Screenshot Comparisons 4K

[Markgway]

Originally Posted by cynatnite

Is the washed out look of the UHD due to the HDR/SDR conversion? Because ugh. I've sampled my UHD here at home and it looks nothing like that at all. In fact, it looks pretty dang good by my estimation.

That's what I was wondering.

I don't have this title, but surely the colour drain wouldn't be present whilst watching it proper, would it?

[BitRate]

Originally Posted by cynatnite

Is the washed out look of the UHD due to the HDR/SDR conversion? Because ugh. I've sampled my UHD here at home and it looks nothing like that at all. In fact, it looks pretty dang good by my estimation.

Yes, it's due to the conversion from HDR to SDR. However, the "4K" transfer is much less colorful than the 1080p transfer due to having a more natural color pallette rather than an unrealistically vibrant one.

[Geoff D]

Originally Posted by Markgway

That's what I was wondering.

I don't have this title, but surely the colour drain wouldn't be present whilst watching it proper, would it?

If you ever see a UHD screenshot look like that - very browny yellow with very low contrast - then basically it's not had any kind of appropriate "transform" applied to it. What I mean (and what Deci was trying to explain in the MI thread ) is that the HDR image actually looks like that in its 'raw' encoded state, and it needs to be played back with the relevant transfer function - the thing that turns the electrical signal back into an optical one - in order to appear normal.

The inherent problem with HDR screen caps is just that: we don't have HDR monitors (mostly) so people are having to apply an SDR conversion to make the caps appear normal. And even if we did have HDR monitors you couldn't then do a side by side comparison with actual SDR caps e.g. regular Blu because they use a different transfer function entirely (the old gamma system) and the monitor would have to switch between both profiles in real time to make it work. Alternatively the SDR images could be converted to display in HDR space, but then we're back to the same issue as with HDR to SDR caps now: how do you make sure that the conversion is properly handled?

The issue is that SDR conversion cannot usually contain the full HDR image without appearing excessively dark, so the dynamic range of the image must be compressed or clipped in order to maintain a reasonable level of brightness. This is why caps are notated as having 100 nits or 150 nits or 320 nits or whatever, that's down to the person doing the caps and the levels they're specifying in the SDR conversion and it makes judging UHD by screenshots a rather uneven process, particularly when it comes to dynamic range (as evinced by the comically blown out highlights seen in some recent caps). The caps do have their uses, but they are most definitely not a true representation of the HDR images.

[slask]

Originally Posted by nickpap7

Synapse is more accurate in terms of colors.

said who?

[Deciazulado]

In the interest of BUHDscience.
Since none of the 4 rock images ^^ are the same frame, being one or two apart and that sequence is a vertigo inducing sequence with the camera swirling around the rock here's a fuller sequence with his' inserted in proper order.
(Since BitRate's UHDs 'r raw HDR, its colors 'r transformed too)

BitRate's cropped BD v

BitRate's cropped UHD (color corrected) v

UHD v

UHD (same frame as BitRate's full frame UHD frame) v

UHD v

BitRate's FHD Blu-ray v

[BitRate]

Amazing work. How did you alter the colors, and since the colors are within SDR, what makes them "correct"?

[Markgway]

Originally Posted by Geoff D

If you ever see a UHD screenshot look like that - very browny yellow with very low contrast - then basically it's not had any kind of appropriate "transform" applied to it. What I mean (and what Deci was trying to explain in the MI thread ) is that the HDR image actually looks like that in its 'raw' encoded state, and it needs to be played back with the relevant transfer function - the thing that turns the electrical signal back into an optical one - in order to appear normal.

OK. I think I actually understand that.

The inherent problem with HDR screen caps is just that: we don't have HDR monitors (mostly) so people are having to apply an SDR conversion to make the caps appear normal....

So UHD caps are effectively useless for colour and contrast but good for grain and resolution.

But here's the thing... if caps aren't good for certain elements and reviews (sometimes) can't be trusted is there anyway of knowing if a disc is worth buying without actually buying it? Even opinions on here can vary wildly.

It used to be that something either looked like shit or it didn't.

[Deciazulado]

Originally Posted by BitRate

Amazing work. How did you alter the colors, and since the colors are within SDR, what makes them "correct"?

The formula.

[Deciazulado]

Originally Posted by Markgway

So UHD caps are effectively useless for colour and contrast

Markgway, no, if they are done properly and seen at near the correct brightness level the shots are designed for (in this case, for a ~ 300 nit computer monitor), the tones and colors within your monitor should look correct as close as possible:

Signal + formula = rendering intent.


Signal


Rendering intent



The biggest limitation of seeing all the tones of HDR as explained in the other thread is that most present monitors (specially computer ones) can't reach the levels of brightness that the format specifies (up to 10,000 nits for highlights) which are up to 100 times brighter (up to +6.67 f/stops higher) than regular 100nit sRGB/SDR monitors. For example on the top of the Rock's plate there's a specular highlight shine that's beyond 100 nits so it's pushing at the boundaries but the rest of the tones should be represented properly in that screenshot.

About color, some of the greens there (and other saturated colors in other images) could be at the extremes too. If they are, they'd be displayed at the limits of the monitor's gamut, but the rest of the green (or less saturated) tones should display at their proper colors. If you have a sRGB monitor the extremes display at that gamut, on wider gamut monitors (like P3 4K / 5K iMacs) this extra saturation will display properly.

[gnj1958]

There's a small increase in sharpness in the 4K if you zoom it out but most normal people don't watch movies zoomed out that much. I guess if you get your kicks watching movies as if you're in the front row at the cinema then congratulations you'll see more detail.

I prefer to watch it at a normal distance myself.

I watched the BD and on my TV it looked really nice. Would the 4K look better? marginally but $30 better? nah.

[BitRate]

Originally Posted by gnj1958

There's a small increase in sharpness in the 4K if you zoom it out but most normal people don't watch movies zoomed out that much. I guess if you get your kicks watching movies as if you're in the front row at the cinema then congratulations you'll see more detail.

I prefer to watch it at a normal distance myself.

I watched the BD and on my TV it looked really nice. Would the 4K look better? marginally but $30 better? nah.

What makes the 4K version worthwhile is its implementation of HDR/ Dolby Vision; its one of the best HDR transfers I've seen.

[andreasy969]

Originally Posted by Deciazulado

Markgway, no, if they are done properly and seen at near the correct brightness level the shots are designed for (in this case, for a ~ 300 nit computer monitor), the tones and colors within your monitor should look correct as close as possible:

Yes, 300 nits HDR=>SDR caps will only be displayed "properly" if one pushes the brightness of the monitor to 300 nits (and will usually also look much better than a 100 nits conversion on calibrated 100 nits), BUT, and that's important, you still have an SDR image then. The colors and brightness are therefore still wrong and, most importantly, the dynamic range is also still limited to the dynamic range of SDR. By giving more room to the highlights (which is what higher nits HDR=>SDR conversions basically do), you will lose sth. elsewhere (resp. get crushed blacks for ex). (If you have a HDR signal that peaks at only 100 nits (or even lower), 100 nits is therefore also the way to go with the SDR conversion.)

I just want to make it crystal clear that even a 4,000 nits HDR=>SDR converted cap is still SDR with the according limited dynamic range and is therefore still far from being close to the original. (but still very useful of course!)

I guess you know that, but I'm afraid others might misread your comment into "300 nits SDR converted HDR cap equals HDR with a peak brightness of 300 nits", which isn't the case. (The way I see it, the higher than 100 nits caps are actually just kind of a "gamma trick".)

[Geoff D]

Originally Posted by Deciazulado

It's the hot HDR Dubai sun

Btw it's not my curve. HDR curve with roll after 250>10000 ->320
I can make you one for 1500 but it be too dark on a browser, not Dubai at all.

Where is the horror artifact, can you point it out?

These caps all have some shocking artefacting (which is so bad it can even be seen in the thumbnails!!):

[Pieter V]

Originally Posted by Geoff D

These caps all have some shocking artefacting (which is so bad it can even be seen in the thumbnails!!):

Dunkirk too:








https://www.blu-ray.com/movies/Dunki...9/#Screenshots

[imsounoriginal]

Part of it is almost certainly due to JPEG compression, too. Dunkirk is WB so no surprises there. But I keep hearing about Ghost Protocol's BD having compression issues (even/especially the Best Buy disc) and I can see it in those caps, but I've never noticed it in-motion. Not saying I don't believe those who say it's there, I just question the severity of it.

[Geoff D]

Goddamn, that last Dunkirk one is a doozy! Banding, blocking, weird vertical lines, holy balls that's terrible.

Originally Posted by imsounoriginal

Part of it is almost certainly due to JPEG compression, too. Dunkirk is WB so no surprises there. But I keep hearing about Ghost Protocol's BD having compression issues (even/especially the Best Buy disc) and I can see it in those caps, but I've never noticed it in-motion. Not saying I don't believe those who say it's there, I just question the severity of it.

Sure, but in this case it's potentially the SDR conversion causing those issues rather than the disc itself being at fault. Applying a certain transfer function to content that wasn't intended for it is one of the fundamental reasons why HDR caps will always come with a pinch of salt (amount of salt may vary). As I said, they really do have their uses in certain cases but all too often the official caps on this site and those on capsaholic are causing more problems than solving them.

[Fat Phil]

Originally Posted by Geoff D

Sure, but in this case it's potentially the SDR conversion causing those issues rather than the disc itself being at fault. Applying a certain transfer function to content that wasn't intended for it is one of the fundamental reasons why HDR caps will always come with a pinch of salt (amount of salt may vary).

Oh, fo' sure. I don't recall the Dubai skyline being one giant macroblock sex party when I watched GP. Seems b-ray.com has got to iron out some kinks in the process. No point providing shots when they look like that.

[Geoff D]

Originally Posted by Fat Phil

Oh, fo' sure. I don't recall the Dubai skyline being one giant macroblock sex party when I watched GP. Seems b-ray.com has got to iron out some kinks in the process. No point providing shots when they look like that.

The funny thing is, someone did actually complain upthread of seeing rampant compression artefacts in the skies in Ghost Protocol, and there were several folks who also got blocking/banding in the skies when watching Dunkirk, which just ties back in to what I said about having the proper EOTF to view this content in, i.e. even having a poorly executed HDR tone map (never mind an SDR conversion) can make these artefacts stand out a lot more.

[Pyoko]

Well, with regards to the luminance, if done correctly the only difference between them would be precision (why they came up with a new EOTF instead of using gamma), even less so if there was a way to capture and view in more than 8 bits.

[andreasy969]

Yes, but "the only difference being precision" resp. having to put 10 bit into 8 bit is exactly the problem. That's factor 4. You just can't do that without losing sth/screwing sth up. And if we had 10 bit, it wouldn't be SDR anymore.