Why 24fps is not enough for HDR movies

[Rasmus Larsen]

I thought this was an interesting article. The section "You Can’t See the Difference – or Can You?" is in line with what I've experienced in front of some 8K TVs. The brain begins to perceive the picture as something closer to reality. It has more depth, especially on contrast-rich displays.

8K TVs Top TV Line-ups for a Reason

[Rasmus Larsen]

It's worth pointing out that the writer, Chris Chinnock, is the Executive Director of the new 8K Association.

A few paragraphs that also caught me eye.


True 8K resolution:
"It is best to look for an 8K TV that does not do sub-pixel rendering, which means sub-pixels are shared to create full color pixels. Traditional RGB pixel architectures with 7680x4320 pixels is best. Each pixel should be able to show the full range of colors and luminance of the display.

LCD TVs that use sub-pixel rendering will have a hard time displaying a single pixel wide grid of black and white (or colored) lines. Since it is the fine details and sharp edges of an 8K display that makes it so impressive, sub-pixel rendering should be avoided."

As you may know, Samsung's latest high-end 4K and 8K TVs sacrifize resolution to improve viewing angles. It's discussed in our Samsung Q90R review (also see picture example).


Color volume:
"Methods to measure this for HDR-capable displays is still evolving, so I would not put too much value on this specification."

I've been saying this since manufacturers started using color volume in marketing. I've talked about ABL - and inverse ABL - on LCD TVs, and how it affects color volume measurements. Also most methods for quantifying color volume does not take into account the absolute black point (only the measured black point). Other concerns are test patterns vs. reality, color and contrast drifting due to zone dimming, ABL, white boost etc.)

And with the 2019 generation of LCD TVs with quantum dot films that are still sold as "100% color volume" certified TVs, despite hitting just 90-95% DCI-P3, it's probably wise for everyone involved to revise thoughts on the matter until better methodology has been developed.

[Steve]

Yes, so for a solar powered video projector, can I use THE SUN as my black point? I should get a lot of good negative contrast values that way.

It's unreal that people do stuff like this. When measuring color you are measuring pure color, not washed out color with a grey reference point as black.

I was thinking the other night, that the bleed through doesn't matter as long as it's low enough that the pupil contraction makes it fall below the visual threshold (difficult with me). So there is room for a value of "perceived color" at different brightness levels (ambient/reflected light, dimly lit, adequately lit, full lit, extra bright (and outdoor sun for outdoor use) each at standard set external brightness levels in nits etc). To give an objective comparison between sets in marketing, and an relative idea of what to expect I'm this situations. I'm not thinking my Q7 is going to hit its color volume targets with the significant more than expected bleed through I'm seeing. As I go above 39 or 40 backlight in dynamic, I noticed direct washing out. I wasn't even looking for it, so I don't know much is below 40. As the black filter seems to be somewhat linear in response, this will go down to zero, as I can see a mild black even at the bottom in a dark room (which is OK, if it was like that at full backlight. More like what a dual cell lcd might be like).

[Steve]

There is another factor. On the Samsung 8k demos, I have seen stores showing them at fullhd or ultra HD instead of 8k. Not only that, they are too compressed and look horrible. Bland surfaces with fuzzy borders. You need high quality sources to get the best benefits out of 8k. Even ultra bluray is too low bit rate for me. We are talking about prosumer compression rates or more. That's around 600mb/s h265 plus. At 300mb/s, it's low. So, some fancy encoding tricks needed for streaming, to give perceived value to 8k content. And h266/267/268/h269/h270, or the Redray codec). You can double, or quadruple, that datarate again for extra quality compared to original lossless cinema camera footage retaining the visual contrast range. 75mb/s+ quality streaming is a must in the mean time to get you to level theoretically at h268. H266 is around, and h267 I think I heard was being fast tracked. I forget if they were fast tracking h268, but that puts them at Bluray quality per pixel at h267, which is what the industry wants, which is like low end prosumer, instead of the much better high end prosumer with 4:2:2 12 bit. The redray codec was originally demonstrated at 7 or so mb/s years ago, and described as flawless or whatever, by cinema people who came to the release event. I called BS in that, that by the time you stream it and to get actual high quality, it would be at similar rates to what we now see. But, the 7 or so mb/s, probably equates to barely visually lossless, which equates to streaming 8k at under 30mb/s. It might have been 9 mb/s with sound data. We had a similar claim the previous decade for Adams platform over here, that when it caner to streaming fell off. That might had been more compression than red ray. As you use more extreme high compression techniques the file becomes more fragile for streaming, so extra data recovery options can come into play, packing it with a lot of non image data. My own simple codec design might out do red ray, at under 20mb/s for 8k, and the design I'm pursuing from the 1990's is aimed at lossless at a number of times more than red ray, but this would need a lot of data recovery options but aimed at getting a lossless 8k stereo image at the same data rate as the Red ray small data rate demo.

[Steve]

Yes, so for a solar powered video projector, can I use THE SUN as my black point? I should get a lot of good negative contrast values that way.

It's unreal that people do stuff like this. When measuring color you are measuring pure color, not washed out color with a grey reference point. I was thinking the other night, that the bleed through doesn't matter as long as its low enough that the pupil contraction makes it fall below the visual threshold (difficult with me). So there is room for a value of "perceived color" at different brightness levels (0 ambient/reflected light, dimly lit, adequately fly lit, full lit, extra bright (and outdoor sun for outdoor use) each at standard set external brightness levels in nits etc. To give an objective comparison between sets in marketing, and an relative idea of what to expect I'm this situations. I'm not thinking my Q7 is going to hit its color volume targets with the significant more then expected bleed through I'm seeing. As I go above 39 or 40 backlight I'm dynamic, I noticed direct washing out. I wasn't even loom mg for it, so I don't know much is below 40. As the black filter seems to be somewhat linear in response, this will go down to zero, as I can see a mild black even at the bottom in a dark room (which is OK, if it was like that at full backlight. More like what a dual cell lcd might be like).

[Rasmus Larsen]

Why you shouldn't buy an 8K TV in 2019

[rogerzop]

It's unreal that people do stuff like this. When measuring color you are measuring pure color, not washed out color with a grey reference point as black.