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

Types of Visual Acuity Testing

a. Minimum Visible Acuity : measures brightness discrimination; the person's ability to detect small differences in the brightness of two light sources. Minimum visible acuity is determined by the brightness of the object relative to its background illumination as opposed to the visual angle subtend by the object.

b. Minimum Perceptible Acuity : measures detection discrimination. Minimum perceptible acuity is concerned with simple detection of objects, not their identification or naming. An example of this type of acuity testing is determining if a child can see and grasp a small candy bead held in the examiner's hand.

c. Minimum Separable Acuity : measures the resolution threshold, or smallest visual angle at which two separate objects can be discriminated. Landolt C, and grating acuity are examples of minimum separable tasks.

d. Vernier Acuity (hyper acuity ): a precise form of visual discrimination still under study. Hyper acuity has been coined to classify the high precision (within a few seconds of arc) with which vernier alignment task can be performed. This level of precision is well above resolution or recognition acuity thresholds.

e. Minimum Legible Acuity: measures the individual's ability to recognize progressively smaller objects (letters, numbers or objects) called optotypes. The angle that the smallest recognized letter or symbol subtends on the retina is a measure of visual acuity. This type of acuity testing is used most often clinically.

f. Snellen Acuity uses a notation in which the numerator is the testing distance (in feet or meters) and the denominator is the distance at which a letter subtends the standard visual angle of 5 minutes. A 20/20 letter (6/6 in meters) subtends an angle of 5 minutes when viewed at 20 feet (6 meters).

http://www.medrounds.org/optics-review/2006/05/31.html

How many megapixels equivalent does the eye have?

The eye is not a single frame snapshot camera. It is more like a video stream. The eye moves rapidly in small angular amounts and continually updates the image in one's brain to "paint" the detail. We also have two eyes, and our brains combine the signals to increase the resolution further. We also typically move our eyes around the scene to gather more information. Because of these factors, the eye plus brain assembles a higher resolution image than possible with the number of photoreceptors in the retina. So the megapixel equivalent numbers below refer to the spatial detail in an image that would be required to show what the human eye could see when you view a scene.

Based on the above data for the resolution of the human eye, let's try a "small" example first. Consider a view in front of you that is 90 degrees by 90 degrees, like looking through an open window at a scene. The number of pixels would be
90 degrees * 60 arc-minutes/degree * 1/0.3 * 90 * 60 * 1/0.3 = 324,000,000 pixels (324 megapixels).
At any one moment, you actually do not perceive that many pixels, but your eye moves around the scene to see all the detail you want. But the human eye really sees a larger field of view, close to 180 degrees. Let's be conservative and use 120 degrees for the field of view. Then we would see
120 * 120 * 60 * 60 / (0.3 * 0.3) = 576 megapixels.
The full angle of human vision would require even more megapixels. This kind of image detail requires A large format camera to record.

http://www.clarkvision.com/articles/eye-resolution.html

Lee, who are you replying to? I am very familiar with both of these.

Just for the record, this:

and this:

If we were to apply this to display / print, it would be only applicable to hemispherical display / print. For your usual flat display / print, calculation is somewhat different.

If that were true then why wasn't it discussed in the Clarkvision link? He did an extremely detailed explaination. I have seen other studies and not one of them said a "hemispherical display/print" would be applicable. Plus human vision is more senstive to perphierial vision than it is looking straight forward.

Turn your logic on, don't mind the actual numbers on these:

Shi*ty illustration but gets my point across:

Hemispherical:


Flat:

RED is too busy selling thousands of 4K cameras. They had 9K prototypes 5 years ago. They will bring those to market when the need arrives. That clear hasn't happened yet.

Hitachi developed typical pre-mastered four-layer BluRay disks five years ago. They also announced that eight layers had been prototyped. They don't sell the media because there is no demand for it yet. 4K will fill that demand.

And I again tell you, the viewing angle is too wide and the seating distance distance is too close. 100 degrees is just crazy. Even THX standards don't recommend exceeding 36 degrees. The eye's central vision (non-peripheral) is relatively small. When the viewing area is massive the eyes have to constantly move to 'see' the entire frame and frequently miss things in the massive glob of out of focus peripheral vision. In addition, looking to the edges of the screen will require a noticeable change in focus. When viewing an image at 30-36 degrees the eyes don't have to move that much, most of the frame stays in the central vision, and the eyes don't need to refocus when looking at the edge of the screen. THAT is why only idiot teenagers sit in the front row of movie theaters. Sure, the image in the front row is huge and immersive, but it isn't good content viewing.

I can guarantee that the Japanese didn't engineer entirely new equipment for the Olympic demonstration. They just don't have that much of it. They took what they already had and adapted the power to match Japanese standards. Besides, someone else pointed out that it was 60 frames - which means 6 bit color - which is fine.

I don't think it was 6 bit colour - I think it was probably 8 bit (per colour channel) colour, (YUV 4:2:0)

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

http://wiki.videolan.org/YUV

YUV 4:2:0 7680*4320*12 bits per pixel*60 frames per second

= 23,887,872,000 bits per second
= 23.89 Gbit/s
Edited by Joe Bloggs - 9/22/12 at 7:19pm

I don't remember making such statements about Super Hi-Vision test equipment and I don't remember saying anything about 4:2:2 being used for Super Hi-Vision. From everything I have read the NHK plan is for Super Hi-Vision to use 12-bit RGB. Do you disagree with that?

I was referring to the point where a person with 20/20 vision starts to see the benefit of 8K resolution and that can be seen using the resolution chart that I linked to.

I certainly don't want a brighter picture, current TV's are far to bright when turned up already. The only improvement in contrast many of us want is much darker blacks which is not going to help much with perceptible / separable acuity.

With the inherent MTF losses in video maximum adjacent pixel contrast is very low, a tiny fraction of full screen contrast, the smaller the details the less brightness and less contrast they have. This is the exact opposite of what the eye requires to see fine detail.

Maximum viewing distance calculations are based on 100% MTF test patterns at high brightness which is totally unrealistic and completely impossible with video or photographs. We need to sit much closer to see the low contrast fine detail that video actually provides.
Edited by Owen - 9/22/12 at 8:38pm

Thousands of 4K cameras? Hardly

Both were BD-Rs which are not used for prerecorded movies (they use BD-ROMs). The 4 layer became the BDXL. The 8 layer is nothing more than a lab experiment.

Well - that is false, They showed their new 8K camera

+1, some words of reason in this crazy discussion.

No - I don't agree with that. You have decided to add a caveat here that you never added before . . . "test equipment."

Again, that chart doesn't even show 8K so why are you using it? This one does:

I think that it matters a great deal that it is Super Hi-Vision test equipment. Did you think that Super Hi-Vision test equipment would be equal to what is currently planned for Super Hi-Vision? Also I would mention that my posts on Super Hi-Vision were in regards to what the NHK is currently planning and that those plans might change in the future.

It is a nice large chart and since it shows the line where someone with 20/20 vision can fully benefit from 4K resolution it also shows where someone with 20/20 vision can start to benefit from 8K resolution. As such I consider it relevant.

What level of visual acuity was used for that chart?

Can I assume I'm not the only one who can see how ridiculously inaccurate the below "chart" is?
Supposedly we can see 4k on an 85" screen at 37' and 8k at 18', we cant even see 2k on a 120" screen at that distance.


Edited by Owen - 9/22/12 at 9:59pm

The only spec that NHK recently changed is to go to 120 fps.

Guess we agree to disagree.

No idea - just found it on the Net.

And yet Richard's chart says you can see 4K from 15 feet on a 125" screen, which you yourself say is impossible

That's because it is impossible.
The chart clearly states that for a 125" screen a viewing distance of 16' is required to fully resolve 2k and 8' for 4k.

The calculations are based on 20/20 vision AND 100% MTF high brightness test patterns which is totally unrelated to video which has about 10% MTF at the pixel level.
Unless you like watching PC generated test patterns with 100% MTF you will need to sit significantly closer to see fine detail in video.

I draw your attention to pages 17 and 18 - it is from 2006:

http://www.hpaonline.com/assets/documents/HPA2006_pre.pdf

A better version - page 8:

http://144.206.159.178/ft/CONF/16427879/16427890.pdf

But the chart says you BEGIN to see the benefits of 4K from 15' on a 125" screen which again you say is impossible. So how accurate is the chart?

In a perfect world with 100% MTF test patterns you “might” be able to see an advantage in 8k at 15” on a 125” screen, but we don’t live in a perfect world, far from it.

I draw peoples attention to the document you linked.

http://www.hpaonline.com/assets/documents/HPA2006_pre.pdf

Look at the page entitled “MTF Characteristics (Measured)”. This shows the theoretical and actual performance of a typical 8k camera. The measurement is in TV lines so its a vertical resolution and for this camera the limit is 4320 pixels. Note that this is for the green channel with better performance than blue and red.

As you can see MTF drops to less than 10% at 4320 (the pixel level) and by definition MTF will also be less than 10% at 7680 horizontal. This means the system has effectively ZERO response at 8k.

30% MTF is considered the usable minimum for cinema and television applications and the system can only resolve about 3910x2200 at 30% MTF. Remember this is for an 8k system not 4k, a 4k system will deliver a little better than half that.

Resolution with 30% MTF is very blurred so only details with high contrast will be visible, low contrast detail will simply disappear, and since calculations used to work out viewing distances for various resolution are based on 100% MTF high contrast test patterns the predictions are very optimistic when it comes to real video.

As I have said before, pixels are not resolution and resolution without high MTF is invisible.

To understand visible resolution you must understand MTF and its affects.

An 8k system can only deliver weakly resolved 4k.

@Lee - that chart is based on 200 pixels per degree - best acuity in NHK tests.

Owen, don't tell me you're not familiar with tests done by NHK. They haven't used computer generated images but images of nature. Conclusion was: best acuity was ~200 pixels per degree. For most of participants that was ~110-120 ppd.

I read it differently. They measured the resolution taking into account MTF and came up with 3200 TV LInes. To put that in perspective:

VHS = 240 TV Lines
LD = 400
DVD = 480 (enhanced for 16x9 TVs)
ED-Beta = 500
35mm Print (1.85 AR) = 850*
1080 HD = 900
2160 4K = 1500
4320 8K = 3200
IMAX 15/70 Print = 4700*

*
http://tristanpipo.com/blog/2005/3/18/last-update-was-sunday-imax-resolution.html
Edited by Lee Stewart - 9/23/12 at 11:15am

That chart is pure fantasy. I am looking at a 63" screen right now and my eyes are 9' away form it. It is 1:1 pixel mapped. I just put up two different backgrounds, one horizontal and one vertical with interlaced 1 pixel lines on a white background. At my normal viewing distance I can't clearly see the distinct lines, but leaning forward a foot (to make it 8' away) brings them into clarity. Your chart thinks that I can distinguish 1080p at 26'. Ridiculous! The other chart that doesn't show 8K is spot on.

Please wait for actual 2160p and 4320p panels for comparisons. If you want more relaxed chart, for those with normal healthy vision, that would be 110-120 pixels per degree:

I just love this. This shows when saturation occurs. For 8K, saturation occurs @ about 120 pixels per degree (8K at 60 degrees viewing angle).

About charts - we have those with extremely low numbers (60 pixels per degree, Carlton Bale), those with extremely high numbers (400 pixels per degree) and this with 120 pixels per degree which seems to be when saturation starts to occur. Yeah, for displays with extremely high contrast & brightness, 300, 400 ppd might apply but for what we have today, I believe 120 ppd is sufficient.

Here is a link to a NHK study in which they did visual acuity tests. In section 3.2.1 "Tests of visual acuity using natural images" the test involved natural images being shown on a flat panel display. The test was to correctly distinguish an image at various resolutions with the original image (at 312 pixels per degree). At about 100 pixels per degree with high detail images the subjects were correct about 90% of the time. At 156 pixels per degree with high detail images they were correct less than 70% of the time. The researchers noted that the ability to correctly distinguish the image was close to the average visual acuity of the group that was tested.

That was in section 3.2.2 "Comparing realness between real objects and images at various resolutions" of the NHK study and it shows that the vast majority of the perceived benefit was reached at about 100 pixels per degree which was close to the average visual acuity of the group that was tested. The researchers said "The results indicated that realness of an image increased as the image resolution increased up to about 40-50 cpd (Figure 10), which corresponded to the observers’ minimum separable acuity, and reached a plateau above this threshold."

That NHK document is from 2006 and the NHK camera used four 8 megapixel CMOS sensors capable of 60 fps. As seen in this 2012 NHK article the NHK now has a 33 megapixel CMOS sensor capable of 120 fps. In my opinion this is a short term issue due to the technological advancement of CMOS sensors.

Where did you hear that 30% MTF was the usable minimum? Also what matters is the camera system that is used to record the video. For example the RED Dragon CMOS sensor is 18 megapixels and the Sony F65 CMOS sensor is 20 megapixels.

That is a strong statement to make based on a NHK camera from 2006. From what I have read up until recently the CMOS sensors in 4K cameras used for Hollywood movies, such as the Red Epic, use interpolation for all of the subpixels. CMOS sensors though are increasing in resolution and the recent Sony F65 is 4K for the green subpixels and uses interpolation for the blue and red subpixels (the first movie that uses the Sony F65 will be After Earth).
Edited by Richard Paul - 9/23/12 at 2:45pm

Any quoted "resolution" that does not specify MTF at that resolution is misleading and tells you nothing.
The graph shows that at 3200 vertical the system has only 20% MTF which is useless, the industry standard is 30% and even that's being generous.
At 30% MTF the camera is only good for about 2200 vertical which is about half 8k for an 8k system. Only very high contrast details have a chance of being visible at 30% MTF, anything less will just be a blur to the eye.

This is an entirely normal and predictable response for an 8k digital camera, the theoretical ideal response is only slightly better and is also shown on the diagram.

MTF tests are done under ideal conditions with everything perfect, real world performance is rarely ever going to be that good. Even slightly imperfect focus due to focus error or depth of field and motion blur result in much less resolution than a perfect still shot under ideal conditions with no motion.