4k consumer equipment is fully available

[jmac698]

A selection of news items detailing the availability of the new standard, UltraHD in 4k resolution, which is 3840x2160 pixels.

First Handheld 4K Camcorder by JVC
"...1/2-inch CMOS imager with 8.3 million active pixels, it delivers real-time 3840x2160 footage at 24p, 50p, or 60p. ... Unlike many high end 4K cameras, the GY-HMQ10 is able to output 4K images to a monitor or projection system in real time with virtually no latency. ... Using MPEG-4 technology and a variable bit rate H.264 codec operating at up to 144 Mbps, the GY-HMQ10 records up to two hours of 4K video to economical SDHC or SDXC memory cards. $4,995, the GY-HMQ10 launches today, with market deliveries beginning in March, 2012."
http://www.cdr-info.com/Sections/New...x?NewsId=32137

Sony BDP-S790 Flagship Blu-ray Player Available For Pre-order with 4k Upscaling
http://www.cdr-info.com/Sections/New...x?NewsId=33029

Nec Announces Its First 4K DLP Digital Cinema Projector
http://www.cdr-info.com/Sections/New...x?NewsId=28574

LG To Showcase 84-inch 'ultra definition' 4K TV at CES 2012
http://www.cdr-info.com/Sections/New...x?NewsId=31999

[Lyris]

Woooo, let's watch our 1080p Blu-ray Discs with scaling artefacts! Awesome ;)

[sumawo13]

Considering the fact that they won't add useful features to Blu-ray like higher bitrates or larger capacity discs, I don't see why they would add 4k. 1080p on Blu-ray already struggles with meager bitrate, I couldn't imagine a 20Mbit/s 4k stream.

[Lyris]

I don't think there is any problem with the max bit rate on BD. 40mbps is enough to give, in many cases, a visually identical result to the source, in my experience. The problems are more related to crap masters and crap encoders, don't you agree?

[Ghitulescu]

2k (1080 for consumers), 4k, 8k and so on have not been designed to provide you more resolution (although this was a big part of the advertising). They were designed to "cover" more space. VHS and then DVD could not "fill in" the giant screens of the '90ies then 2000ies. There is a big stupidity to pack more pixels than the human eye can discerne.

[kolak]

We are still MILES away from human eye possibilities- going to take years until we can outperform eye :)

[scharfis_brain]

I've got a 3840x2400 pixels 22" LCD on my desk.
It features a whopping 204dpi (9.2 Full-MegaPixels)

with normal viewing distances 50cm to 1m there aren't any pixels visible at all.
ordinary bluray contents needs to be upscaled 4x :p

[Ghitulescu]

We are still MILES away from human eye possibilities- going to take years until we can outperform eye :)

There are 6 millions cones - so a panel having 6 megapixels (x3) which is placed remotely than the fovea can accommodate, is perceived as continuous (no pixels, no lines).

[Didée]

What about interference & moiré patterns? :)

[Ghitulescu]

What about interference & moiré patterns? :)

We learn these avisynth filters during our infancy :) together with the upside-down image reversing.

[SeeMoreDigital]

I've got a 3840x2400 pixels 22" LCD on my desk.
It features a whopping 204dpi (9.2 Full-MegaPixels)
I used to have a 22" Viewsonic with that kind of resolution... The images it displayed looked freakishly alive ;)

At some point, I'd like to have a look at the 55" Toshiba 55ZL2 "glasses free" 3D TV which offers a native resolution of 3840x2160 pixels.

[turab]

We are still MILES away from human eye possibilities- going to take years until we can outperform eye :)
You need to consider viewing distance and screen size. Based on the average viewing distance, which is about 9 feet, 1080p looks best on a 69-inch screen according to this article: http://hdguru.com/lechner-distance-the-number-you-need-to-know-before-buying-an-hdtv/21/. I'm not sure how that number was derived, though, so YMMV.

4k is reasonable, but I can't figure out the rationale behind Ultra HD.

[Mug Funky]

moore's law has hit display systems it seems.

4k is not at all interesting from a viewing point of view. it's okay for VFX work, but the data requirements are actually insane, so it's a pain for all concerned.

you've plowing into the limits of the lenses and you're well beyond the resolution of modern film stocks at 4k (3k is the point of diminishing returns based on my tests with super35, kodak 5201 and cooke superspeed lenses - the state of the art of 2008. this was on supersharp landscape shots, too).

before anyone points out the size of the crystals on film, try looking up MTFs and diffraction first :)

with digital cinema, we're sharper than we've ever been, but to make a picture that people want to look at, you have to shoot decidedly shallow. only the major edge of the subject will have detail up at 4k, everything else will be drastically less. landscapes are a different matter, so movies like Baraka would be really amazing in 4k. certainly not much else though.

[jmac698]

This is well answered in http://www.cst.fr/IMG/pdf/35mm_resolution_english.pdf

The best subjective resolution of a test pattern displayed at an actual theatre after the standard release process, was 875l/ph.
I believe that's 4.6MP but I'm not quite sure how to convert.

[Lyris]

I would never say no to higher resolution.

That said, I don't think there is much wrong at 1080p - and I watch on a 123" screen.

I'm more concerned with crap transfers making their way onto BD...

[Mug Funky]

the 35mm printing process is dying fast though.

digital projection takes away not only the blur of optical printing stages, but the jittery gates and floaty film transports that smear detail over time. you'd be able to resolve more in a cinema with DCI.

that said, watching Blade Runner in 4k only revealed more grain to me, and sharper titles. i strained like hell to try and see pixel edges in the cinema, but just couldn't. 2k I can on titles, because they're often aliased for some reason (i never did my titles like that, but there you go i suppose).

i hadn't seen that paper though - i'll use that as ammo when someone gets nostalgic to the point of inaccuracy :)

[jmac698]

It seems that lens are the limiting factor in cameras.
You have all the data at hand, but take the green-yellow light and f/8-f/11 aperture values as a reference. It represents a realistic, not too demanding case. Consider a 35mm system with a lens at f/11. At best, the maximum resolution you will get is equivalent to 16 MP, even if your camera has 22 or 25 MP.

http://luminous-landscape.com/tutorials/resolution.shtml

There's all sorts of ways that microscopes get past the resolution limit, in fact there really isn't a limit anymore. Some techniques include near-field reconstruction and negative index of refraction materials. Light has been focussed with -ve index materials!
As far as I understand it, we'll one day be able to see perfectly all across the universe - please someone correct me if they know better.
There's even more ways to make use of light beyond the diffraction limit, like double-patterning in chip manufacture.

You need some care in scanning film, from what I've been reading, there's 9 layers of grain in color film and if you focus on one layer you see a lot of grain, the worst grain even as there's a combinatinon of fast and slow layers with different grain sizes. What you'd want is to combine all the layers sharply to reduce noise. It also seems that you could get a sharper image from the brighter areas but it will be noisier. I think there's improvements possible to scanning. Grain comes from dye patches which develop from the unprocessed silver crystals.

[Biggiesized]

you've plowing into the limits of the lenses and you're well beyond the resolution of modern film stocks at 4k (3k is the point of diminishing returns based on my tests with super35, kodak 5201 and cooke superspeed lenses - the state of the art of 2008. this was on supersharp landscape shots, too).
Hmm... I thought that 5212 (100T Vision 2) was the sharpest stock that Kodak released.

Did you do any testing with it?

[Mug Funky]

only 7212 i'm afraid - the 16mm counterpart.

it's an unpopular stock because it looks like shite. really grainy, much more so than you'd expect from 100 ISO. 50D is certainly easier on the eye.

the main sharpness losses on film are in the red channel, as that's where the diffraction hurts most. this has the advantage of softening skin tones and blemishes though, which saves on make-up :)

[edit]

the testing i did was camera tests for a feature. they were going to cut between 5219 (vsn3 500T) and 5201 (vsn2 50D), and wanted everything done in 6k (!!!), but we talked them down a touch. i must say, the 5219 was impressive as hell. with a careful grade you could cut seamlessly between the two. it helps that they liked to shoot bright and grade dark though.

[pandy]

There are 6 millions cones - so a panel having 6 megapixels (x3) which is placed remotely than the fovea can accommodate, is perceived as continuous (no pixels, no lines).

However distribution is not so regular x,y grid as in display or light sensors - a bit quasi-random, overlapped , more recreate hexagonal than rectangle. So there plenty place for resolution improvements based on http://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem#Reconstruction

[Biggiesized]

only 7212 i'm afraid - the 16mm counterpart.

it's an unpopular stock because it looks like shite. really grainy, much more so than you'd expect from 100 ISO. 50D is certainly easier on the eye.

the main sharpness losses on film are in the red channel, as that's where the diffraction hurts most. this has the advantage of softening skin tones and blemishes though, which saves on make-up :)

[edit]

the testing i did was camera tests for a feature. they were going to cut between 5219 (vsn3 500T) and 5201 (vsn2 50D), and wanted everything done in 6k (!!!), but we talked them down a touch. i must say, the 5219 was impressive as hell. with a careful grade you could cut seamlessly between the two. it helps that they liked to shoot bright and grade dark though.
:thanks:

*Necro-res*

What scanner did you use in post to compare the stocks? Not all scanners are created equal. :sly:

Secondly, did your exposure affect the sharpness of the stocks? For instance, does overexposing by a stop, to create a "thicker" negative, help sharpness in any way? I know grain levels will be less apparent.

[Revgen]

I couldn't imagine a 20Mbit/s 4k stream.

I could if they would quit having ridiculous restrictions on how the codec can be utilized.

See Shon3i's thread for more details. LINK (http://forum.doom9.org/showthread.php?t=154533)

Even so, I don't believe that they would use meager bitrates on 4K discs anyway.

[jmac698]

Well you can have your answer now, as 4k discs have been released.
http://www.cdrinfo.com/Sections/News/Details.aspx?NewsId=35024

The new 4K Ultra HD Video Player is a hard-disc server that connects to Sony's 84-inch 4K LED TV (XBR-84X900) allowing users to view 4K resolution movies and short form 4K videos.

Available as a bonus loaned exclusively to U.S. customers purchasing the Sony 4K LED TV (MSRP: $24,999.99), the video player comes loaded with content, including both full length Hollywood features and a gallery of videos.

Pre-loaded onto the Video Player will be the following collection of 10 full-length feature films presented in native 4K for the home:

- The Amazing Spiderman
- Total Recall (2012)
- The Karate Kid (2010)
- Salt
- Battle Los Angeles
- The Other Guys
- Bad Teacher
- That's My Boy
- Taxi Driver
- The Bridge on the River Kwai

In addition to the full length features, the Video Player will come pre-loaded with a gallery of 4K video shorts, including Red Bull Media House's exclusive 4K videos The Athlete Machine: Red Bull Kluge and four acoustic performances of songs from Red Bull Records' Heaven's Basement's new record, "Filthy Empire." Additionally, Sony's 4K Ultra HD delivery solution is designed to be updated with additional 4K titles and video clips.


I will also add that Total Recall and The Amazing Spider-Man were shot with Red Epic digital cameras and Total Recall, at least, used Panavision anamorphic lenses. These are some of the first digital 4k theatrical releases using Red.