8K by 4K or Octo HD - the real SUHDTV technology - Page 6

http://www.avsforum.com/t/1428686/8k-by-4k-or-octo-hd-the-real-suhdtv-technology#post_22391403

^^ You are essentially saying what I said back there What do you think is the "natural" sitting distance from the TV, simply based on the non AV savvy friends you have, and how big a city dweller's wall can be? ie if the wall can hold 84" comfortably and people sit 10' away, can 8k really be perceivable?

In addition we also have to take into account that our perception about 4k or 8k TV will likely come with other enhancements besides higher resolution, for eg higher bit depth, wider chroma and better gradation processing. Sharp's 4k demonstration also manipulates the image to have a more realistic 3D feel by blurring images further away, which is familiar to our eyes' perception of depth
Edited by specuvestor - 9/20/12 at 1:50am

My friends sit from 10 feet to about 14 feet away from their HDTVs. Some are 32". Some are 46". One has a 60" HDTV. A city dwellers wall (living room) can be any size (length) with a minimum of 8 feet high.

IMO - yes definitely. You are talking about a format that delivers 33 Megapixels per frame. Unlike 4K which is 8.3 MP compared to HD's 2 MP.

The problem at that distance would be a smaller 4K TV versus a same size HDTV - say 60". I don't believe the difference wiould be noticeable.

If they don't make a substantial increase in the dynamic range, then the increased color depth will not be very noticeable to viewers. The human eye does much better with B & W than it does with Color when it comes to noticing changes:

."The retina contains two types of photoreceptors, rods and cones. The rods are more numerous, some 120 million, and are more sensitive than the cones. However, they are not sensitive to color. The 6 to 7 million cones provide the eye's color sensitivity and they are much more concentrated in the central yellow spot known as the macula. In the center of that region is the " fovea centralis ", a 0.3 mm diameter rod-free area with very thin, densely packed cones."

I have not read all the posts, but isn't something missing in the discussion in relation to the number of pixels and screen size? It seems we are all assuming the same pixel pitch (distance between two pixels), and conclude that more pixels means bigger size. But if we have higher resolution screens with smaller pixel pitch, the screen can be smaller.

If we take the iPhone retina display with 326 pixels per inch as the finest resolution that the eye can detect, then a 8K screen of 7680x4320 pixels will have a screen size of just 23.6" by 13.3". However since we view a TV much further away than a phone, we can consider instead the iPad "retina display" of 264 pixels per inch, which will result in a TV screen of about 29" by 16".

Looking at it in reverse, a 46" TV (roughly 40" by 22.5") with 7680x4320 pixels will have 192 pixels per inch. The viewing distance of a TV is further than for a tablet, so will this still be perceived as a "retina display"?

Anyway my point is that 768x4320 does not necessarily mean a huge TV. With the pixels closer together, the picture will be sharper and will be like real life.The manufacturing technology is here - whether they will be produced is a question of cost versus demand.

By the way, the new 15" MacBook Pro has a "retina display" with resolution of 220 pixels per inch.

This is 2012. NHK says they will be releasing 8K in 2020 or sooner. So we are already in your 6-10 year time frame and "most consumers" are not buying large size HDTVs.

Why all the confusing discussion of screen size and viewing distance when the only significant factor is viewing angle?
Most people find 30 degrees horizontally to be optimum for comfortable viewing and for this reason cinemas are generallydesigned with central seating around this angle. Some people may like to sit in the front row for a wide angle, head turning view and some like to sit in the back row (but watching the screen is not always their primary interest!).
Assuming one has 20/20 vision one can only resolve 1800 elements in a 30 degrees angle making the 1920 pixels of HDTV as clear as such a person can see. Doubling or quadrupling pixels will add nothing but add expense.
Some younger people may have better than 20/20 vision so they might see 4K as a slightly sharper picture but 8K seems totally unnecessary in home use except as a way for manufacturers to sell ever more expensive equipment.

Nope. First, Snellen chart results slightly underestimate (few percent) visual acuity related to displays. Next, about your "maybe some have better than 20/20". Here's a chart with average acuity related to age:

You can't use Snellen 20/20 as a limit. For general public, 0.5 arcminutes for center-to-center pixel spacing can / should be used.

Now about this: Correct number would be 1842 elements horizontally (display is not a hemisphere, it is flat).

We are not going to agree it seems. I was generous in using the 1800 element figure for 20/20 vision since this only applies to high contrast black to white elements and not to lower contrast images where angular resolution becomes significantly worse.
You state
For general public, 0.5 arcminutes for center-to-center pixel spacing can / should be used.
Where is the data to justify such a demand?

Human vision

http://en.wikipedia.org/wiki/Minute_of_arc#Human_vision

Even though Roger Clark from Clarkvision cites sources for 0.3 arcminute spacing and is using that value for display & print resolution calculations in his article Notes on the Resolution and Other Details of the Human Eye, for great majority of us, 0.5 arcminute pixel spacing should be used when making such calculations.

Using only Snellen chart results for determining visual acuity in relation to electronic displays is not such a great idea for several reasons:

Created for different purpose altogether
Many times (I have experienced this myself), when you successfully read all the letters on 20/20 line, your doctor will end your examination concluding further testing isn't needed since 20/20 is perfectly healthy and his / her job is to spot when something is wrong, not to measure just how great your vision is. No difference between 20/20 and 20/6.3 in eyes of a doctor - both are perfectly healthy thus once you reach 20/20, examination is usually ended.

Not precise enough
Every row of letters subtends different angle. Difference between those is not very fine but is not very gradual but sudden.

Very different from normal display image
In order to successfully pass every line on Snellen chart, you have to be able to correctly identify a letter in question. You can see that there are slits on a letter but you won't pass unless you can correctly identify the letter.
Darrel Hopper from Air Force Research Laboratory shares some of the difference between what Snellen chart covers and what human visual system / ideal display covers:


As such, Snellen chart results are poor representation of visual acuity in relation to electronic displays.

I just don't think people are thinking clearly about 8K. It just isn't going into the home. The infrastructure is too costly, the benefits are too small, and outside of a few enthusiasts who think that more is always better, there is no demand for it.

First off, actually making it happen. A 32MP, 10 bit color video stream at 120hz is just stupid-huge. It takes up too much spectrum to broadcast (or transmit from sat or across cable) and requires media far beyond anything foreseeable. Getting 4K@60 onto a quadruple layer BD can be done. The video will be 8x bigger using 8k@120. Adding deeper color and we are looking at longer movies tipping the scales at 1TB. The engineers can check me on this, but we will need a completely different kind of media for this - probably holographic - and that isn't anywhere near coming soon. The spectrum that TV broadcasters have in the US will not hold it unless it is compressed into noise. Cable and Sat distributors won't have this kind of space either. In fact, I doubt 4K will be deployed to broadcast or pay TV.

Second, it all comes down to size. People won't sit close enough to smaller displays and most won't want huge displays in their homes. For example, I'll compare full HD, 4K, and 8K in my setting on a 120" screen - which is bigger than 99.99% of home displays.

• 1080p - full perception of resolution @16'
• 4K or QuadHD - full perception of resolution @8'
• 8K (NOT Octo HD) - full perception of resolution @4'

Who is really going to sit 4' away from a 120" screen? What is that 100 degrees? Sure, when they show it at the trade shows people walk right up to the screen to see the detail, but that isn't how the consumer actually watches TV/movies/sports.

Third, now I am not saying 8K isn't going anywhere. It is the new 70MM. Movies will be filmed at it and big theaters will be outfitted with 8K projectors. Why? Well, because a nice 600" screen at the multiplex looks perfect when the view is 20' from the screen. I am sure that it is no coincidence that 70mm film has about the same resolution as 8k (and 35mm film is about the same as 4K).

So my best guess is that theaters will be 4 and 8K, home recorded content will be 4K, and general distribution will be 1080p. And the following generation won't be using screens - it will be replaced by VR-type displays.

In just a few years 4K displays will fall to today's 2K prices.... at which point it will make sense that "your next purchase" be 4K to best resolve upscaled 1080p sources and future/growing 4K source (4K video games will probably be the first thing to arrive since they're rendered in real time by the player and don't require a new "format" or delivery stream).

You can disagree with many things but you can't say "8K isn't going into home". It goes, there is not question about it. Real question is when. Do you seriously believe that in 200 years (assuming world doesn't come to an end before) we won't have huge displays (coupled with content) with perfect image even when you're 3 inches from it?
You're using 1 arcminute for center-to-center pixel spacing. Plus, you don't have to appreciate whole resolution to be able to tell the difference between the two.

All valid points . . . and yet NHK (Japan) says they will begin broadcasting 8k in 2020 or sooner. They are not getting involved in 4K. They are going from HD to 8K.

NHK can flush billions down the toilet, just like they did with MUSE. The jump from SD to full HD was done incrementally, that is true, but the move to 4K will be sudden. So complaints about 'incremental' improvement are fairly bogus when the move from 1080p30 to 4Kp60 (or even 120) is quadrupling the resolution and doubling or quadrupling the frames. And color depth is being enhanced at the same time. trying to double-jump to 8k - increasing the already acceptable number of pixels by a factor of 16 is just madness.

No, by that time we will have the image projected directly into the eye. Big screens are still a highly niche item. And it isn't even cost. People don't want some big screen crapping up their house. I had to get a drop-down screen with the case completely hidden in order to get past my wife. Right now people happily pay MORE for a 55" LED LCD display when they could get a decent projector and a large screen for peanuts now. The desire to have a big, beautiful screen is not a big desire for the general public. Couple in the serious increase in costs and distribution issues and it is a non-starter. The future is tilting toward digital distribution and a single 8K data stream is greater than a gigabit. That just isn't going to fly. I would also remind you that there will be very little content at that resolution. Only a subset of existing film was shot exclusively at 70mm.

And I used 1 arcminute because that is what my eyes confirm. I also use my display as a monitor and I know exactly what the distance of greatest resolving power and detail is.

NHK doesn't think so. That is what they are committed to.

Super Hi-Vision 8K TV standard approved by UN agency

http://www.bbc.co.uk/news/technology-19370582

According to DisplaySearch, the 50+" segement of the market (FPD) is 16% and growing. I wouldn't call that niche.

A major improvement in video compression is coming very soon. Called H.265 HEVC. It already has 4k and 8K in it's specifications:

http://en.wikipedia.org/wiki/High_Efficiency_Video_Coding

There will be further improvements in the future. Remember, we are talking about 8 years from now (8K in Japan).

As far as 8K content, it will be made using 8K cameras. Keep in mind that the USA will probably wait on implementing 8K. We will once again go the "step route" with 4K first.

I am not sure that's true.
I'm not going to get into how you measured it, I don't need to. What is important is not anecdotal evidence of a individual. We have studies, we'll have more of them. I wrote about this in length before.

It is true - for the camera negative. But the resolution drops each generation you step down from the CN.

Rather than theorize about what can and cant be perceived by the individual, here is a simple demonstration that anyone can use to test how 8k v 4k will affect them.

The image below is a 430 pixel high crop from a 36 Megapixel Nikon D800 image taken under ideal conditions with optimal lens settings and no motion. 430 pixels is about 1/10th the height of 8K but fits easily on just about any PC monitor at 1:1.
View this image at 1:1 on your monitor and measure the image height (not the screen). Now view the image at 10 times that distance to simulate viewing the full image a 1x picture height.

1x picture height is like viewing a 100” diagonal 16:9 screen from 4 feet or a 200” screen from 8' and provides an 84 degree horizontal viewing angle, almost 3 times as wide as recommended for cinema presentation. Personally I find 8' a minimum distance for reasonably good sound presentation, 4' is unusable IMHO so a 100” screen would be too small.

The left side of the image is a straight crop from a uncompressed RAW image, the right side has been scaled down to 4k to strip off all resolution above 4k, and then scaled back up to 8k so that it can be compared side by side with the original image. The upscaling cannot put back any of the resolution lost during the down scaling process so the comparison is fair. More pixels do not mean more resolution, in fact video never has the visible resolution its pixel count would indicate, its always much less.




Here is another comparison, this time from a Canon 5D Mk11. The crop is 852 pixels high or about 1/5th the height of 8k. It needs to be viewed from 5 times image height (when displayed 1:1) to simulate viewing the full image at 1x image height.




With both images I can see a small but noticeable difference between 8k and 4k at the equivalent of 1x full image height, however there are some extra factors that need to be considered.
Anyone who has used a very high res DSLR camera will know that getting a perfectly sharp focused shot is very challenging and requires everything to be just right. Getting the same performance from a movie camera with camera or subject motion will be effectively impossible and even tiny focus errors will drastically affect resolution.

Its a LOT easier to see small differences in still shots compared side by side than it is with full motion video. I seriously doubt that anyone could reliably pick 8k from 4k at 1x image height in a double blind test with full motion video.

Since a viewing distance of 1x image height is impractical with anything but a huge screen I question the usefulness of 8k in the home. I would prefer to have 4k with more color graduations (10bit), 4:4:4 color encoding, higher frame rate and less compression.

People can view these images and work out the viewing distance required for 8k to be of value to them, bearing in mind the still images above will show more difference than any video content.
Edited by Owen - 9/21/12 at 4:56am

All side-by-side tests with 70mm film projector and 4K DLP projector has shown that 4K is the equivalent to 5perf 70mm (same as the 16:9 portion of IMAX 15perf 70mm) both from 70mm scans (scanned at 8K and downsampled) and from 4K cameras (captured at 5K and downsampled).
35mm is about 3.5megapixel at the most, but scanned at 6K and downsampled to 4K can give a pretty good 4K picture digitally projected, but not as good as 70mm scans or digital captured for 4K delivery.
8K is quite much better. It is really at 8K the images are really significant better than than 2K and 4K.
But it of course depends very much on the 8K equipment used and the lighting, lenses, motives and postproduction.

Till now, there hasn't been good enough 8K equipment to really show off what 8K is capable of.
But still, the reports from people that has seen the Sharp 84" 8K prototype TV have all remarked on how much 3D depth there are in some of the scenes, to the degree that they say that if we had 8K TV's, the interest for 3D would fast diminish.

Notice that the material demoed at the Sharp 8K was filmed with the previous generation of NHK's 8K cameras. Same material was shown at IFA 2012.

The 8K material filmed at the London Olympics with NHK's new 8K CMOS cameras, projected on big screens with JVC's "8K" E-Shift projectors, so even those demo's didn't recreate a full true 8K images on the screen.

Next time time the Sharp 85" 8K TV prototype is demoed, we can expect them to use the new 8K material from the London Olympics and we will be a step closer to see what 8K will look like in the future.
But until now it has only been steps on the way to real quality 8K.
Of course the most sensible for the movie industry would have been to utilise the 4K format for many years in the cinema before 4K enters consumer home space.
But that is a opportunity the industry have missed and don't do much at all to rectify, even if more than a quarter of the screens in the US alone are fitted with 4K projectors.
Having lost that opportunity they could rush for 8K, but that won't happen either.
Nobody are building 8K digital film cameras or real 8K digital projectors at the moment.

I am sure VR-type glasses will be popular at some point in the future when they become smaller. The highest resolution VR glasses now are the Silicon Micro Display ST1080, but they are still only 1080p for each eye, even if the pixel-par-inch is an impressive 2976.9 PPI from the 0,74 display.

There is also something to be said about having image display so close to the eye.
For many that is unpleasant and some space between oneself and the screen is preferable.
Because of that, big screens in theatres will always exist, even though the projectors will disappear.

One should remember in all the continuous discussion display on seating distance and angel of view that it is only "half the story".
The real production of image quality is on the capture and post production stage.
As long as that is done quality in mind (high resolution capture resolves details better than low resolution capture), the only part the display play is its ability to reproduce the quality 1:1 that is delivered.
Then seating distance and angel of view will be of less interest.

Today's discussions of high resolution moving images is like going back in time to the days when new high resolution films was possible to watch regularly in movie theatres all over the world.
High resolution movies was lost on the way with the demise of horizontal 35mm (Technirama and VistaVision) and all the 70mm formats, and now we are slowly regaining the possibility to watch high resolution films again some day in the future.
Edited by coolscan - 9/21/12 at 5:24am

Film has quite a high limiting resolution but MTF (contrast) at high spatial frequencies is quite low. The eye needs high contrast (MTF) to see high resolution, without high MTF high resolution is lost to the eye.
Perceived image sharpness depends of high MTF at medium spatial frequencies, very high "resolution" is not required. That's why movie producers concentrate one high MTF at relatively low spatial frequencies well below 4k as thats what the viewer needs in a cinema environment.
Digital cameras have very good mid range MTF, thats why 2k digital camera footage can look very sharp, sharper than 35mm film which has a higher limiting resolution but lower MTF.

In my observations most film titles are not significantly limited by 2k, especially scope titles shot with an anamorphic lens. This is evidenced by how much sharper IMAX titles (shot which spherical lens) can be when delivered at 2k.
Edited by Owen - 9/21/12 at 6:27am

Optics make the image appear to be at infinite distance (infinity focus). Your eyes can't get more relaxed than that.

@Owen - simulation is simulation. Physical pixel structure is different on simulated images and on real lower-resolution display. Why haven't you used nearest neighbor for upscaling, like this:



Edited by Randomoneh - 9/21/12 at 7:25am

I just wanted to make it clear that the ITU was a different organization and I do understand that there were many companies/organizations involved in the ITU study group on UHDTV. I prefer the ITU as an organization since anyone can freely download their standard documents while SMPTE charges the general public $75 to download a single standard document. In my opinion that is a bit absurd in the year 2012.

The main issues were the lack of consumer 4K computer monitors and a connection capable of the needed bandwidth. DisplayPort 1.2 HBR2 is capable of 4K at 60 fps (with 30-bit color) so the computer industry now has a suitable connection and it has been included on a wide range of recent video cards from AMD (from the AMD Radeon HD 7970 to the AMD Radeon HD 7750). I think the only issue at this point is the release of consumer 4K computer monitors.

I have read many articles on UHDTV and I have never seen one that uses the term "Octo HD". irkurk, could you post a few links to the articles that use that term?

Just my opinion but I think we will see ATSC 3.0 released in 4 to 8 years and that it will likely offer an improved modulation system (30 to 60 Mbps in a 6 MHz channel), support for new audio/video codecs (such as HEVC), and support for 4K UHDTV video. ATSC was the first digital OTA broadcast standard for HDTV so it is old compared to several of the more recent OTA broadcast standards.

You must distinguish between capture (film&digital) and delivery after post production.
Both film and digital capture format have low contrast regardless of resolution to capture as much dynamic range as possible.
Desired contrast and electronic sharpness is then added in post. I agree that high contrast increase the perceived sharpness.
Low resolution, f.ex. 2K capture requires more added electronic sharpness added in post than capture done at a higher resolution.
5K or 6K capture for 4K or 2K delivery will have a higher organic sharpness and avoids the Edge Enhancement that we so often see in movies released on BD.
The reason for this EE is that previously one thought that 2K film scans and 2K cameras was sufficient for 2K delivery. When the material is authored for BD, the production people deem the material to be "too soft" and over-sharpens the material.
Now one has learn that film schould be scanned at 4K for 2K delivery and at least at 6K for 4K delivery.
The same with digital cameras that now have gotten sensors with higher than 2K resolution for 2K delivery.
That those cameras also will be used for up-converting to 4K delivery just show how little quality consciousness and care for image quality there are in the movie industry. What I see complained about from people that work in the industry is that there are mostly stupidity, hearsay, fanboyism, incompetence and money that is the reason for choosing a capture format for a movie.
The choice of equipment based on the desire for the best technical Image quality capture and "future proofing" are more rare. The 2K digital cameras does a lot of in camera processing (including electronic sharpening) before the images are stored on the capture cards/tapes.
This processing doesn't happen for 35mm film before post production stages.
That' why the 2K cameras can look sharper.
Similar to the 35mm film format happens to the digital RAW capture formats on those cameras that doesn't pre-process the RAW in camera. Stating as you do; Have you seen much movies shot on Sony F65 (6K'ish) or Red Epic(5K) RAW properly processed, or film scanned at 6K, projected with the latest Barco or Christie Series-2 4K projectors (Sony 4K SXRD/Lcos is softer)?
I don't see the evidence at all in your estimates, because you forget an important point.
That is that IMAX titles usually are scanned at 8K and down-sampled to 2K.

Which shows that very high resolution capture matters very much compared to 2K capture.

By the way; Estimates done for 35mm capture film (not release prints) is that it has from 1.5 to 4 megapixel resolution, depending on type of film used and shooting situation.
Edited by coolscan - 9/21/12 at 10:23am

1. NHK wants to sell 8K cameras to compete with REDs pending 8K cameras. There is a use for that resolution in theaters and specialty situations - not in the home. Trust me, NHK will backtrack and soon start talking about their 4K options in the consumer segment.

2. And the overall display market is shrinking - down 4% YOY. The only growth is in the 40-65" category which has done well due to price drops. And you can hardly call those big screens. To see the full 8K resolution on a 60" display you have to be 2' away from the screen! However, you do start seeing improvement at 4' - which is still too damn close.

3. HEVC can at best double the compression versus h.264. But remember we are increasing the resolution by a factor of 16, the fps by a factor of 4, and the color depth by factor of 2 (which is often forgotten, moving to 10bit color needs a lot more data). HEVC can get 8K@120 down to an 800 megabit stream - and that is still using 8 bit color.

I would also point out something else. Getting the data from the player/decoder to the display. That is an uncompressed signal (currently). Let's do the math:

Resolution x number of color channels x frames per second x color bit depth = number of bits per second of raw uncompressed video

So a ten bit 8K120 would be (7,680×4,320) x 3 x 120 x 10 = 119,439,360,000 bits per second or about 120Gbs. That would currently require TEN HDMI 1.4a cables to carry the signal. Considering that HDMI connection speeds have only doubled in the last 10 years, I just don't see increasing by 10x in eight.

Your numbers are not even close to reality:

http://www.bbc.co.uk/blogs/researchanddevelopment/2012/08/the-olympics-in-super-hi-visio.shtml

You have no data to support that claim. You are still at 1 arcminute. Click here.

The article specified nothing about color depth or, more importantly, frame rate. Remember, NHK was making promises about 8K120 with 10 bit color - which I based my calculations on. The Olympic demonstration could have been at 24fps and 6 bit color. That would drastically reduce their transmission needs.
(7,680×4,320) x 3 x 24 x 6 = 14Gbs The compressed video portion of the signal would be about 94Mbs

Except it specifically says 24 Gbps uncompressed and 280 Mbps compressed, so your guess is incorrect.

Human visual acuity of 1 arcminute is standard knowledge. 20/20 vision is by definition the ability to distinguish 1 arcminute. Just because some people have 20/10 vision doesn't mean 20/20 isn't the standard.