8K by 4K or Octo HD - the real SUHDTV technology - Page 7

http://www.trustedreviews.com/news/nhk-s-super-hi-vision-plans-could-bring-ultra-hdtv-as-soon-as-2016

(7,680×4,320) x 3 x 30 x 8 = 23.9Gbs - Bingo! They are doing 8Kp30 at 8 bit color depth. So they are going to have to quadruple their frame rates and double their color depth to meet their own promises.

That article went on and on about the computer power required to do this. Now remember that what they are promising is 8x more intensive. And they think they can deploy this to the consumer in 8 years.

Don't you get the same result when you substitue 25 fps and 10 bit color depth?

BTW - if you read the article above your post, NHK is now saying they think they can deliver 8K in 4 years.

Into Movie Theaters.

Remember, their standard includes 22.2 channel audio too. They are trying to beat RED to be the camera/projector provider for digital cinema that can actually reproduce 70mm/IMAX quality for events. This is not going into the homes of regular consumers since it requires a really big screen. That doesn't mean there won't be some rich guys installing them in their mansions. Those same people also own medium-sized jets. Everyone just needs to get over 8K in the home. Seeing it in the theater might actually make it worth braving the masses at the cineplex.

4K is the real deal. It can provide the current cinema resolution in the home. It can fit on quadruple layer BluRays (which new players are beginning to support). It is ideal for 60-150" screens - which homes can have. Combined with the latest compression there is even some chance of getting pay-TV providers to provide a few channels of it. And most notably - most previously filmed content can get their highest quality scan at this resolution. That means when you buy Lord of Rings @4K you know that you have the maximum level of detail. They will have to remake the movie to get higher.

My actual fear is that people and companies will balk at 4K right now because they think they will need to go to 8K. The market needs to acknowledge that 8K isn't a consumer technology and focus on 4K in the here and now.

No, that gets 24.9. And would also be really stupid. This is the Olympics - which by definition involve lots of motion. Some very rapid motion. 30fps seriously sucks for sports. Hell, I bet a 4K display at 120fps and 10 bit color would actually look a lot better to the viewer. I know from experience that 720p60 looks better than 1080p30 at sporting events.

It doesn't specifically say that does it?

Trying to beat Red? They already did. Red has shown no 8K camera in fully working condition in a real environment.

Sharp showed an 85" SHV display. That is one inch larger than the LG/Sony 4K TVs that will go on sale this year.

It''s so easy to say . . . 4 layer BDs. But in reality the manufacturing of such a BD-ROM has never been done before. No one knows the yield rate will be. No one knows the reliability rate when it comes to mass produced players. Sorry - BDXLs are not the same as BD-ROMs when it comes to manufacturing. They are BD-Rs not BD-ROMs

8K will do nicely on your 60" to 150" sized displays. They will be selling 4K TVs without any 4K content. The idea is to upscale HD to 4K. Same thiing can be done with 8K until content is made available.

4K is nothing more than a way station on the road to what many feel is the final stop . . . 8K
Edited by Lee Stewart - 9/21/12 at 1:46pm

Except your 6 bit color depth isn't part of the SVH specs. And you keep forgettng the demos were done in the UK. They don't use 24 and 60 Hz due to their 220V power. They use 25 and 50 Hz.

24.9 Gbps is close enough to the stated 24 Gbps in the article.

This si utterly rubbish!

First of all; NHK is a the governmental owned TV station in Japan, like BBC is for the UK.
They don't manufacture and sell equipment.

The newest 8K CMOS Broadcast cameras (one big and one smaller) NHK used at the London Olympics where built by Hitatchi.
RED announced a 9K 645 (medium format) digital film camera in 2008, as a part of their future timeline. Since then they have not mentioned it with a word except saying it will be made at some time in the future.

Broadcast cameras and and digital cinema cameras are two very different technologies. Digital cinema cameras can not be used for broadcast at their native resolution and broadcast cameras can not be used for digital cinema because of the image process in camera. The main 8K camera NHK used in London is a big beast of a camera, totally unusable for modern cinema production.

Pansonic (145" 8K plasma), Sharp (85" 8K LCD) and JVC ("8K" E-Shift Lcos projector) build the displays NHK use in their UHDTV tests.

Try not to post scenarios that you have dreamt up in your head that will only exist there and nowhere else.



8K UHDTV: How do you send a 48Gbps TV signal over terrestrial airwaves?
Edited by coolscan - 9/21/12 at 2:17pm

I'm going to send you here for the second time. While you're there, you can visit here for some extra information about average acuity (by Snellen).

This time, come back with an argument. It would be embarrassing if I had to correct you every time you claim this.

Good idea, nearest neighbor is very sharp but a little harsh.
Here are the same images using nearest neighbor.
For me all the images look effectively identical at the equivalent of 1.5x full picture height and since thats an impractically distance for movie viewing I dont see much point in 8k as a display format.
Video looks a lot better than a still image so small differences will never be noticeable.

View at 10 times image height to simulate full image at 1x picture height.



View at 5 times image height to simulate full image at 1x picture height.

Edited by Owen - 9/21/12 at 8:04pm

Here is some additional specs on NHKs new 8K camera:

http://gizmodo.com/5944270/how-4k-tv-works

http://www.avsforum.com/t/1428971/4k-bluray-seems-closer

As consumers we only care about the delivery format, the pixel count of the capture system is a secondary issue.
MTF and dynamic range are completely separate issues. MTF quantifies the amplitude v frequency response of the system not the overall dynamic range.
All digital capture systems have a falling high frequency response that must drop close to zero at nyquist (spatial frequency equal to the pixel grid) to avoid out of band frequencies entering the system and causing distortion. The low pass filters used (both optical and digital) cannot have an infinite cut off, capturing with many more pixels than is required for final output allows better filter performance and higher in band response when scaling down to the lower output resolution.

Lenses, film and film scanners also have an MTF response (falling high frequency response) and the final output is the sum of all MTF losses in the chain. With digital cameras the additional losses of film are avoided which allows higher mid band MTF which is what defines a sharp picture.
No matter what capture system is used a digital image can only have about 60% of the visible resolution its pixel count would suggest, above that MTF is too low for detail to be visible.

The reason IMAX or high res digital looks sharper than 35mm film when encoded to 2k is because mid frequency MTF (well below 2k) is higher, the limiting resolution is the same in all cases and set by the 2k format.
Since most 35mm film titles are significantly softer than IMAX or high res digital when encoded to 2k its clear the 2k format is not the major limitation for 35mm film source.
"Resolution" and image sharpness are separate issues. A high resolution image with low MTF will look soft while a lower resolution image with high MTF will look sharp. High resolution without high MTF is pointless for movies.

I am all for high res capture systems as they do offer tangible befits, the issue here is 8k v 4k as a display format and I just dont see 8k being worth it for reasonable screen size to viewing distance ratios. A 200" diagonal screen at 8' is not reasonable or practical IMHO.


Here is an example of what falling MTF at high spatial frequencies does to visible resolution. The top half is 100% MTF (perfect), the lower half is a typical representation of the performance of a digital capture system, at the pixel limit (nyquist) MTF drops to about 10%. The limiting "resolution" is the same in both cases as you can see the fine detail if you look closely, but stand back from the screen and the resolution you can actually see changes dramatically.

Remember, most details in the real world dont have 100% relative contrast to begin with, 30% is likely more typical. Factor in 50% or more MTF loss and a 30% contrast difference between details in the original image becomes 15% or less in the video which is useless.



For those who want a better understanding of "MTF" I suggest this tutorial.
http://www.normankoren.com/Tutorials/MTF.html
Edited by Owen - 9/21/12 at 10:00pm

But an 85" 8K diagonal display at 6' is both reasonable and practical.

BTW - thanks for the info. Very informative.
Edited by Lee Stewart - 9/21/12 at 10:02pm

That's only 54 degree viewing angle, barely enough for 4k which is good to 60 degrees plus. 8k needs 120 degrees plus, thats like viewing the 85" from 3'.
85" at 6' is 1.7x image height, if you can see a difference between 4k and 8k at that distance you have exceptional vision. Look at the images I posted above, view the small one at 17 times its height and large one at 8.5 times to get 1.7x effective full picture height performance.

How do you get good sound integration at that distance, even at 10' its difficult.

Glad to be of assistance.
Edited by Owen - 9/21/12 at 11:18pm

For the SHV Olympics demos (that were done in the UK - using NHK's cameras), they used 60 fps.

http://www.engadget.com/2012/07/31/super-hi-vision-eyes-on

IMO, the "recommended" 100 degree viewing angle for 8K is based on very large screens (300+inches). Do I think people are going to sit 3 feet from an 84" 8K display to reach the 100 degree viewing angle? No. I do not. Nor do I believe you will lose seeing the extra resolution 8K provides over 4K by sitting back 6 feet from that size of a display. It's just too short of a distance.

I don't see the problem. Just use smaller "screen channel" speakers aimed a bit towards the "sweet spot" seat. It's only an 84" image. It doesn't need a sound system suited for a 200" image.
Edited by Lee Stewart - 9/21/12 at 10:46pm

OK - thanks for that. So somewhere there is a math error (mcnabney's formula?) because they sure weren't using 3 bit color depth.

Joe - did you see one of the 8K demos during the Olympics?

I'm guessing his calculation assumes RGB colour, and not something like 4:2:0 colour or whatever chroma subsampling they used

edit: here's another article link
http://www.tvtechnology.com/article/broadcasters-show-the-big-picture-from-the-olympics/214799
"Production begins with the acquisition of content on NHK prototype Super Hi-Vision cameras, at 48 Gbps"

(and the BBC article says uncompressed at 24 Gbit/s) No
Edited by Joe Bloggs - 9/22/12 at 1:52am

I corrected my previous post, it now reads.
"That's only 54 degree viewing angle, barely enough for 4k which is good to 60 degrees plus. 8k needs 120 degrees plus, thats like viewing the 85" from 3'.
85" at 6' is 1.7x image height, if you can see a difference between 4k and 8k at that distance you have exceptional vision. Look at the images I posted above, view the small one at 17 times its height and large one at 8.5 times to get 1.7x effective full picture height performance."
Remember, you are very unlikely to ever get video as sharp as a still photo and differences that are visible in stills wont be in video.


I have owned a 70" TV for over 5 years and the sound situation is not as simple as you think.
A much larger screen viewed from 10' or more works much better, thats why I have a projector.
Edited by Owen - 9/21/12 at 11:30pm

That is the first time I have ever heard that 8K requires 120 degrees plus. I am going from this on optimum viewing angles:



And like I said, those viewing angles are not "cast in concrete." They are considered "optoimum" for large displays. Just like there are "optimum" seats in a commerical movie theater - maybe 25% of all seats. The other 75% are either too close, too far or too off-center. But they still sell out theaters don't they?

BTW - there are no Seating Distance Police.

I looked at your images above. Not sure I looked at them properly though. I zoomed my display (20") to 300%. The apartments - the 8K were still viewable with no artifacts while the 4K version was a total mess. All pixelated and blocky. As far as the woman's face, there was little difference between them and IMO, has to do with such a closeup type photo.

I have owned numerous HT setups, all with multichannel sound setups. From 27" all the way up to 144." The biggest mistake I see that many people make is having a sound system that is bigger than their display. If you can't design the sound system to match both the size of the display and sweet spot seat, then you need to do a lot more research.
Edited by Lee Stewart - 9/22/12 at 12:27am

30 degrees is normally considered the correct distance for 1080 (2k) therefore 4k=60 degrees and 8k= 120 degrees. Whoever made that diagram you posted cant add up, double 30 is not 55.
I have tested my vision many times over the years and when viewing test patterns with 100% MTF my eyes conform well with what is considered normal visual acuity. Obviously some people have better vision than others but who wants to spend their time looking at test patterns?
When viewing real video content that will never have 100% MTF at high spatial frequencies no matter how many pixels it uses the limits are lower. In fact the higher we go in resolution the lower MTF becomes as lens performance becomes a limiting factor at very high spatial frequencies.
Remember resolution without high MTF is invisible.

As for the sound issue, I agree its important the sound match the picture, big sound with a small picture is not ideal and the reverse is also true
Having said that the brain can easily tell the difference between an 85" viewed from 6' and a 170" viewed from 12 ', the viewing angle is the same but the viewing experience is not. The same goes for sound, near field listening at 6' is different to 12'.

I would sit 12 from a 170" screen but not 6' from a 85", others may feel differentially, thats their choice.


No you are definitely not looking at them properly. Zooming to 300% is liking looking at you projector screen though a set of binoculars, its silly.
Follow the instruction I gave. Display the images 1:1 (download them and load in photo editor to be sure you have 1:1). Then measure the height of the small image on screen and view the image from 10 times that distance to see what a 8k and 4k image looks like when viewed at 1 times picture height or about 84 degrees viewing angle, closer then just about anyone would ever want to sit.. Remember the full image is 10 times larger and wont fit on just about any PC monitor so a crop is used.

The larger image must be viewed from 5 times its height because it is twice as large. If it wont fit at 1:1 (100%) on the monitor at cant really be used.

Double the above distances if you want to see what 2 times image height looks like (about 42 degrees viewing angle), most people find even that too much for comfort. Its like viewing an 85" screen from 8'.

Anyone who can tell 8k from 4k at 2x picture height (42 degrees) deserves to win a prize.
Edited by Owen - 9/22/12 at 2:57am

First thing's first. Mathematically, you get wrong result if you do this: perfect number of degrees = number of pixels / number of pixels per degree (related to visual acuity). Why? Because display is flat and not a hemisphere where every part of display is located at the same distance from eyes. Please don't do this. Example: 7680/110 = 69.81. Wrong.

Now for acuity. If you are conducting tests yourself, you have to have a sufficient contrast. Acuity depends on contrast too. Also, you have to use nearest neighbor algorithm. Yes, it's harsh but that's the way real lower-resolution display works.

Wrong way of simulating [angular] resolution:

View at 100%.

Right way of simulating [angular] resolution:

View at 100%.

From NHK:


Average acuities of participants are next:


Therefore, NHK concludes that most of the benefit for average consumer will be seen up to [angular resolution of] ~110 pixels per degree. (Not 60 ppd as 20/20 would suggest, by the way 20/20 isn't even average acuity).

Two very important questions:

1. Will effective acuity be higher in the future due to very high-contrast displays making today's calculations obsolete? Very likely.
2. Will results of the tests conducted with real double-density displays yield different results than results we have today (simulated resolution)? Very likely.

Edited by Randomoneh - 9/22/12 at 4:53am

Sorry - the diagram is correct, it is you that is wrong. Right from JBC:



Did you see any of the 2012 Olympics demos of 8K in London or Wash. DC?

And if all you had was an 85" - how far would you sit from it?

I don't see how looking at images on a 20" PC monitor set to 1280x1024 is going to accomplish anything.

The original goal for Super Hi-Vision was that it would have an uncompressed bitrate of 72 Gbps (8K at 60 fps with 12-bit RGB) as seen in this 2010 NHK article but due to technical limits the NHK uses 24 Gbps (my guess is that they use 8-bit 4:2:0 YCbCr). Also the current goal for Super Hi-Vision increased due to the decision to use 120 fps and as seen in this more recent NHK article the uncompressed bitrate is now 144 Gbps (8K at 120 fps with 12-bit RGB).

Agreed, and I have read several articles that said it was at 60 fps. My guess is that the NHK used 8-bit 4:2:0 YCbCr since that would give an uncompressed bitrate of 23.8 Gbps.

A person with 20/20 vision only starts to see the benefit of 8K resolution on a 84" display at a distance of about 6 feet. As such there is a large display size range where a 4K resolution will make more sense for the average person. For example in a BBC London survey the average TV viewing distance was 2.7 meters. At some point in the future I think that 8K UHDTV will make sense as a consumer TV standard but I do wonder whether it will happen in the next 20 years.

The NHK is simply taking into account that for large displays at close distances the center of the display is closer to where you are than the sides of the display.

Fair enough, I did not factor in the super close viewing distance and its affect on distance to screen edges.
It goes to shows that a small difference in viewing distance is critical at such high resolutions. At greater distances that's much less relevant and I seriously doubt many people will be prepared to sit at 0.75 x image height. I like a big image and wide viewing angle but that's insane.
When we have to worry about not being able to resolve detail either side of the screen because its further away then the center we have lost the plot IMHO. Sit a bit further back and we cant resolve any of the screen which is what I have been pointing out.

No, is it better than a still image form a top grade DSLR?
Certainly not the 3' recommended by NHK.


The size and resolution of the monitor is irrelevant, its all about pixels per degree of vision.
If you cant even be bothered to look at the images in the way I requested why comment?
Hopefully others will take more of an interest, they might learn something.
Edited by Owen - 9/22/12 at 7:47pm

Overall screen contrast will improve but the drop in MTF at high spatial frequencies in the video source will remain unless there is a breakthrough in lens technology.
Existing displays have 100% MTF right out to the pixel level so there is no real limitation in that department for displays already.

The 85" prototype|proof-of-concept|"roadshow" 8K4K UHDTV from NHK pictured at this 2011 EBU demo (link) looks to be 5 or 6 feet from the front row seats . . . although placement may be dictated partially by the need to allow second row viewers a clear sight line...?!
_ _
Note also the experimental integrated audio system: This set is designed to provide full 22.2 channel surround sound playback in rooms where installing 22 separate satellite speakers is impractical; instead, it uses 100+ "small" speaker units in a frame around all four sides of the display, plus some proprietary active-DSP|soundbar-style technology, to deliver "the same" playback soundfield to the audience. (It's not clear from the article whether or not there there were also separate subs or some speakers placed in the rear corners of the room.)
_
Edited by SoundChex - 9/22/12 at 11:51am

Hey - didn't you say they weren't going to use 4:2:0 for SHV. That only 4:2:2 or 4:4:4 would be used.

I don't understand why you are showing a recommended seating chart that stops at 2160 when you talking about 8K/4320. The recommended viewing distance from NHK on an 8K 84" TV is .75 picture height. PH = 44" x .75 = 33" - just under 3 feet - not the 6 feet that you claim.

Do you know how minimum perceptible / separable acuity works? If brightness of a point / line is high enough (assuming contrast is high enough) - size / spacing doesn't matter.