HDR heatmaps and torch mode UHDs

[Synchg]

yes so how does one go about finding out if the HDR grade is an automatic one or one done under director's/cinematographers eye. there should be logo on case methinks

[Pyoko]

Originally Posted by Synchg

yes so how does one go about finding out if the HDR grade is an automatic one or one done under director's/cinematographers eye. there should be logo on case methinks

I don't think the problem is automatic vs. supervised as much as the fact that there's only so much you can do working from an already graded SDR master. Apart from Fury Road, which is kind of borderline since the BD and UHD releases were just 6 months apart, does anyone know of, or suspect, any catalog DI titles that were regraded from scratch from the raw source footage?

[Geoff D]

Originally Posted by Pyoko

I don't think the problem is automatic vs. supervised as much as the fact that there's only so much you can do working from an already graded SDR master. Apart from Fury Road, which is kind of borderline since the BD and UHD releases were just 6 months apart, does anyone know of, or suspect, any catalog DI titles that were regraded from scratch from the raw source footage?

Am I to infer from that comment that you think that most HDR catalogue grades are of the upconverted SDR variety?

[Pyoko]

Originally Posted by Geoff D

Am I to infer from that comment that you think that most HDR catalogue grades are of the upconverted SDR variety?

I don't know, it's all I've seen, but I have less than a handful of them so my sample size is really too small to form any hypotheses.

[Geoff D]

I think it's just the luck you've had to pick out more movies than not that've had an 'SDR upconversion', I wouldn't be as effusive as I am about UHD if everything was just a flat upconvert of SDR with no extra highlight detail.

In the case of Lego Batman, see the article below. They worked a hell of a lot on the movie in the SDR style (though in the wider gamut, as afforded by theatrical finishes) and then ported it over to HDR, so I'm not surprised that it's not as expansive for HDR as people might want/assume it to be. To that end, they actually address the point of it being better to create the material in HDR first and then downgrade the SDR from that, rather than start with SDR and uprate the HDR from there.

https://wwwm.creativecow.net/interst...et/article.php

But even so: Lego Batman is a feast for the senses in 4K HDR and even if the HDR is 'fake' it's still got all that colour which trounces the SDR Blu-ray, and it's also got much more of a grainier texture on the UHD disc which ties into the kind of "'70s movie" aesthetic that they were trying to get. (It still cracks me up that they animate movies with all this anamorphic bokeh, depth of field and whatnot.)

[Pyoko]

Upconverted from SDR of course, but even if there's no additional detail in the highlights the HDR does wonders for making them feel less clamped. The SDR grade had a good amount of compressed highlight detail and didn't clip excessively so that no doubt helped, having more to pull from for the HDR expansion.

Average brightness is too high, and some parts are excessive, the holo-interface at the command deck in particular bothered me. Had a blast watching it though, enjoyed it much more than I remembered from when the movie first came out. A very fun UHD.









Album

With a similar highlight roll-off the HDR version would on average be equivalent to the SDR version with peak white at 199 nits (Gamma at 2.40), based on 20 samples. (Min: 119, Max: 265, Median: 196)

[Pyoko]

Originally Posted by Geoff D

But even so: Lego Batman is a feast for the senses in 4K HDR and even if the HDR is 'fake' it's still got all that colour which trounces the SDR Blu-ray

Oh the UHD is most definitely glorious and the BD doesn't even compare. As much crap as I might seem to give upscales and upconverts I can enjoy them greatly if they make good use of what they have. There was a good local multi-buy deal so I stocked up on some more modern and "pop" titles than my usual fare. Probably Hellboy or Prometheus next.

[Geoff D]

No worries, I got me a liking for several discs that would appear to just be the laziest converts from 2K SDR. Even that there fake-assed Moon got some love: https://forum.blu-ray.com/showpost.p...&postcount=287

Re: Pac Rim it's the opening up of the low end that I appreciate the most (not unlike Moon, funnily enough), the blacks are crushed AF on the SDR while they 'unclamped' a TON of shadow detail for the UHD pass. This can be plainly seen even in your heatmaps, like the one of Yuriko in the suit there. So even though the highlights aren't doing anything, the extra detail in the lower end helps to prevent it from feeling too "narrow" in HDR.

[Kris Deering]

Many of the grades being done are the opposite of SDR upconverts. Most are done in HDR from the start with a SDR trim pass done after the fact. I just think there isn’t a lot of range on the capture side that truly justifies massive amounts of true dynamic range.

[Pyoko]

Originally Posted by Kris Deering

Many of the grades being done are the opposite of SDR upconverts. Most are done in HDR from the start with a SDR trim pass done after the fact. I just think there isn’t a lot of range on the capture side that truly justifies massive amounts of true dynamic range.

For many newer titles and new film scans yes, for the DI catalog titles, I'm not yet convinced. The SDR and HDR having the same amount of highlight detail in and of itself is no proof, as you say it could have been limited by the camera capture and it's quite possible they managed to compress all of it for the SDR grade.

However, unless you knew exactly the curves that were applied to go from the scene-referred linear/log source to the subjective, graded SDR world, and reversed those, then going from that graded SDR to HDR would very likely incur a certain loss of variation and steepness in the upper end. Mostly because this kind of thing is difficult to eyeball after the fact, but it also might not be practically possible to reverse it to such a degree due to the bitdepth. This loss and similarity to the SDR version is something I've observed in all of the DI catalog titles I've looked at so far, but none of the straight-to-HDR titles.

[Pyoko]

Nothing particular, just wanted to confirm what I saw when I watched it. The UHD is brighter but the highlights are as compressed as on the BD. Also spotted some broken pixels in the red light on that last shot.








Album

With a similar highlight roll-off the HDR version would on average be equivalent to the SDR version with peak white at 238 nits (Gamma at 2.40), based on 11 samples. (Min: 139, Max: 320, Median: 246)

[Pyoko]

Do I even need to say it? I'll say it anyway. SDR conversion. Not the most boosted HDR either. Hardly the biggest upgrade, but it's OK, I don't know if it's just 2K vs HD but they managed to squeeze out a bit more fine detail and sharpness here. The BD is a bit dark, that first shot is clipping badly yet still only reaches 50% brightness.








Album

With a similar highlight roll-off the HDR version would on average be equivalent to the SDR version with peak white at 285 nits (Gamma at 2.40), based on 25 samples. (Min: 211, Max: 372, Median: 289)

[Geoff D]

Bah. Get Out still looks far better than the BD to me.

[biggquis]

Okay, what does this mean... in layman's terms?

With a similar highlight roll-off the HDR version would on average be equivalent to the SDR version with peak white at 285 nits (Gamma at 2.40), based on 25 samples. (Min: 211, Max: 372, Median: 289)

[cms382]

Pacific Rim (and Prometheus to a lesser extent) looks like another one where the color volume required can really be a challenge for tone-mapping on inferior displays. And grading the clipped areas hot is probably the correct choice for that era of digital filmmaking.

[Pyoko]

Mucho improved over Prometheus with true HDR this time, though not with the greatest range. I still like Prometheus much better as a movie, more interesting scenes, locations and characters. This was such a re-tread, although Fassbender is great as always.

The shot with the green laser is interesting as you can clearly see the outline of a "ghost gamut" larger than P3, most likely the gamut of whatever display it was mastered on.








Album

With a similar highlight roll-off the HDR version would on average be equivalent to the SDR version with peak white at 182 nits (Gamma at 2.40), based on 18 samples. (Min: 171, Max: 221, Median: 177)

[Pyoko]

Originally Posted by biggquis

Okay, what does this mean... in layman's terms?

It's basically trying to do a rough estimate of how bright the HDR/UHD version was mastered compared to the SDR/BD. HDR is designed on absolute values, so if the content says a pixel should be 537 nits in brightness, then on a reference display it will show up as 537 nits.

SDR on the other hand is a relative scale. The pixel is defined on a scale from 0% (completely black) to 100% (maximum white) and then the actual brightness output depends entirely on the settings of the display: the LCD backlight or OLED light combined with the gamma setting. Recommended in a dark room is for maximum white to output around 100 nits and gamma 2.4 is common.

In terms of the HDR this means they effectively decide for you at the time of mastering how bright it will be on your display, to a greater degree than on SDR. Because the whole HDR scheme is more complicated with more variables it's not always possible to change this as easily on the display side (at least not with known good and expected results), which can be an issue if the HDR content was mastered too bright or too dark.

As for the numbers, in that particular case if you were to play the BD, and you set your display to gamma 2.4 and maximum white to output 285 nits, then you would get something that roughly resembled the UHD in average brightness on the low-to-mid end, except the UHD would also have much brighter highlights on top of that because that's what HDR is about.

So if you could attenuate the highlights of the UHD to the same degree as they are on the BD (this is what is meant by highlight roll-off), which is something completely theoretical and not something you can or would want to do, then they would look similar with those settings.

[lgans316]

Mad Max Fury Road Storm scene can hurt your eyeballs. It's similar to BvS final fight.

I do like bright images but I think constant exposure to very bright contents is not good for your eyes on a long term. No wonder most of the UHDs are capped at 1,000 nits.

[nissling]

Originally Posted by Pyoko

SDR on the other hand is a relative scale. The pixel is defined on a scale from 0% (completely black) to 100% (maximum white) and then the actual brightness output depends entirely on the settings of the display: the LCD backlight or OLED light combined with the gamma setting. Recommended in a dark room is for maximum white to output around 100 nits and gamma 2.4 is common.

Very true overall.

I'd only like to point out that SDR relies on that the display handles the signal much like a CRT. On LCD, OLED, plasma etc the EOTF is an added mathematical algorithm. Essentially like a LUT. Without it, SDR would look very flat and dull. A CRT monitor however behaves this way natively, so in practise you could say that modern displays are actually emulating those old tubes when when it comes to SDR.

The old Sony BVM monitors like D24 usually had an EOTF of around 2.6-2.7 when calibrated to BT.709/D65 with white at 100cd/m² and black level set right. It's rarely practical to completely simulate a grading suite in a home enviroment tho realistically speaking.

Your analyzes are great, by the way! Always very interesting to see how titles are getting the HDR treatments as the results vary greatly from title to title.

[andreasy969]

Originally Posted by lgans316

I do like bright images but I think constant exposure to very bright contents is not good for your eyes on a long term. No wonder most of the UHDs are capped at 1,000 nits.

IMO there's actually nothing wrong with going beyond 1.000 in general, if used properly as in for very small highlights only. Since perceived brightness (brightness is the perception) doesn't increase linear with increased luminance (the factual value), 10.000 nits also doesn't (or shouldn't - never seen 10.000 nits ) look 10 times brighter than 1.000 nits. (which I think is also the reason why the lower nits values come with much higher precision btw)

1.000 nits APL is another story though - "The Meg" says hello.