit and 8-bit, right?

There's also another thing I don't understand and I've seen in TV reviews. And that is testing peak HDR brightness in a predefined scene (frame). Let's say from a TV show or movie. I don't understand how this is a measure of "real-world brightness" and not a measure of EOTF tracking capabilities and tone-mapping approach of the TV/manufacturer.

If the brightest spot (cluster of pixels, not pixel) on that "real-world" frame is, let's say, supposed to be 1200 nits and TV1 displays this at 650 nits and TV2 at 800 nits, this is not necessarily a real-world brightness test to me.

Maybe TV1 can have a maximum (let's say) 10% window output of 1000 nits, and TV2 900 nits. It's just that tone-mapping has different approaches here, that I think border on subjectiveness more so than anything else.

As above, TV1 would have more [compressed] "space" for highlights even higher than that 1200 nits cluster, while TV2 would have even less "space" and even more compressed.

Have I understood this correctly?

Also, maybe this is for another post, but from what I've read (white papers), I gather that color volume is extremely important now with HDR. Like cement is for bricks. Because 1000 nits of washed-out red color is meaningless compared to 1000 nits of properly satured red (I guess this would border on neon aspect look, now that I think about it )

EDIT: Also, in my previous post - it's not to bash on edge-lit LCD's (I've owned one, actually). I think their success is very scene dependent. If we've got a bright lamp on the left side of the screen and the APL for the remainder of that frame is around 150 nits, than edge-lit can do a decent job. Sure, the entire vertical stripe where the lamp is will be brighter, but 85% of the remaining image will still be good to go... I mean, it's still HDR, for sure in this instance.

I don't think those TV reviews are saying that that's what the content is actually doing, they're stating what the TV is doing within the limits of its tone mapping and relevant backlighting/ASBL/ABL brightness issues. By 'real world' they mean what the TV can achieve unrelated to any manufacturer-provided bumpf that states it can do one squillion nits or whatever.

[Stacey Spears]

Originally Posted by Penton-Man

Clearly, the individual home user experience for episode 3 of Season 8 GoT was dependent upon what particular display device was used, e.g. https://forum.blu-ray.com/showthread...o#post16366694 , the ambient lighting , as well as, significantly, the deliverable, e.g. vis-à-vis DirecTV, cable, HBO Go, etc.

The respective post house and colorist are well experienced….https://www.siminternational.com/ and I would have expected them to have anticipated the issues many home users experienced with the various distribution channels in handling the challenging content of epi 3. I choose not to fault Joe, the colorist who graded the episode, as colorists can only advise.

b.t.w., I’ve since been informed that apparently the JPEG 2000 delivery presentation at the special screening at the TCL Chinese Theatre in Hollywood looked fine, but I’m waiting on a 2nd opinion for verification.

It is my understanding that the display Joe used for the grade was calibrated at 2.2. gamma, so if you are calibrated for BT.1886, then it would have appeared too dark for you. At 2.2. gamma, I thought it looked fine. So I believe an HDR version would have looked fine too.

On the other side, the resolution of that episode is the biggest disappointment for me since Lock killed Jacob on Lost. Like Jacob, the Knight King was mystery with no answers. But then I digress.

[Scarriere]

Originally Posted by LordoftheRings

Did you thank few arrogant members who keep insulting me.
Then take a rest. The dating advice thread might provide you with pictures of your real interest.

I don’t think there’s pictures in that thread. Just like you with the pictures of “your gear”.
I said I wouldn’t quote you unless you lied. You invite it when you do.

[sapiendut]

Originally Posted by Stacey Spears

It is my understanding that the display Joe used for the grade was calibrated at 2.2. gamma, so if you are calibrated for BT.1886, then it would have appeared too dark for you. At 2.2. gamma, I thought it looked fine. So I believe an HDR version would have looked fine too.

On the other side, the resolution of that episode is the biggest disappointment for me since Lock killed Jacob on Lost. Like Jacob, the Knight King was mystery with no answers. But then I digress.

Personally I find the biggest problem is people watching the episodes with too much lights on and/or viewing using illegal IPTV. None of my clients had the problem. Even my friends who don’t have their TV calibrated and set their TVs to movie mode and watched it in low light areas don’t have that problem either

https://www.wifihifi.ca/LatestNewsHe...ourFault!.html

[mrtickleuk]

Originally Posted by Geoff D

And consumer OLEDs do better at managing tones in darker parts of the image, LCDs do better at managing tones in the brighter parts of the image, as nicely summed up in this SMPTE slide:

Thanks for your very informative post

What does "ACES" mean in the graphic, please? Is it another tricky one like Ate Kay?

[mrtickleuk]

Originally Posted by Stacey Spears

It is my understanding that the display Joe used for the grade was calibrated at 2.2. gamma, so if you are calibrated for BT.1886, then it would have appeared too dark for you. At 2.2. gamma, I thought it looked fine. So I believe an HDR version would have looked fine too.

Oh! First, thanks for the information.

Second, that's a big shock for me. My understanding was that gamma used to be "the wild west" and you had no idea what each production would have been created at, 2.2, 2.3 etc. But finally (very late indeed), the introduction of BT.1886 meant everyone had a standard to use for grading SDR. So why not use it?

On the other side, the resolution of that episode is the biggest disappointment for me since Lock killed Jacob on Lost. Like Jacob, the Knight King was mystery with no answers. But then I digress.

Whatever Happened, Happened. You Cannot Change Time - YouTube

[LordoftheRings]

The list will quickly grow to one million ...
• https://www.forbes.com/sites/paultas...rr-martin/amp/

Lol, what a show! ...Send me a flaming dragon.

[Wendell R. Breland]

Originally Posted by mrtickleuk

What does "ACES" mean in the graphic, please?

Academy Color Encoding System

[Geoff D]

Originally Posted by mrtickleuk

Oh! First, thanks for the information.

Second, that's a big shock for me. My understanding was that gamma used to be "the wild west" and you had no idea what each production would have been created at, 2.2, 2.3 etc. But finally (very late indeed), the introduction of BT.1886 meant everyone had a standard to use for grading SDR. So why not use it?

I keep seeing people say that 1886 is the 'standard' now but I see no real evidence of it. From indie disc producers to the major studios to TV broadcasts, it seems as if they're all still doing something different which is why I like my SDR gamma at 2.3: not so dark it knocks out all the shadow detail, not so bright it makes everything look too thin. It's juuuuuuust right.

[Penton-Man]

Originally Posted by Geoff D

While good contrast is a fundamental cornerstone in getting good HDR, it's disingenuous to claim that it is the only thing needed for good HDR because it's a system built that is built around absolute luminance. That's a fact, and all these things like dynamic metadata and HDR Optimisers are merely agents of compromise when you really stand back and analyse them.

The same is very much true in the reverse though, that it's not solely about high luminance either, and contrast is perhaps still the most important part of the equation but this is why having a combination of great contrast and great brightness unfettered by ASBL and ABL is the utimate goal for HDR and is why I wouldn't trade my ZD9 for anything right now; I'd rather have good contrast and great luminance rather than great contrast and good luminance. Of course it's flawed up the wazoo, it's an LCD TV, but time after time after time I keep watching stuff on it and it still takes my breath away (do I get a point for a Top Gun reference?), it's extraordinary how over-engineered this TV was at the time.

Indeed you do, plus I think you may now be ready for more - https://www.ziprecruiter.com/jobs/av...st10&mid=28896

[Penton-Man]

Originally Posted by LordoftheRings

Penton-Man must be proud of us; he can relax now

better yet , nothing like playing golf to relax and watching the b



ball remain straight rather than sway right or left with a slice or hook. Otherwise, if there’s some twerking, one might be inclined to stop at 9 holes and not complete the full 18.

[Penton-Man]

Originally Posted by Stacey Spears

It is my understanding that the display Joe used for the grade was calibrated at 2.2. gamma, so if you are calibrated for BT.1886, then it would have appeared too dark for you. At 2.2. gamma, I thought it looked fine. So I believe an HDR version would have looked fine too.

If that be the case, I’m thinking that since his proposal was essentially for ~ 2.35, https://forum.blu-ray.com/showthread...ma#post3103648
Charles, likewise, couldn’t see didley squat for episode 3 either.

This gamma talk reminds me of a previous free-for-all discussion about display gammas back in the day - https://referencehometheater.com/201...gamma-correct/

[Penton-Man]

Originally Posted by mrtickleuk

What does "ACES" mean in the graphic, please?

An example of it being incorporated into the workflow –
http://digitalcinemareport.com/artic...e#.XN8ZqchKiUk
that said, some folks believe that using any of the ACES based working spaces have some quirks when using traditional lift-gamma-gain controls that give up a lot of precision, which is why some shops choose to use Rec. 2020 color primaries with a PQ transfer function as their working space.

Originally Posted by mrtickleuk

Is it another tricky one like Ate Kay?

I like Kay , he's another vocal supporter of Top Gun: Maverick

[brps3]

Originally Posted by Penton-Man

not 3, like the banditos….

Good looking F-5

[Penton-Man]

Originally Posted by LordoftheRings

Penton-Man must be proud of us;

Quite so . You guys are doing sufficiently well without me.
2,439,081 – 2,433,563 (from this time - https://forum.blu-ray.com/showthread...r#post16408436 ) equals a tad under 2,000 views/day.

I feel no concerns about taking additional posting sabbatical beyond the completion and celebrations following stage 7 - https://www.amgentourofcalifornia.co...-7-may-18-2019

meanwhile, local lurkers, mark your calendars….https://www.meetup.com/Los-Angeles-C...nts/261504585/

[mrtickleuk]

Originally Posted by Wendell R. Breland

Academy Color Encoding System

Wonderful, thanks! I'm not a fan of unnecessary TLAs and FLAs

[mrtickleuk]

Originally Posted by Geoff D

I keep seeing people say that 1886 is the 'standard' now but I see no real evidence of it. From indie disc producers to the major studios to TV broadcasts, it seems as if they're all still doing something different which is why I like my SDR gamma at 2.3: not so dark it knocks out all the shadow detail, not so bright it makes everything look too thin. It's juuuuuuust right.

Yes, this recent conversation has certainly been interesting to me. I followed the links to Mr Poynton's 2010 PDF where he appealed for a standard. It is prefaced at the top with an extra note dated 24th June 2018, which says that "BT.1886 was adopted about a year after this pleas was made". Just not adopted by him, apparently (!).

I like your idea of using 2.3. That's great. I think that the next time I calibrate I will set my day mode to 2.3 rather than 2.2. I rarely use the day mode anyway, so having 2.3 will be a good "backup" mode to have ready . I normally watch in the 2.4 Night mode.