Some people are still being told that 200:1 is all we can see.... - Page 17

Chris are the 'folks' at Brightside ignorant? The scientific evidence referenced in this very thread, proves beyond any doubt that 100,000:1 is ridiculous!

http://proav.pubdyn.com/Tech_Apps/Ma...rallaxView.htm

You can reach him at pete@hdtvexpert.com. Darin

Please provide a reference for the Brightside info (Thanks in advance). Anyone who made such a statement does not comprehend the complex mechanisms of the HVS, including the obvious contribution/function of the eyes Iris

The has been no scientific reference that says anything of the sort. No, indeed it is you who has displayed astounding ignorance.

Darin: If I understand you correctly, you don't care whether or not you can different levels. So let's hypotheitically create a high CR scene that is projected with say 300:1 and 30,000:1. How many levels of brightness do you think you can resolve under both conditions? The squares test tells me that on my monitor I cannot resolve it from display black. I don't need the 30,000:1 display to create that effect and before Chris kicks in with a display problem, the square is nicely rendered. Are you saying that the 30,000:1 display will help me see that square? Frankly, I don't have the full answer to that. I would assume that the high CR display would extend both black and white levels. So maybe I could actually distinguish the 2% square against the background. But what about 0.5% brightness squares? Or 0.05% squares? It's not really easy to come up with these values and therefore I would love to see some publication that addresses this issue under defined experimental conditions.

It is even more difficult to compare displays, because brightness certainly plays a role. A display with twice the brightness will generate an image that looks more contrasty than a dimmer display even if they were capable of creating the same CR.

The question of what's the maximum CR is very different from what's the minimum luminance difference we can see. I've said this many many times, it's consistent confusion of JND issues with maximum CR issues.

Wow, I'm glad to see there's another person with extraordinary visual capabilities! Are you also capable of functioning at night? That's great! Well, at least there's two of us on earth who can go outside at night. Too bad everyone else can't.

So 2% brightness vs 100% brightness is like the sun? The 2% square simply vanishes on 700:1 (nominal) display. You ought to get out more to see the sunlight

There are 2 issues I see with your experiments as you have posted:

1) they involve motion; you move a mouse pointer, or move your hand for a shadow puppet, you will immediately get luminance adaptation, which can occur in < 100 ms, without any iris adaptation!

http://www.iovs.org/cgi/content/full/42/11/2721
2) Your experiments are not double blind. A lot of times we "see" what we expect to see

We can remove iris adaptation, which takes seconds to minutes from these discussions. You will still get luminance adaptation, in 2 to 100 ms if there is a change in the precept (visual frame).

Here we go again, you're imposing limitations. We're interested in the MAXIMUM, not some other limited example where in that particular situation things may be less than the maximum. As I exaggerated before, most people are capable are moving their eyes around. We don't have them glued in our head and we're not strapped down to a gurney. When you view a single scene, you can move your eyes around.

I suspect that we could come to some reasonable approximation of a top end CR number for video displays if we could start from some established scientific baseline. We seem unable as a group to even do that.

I suggest we start with 2 common "facts" in the HVS community:

1) The retinal ganglion cells can output a luminance range of <200:1 per precept (visual frame). The CW in HVS research is ~100:1.

2) Luminance adaption (contrast gain) can occur in as little as ~1/10 sec.

If we put those numbers together with the video frame rate of 1/60 sec maybe some like Darin can come up with a number that most of us thinks makes sense.

Although that depends on the average light level. For dark scenes it takes tens of seconds to minutes. So a short strobe of light will rest the HVS.

How did the topic of this thread get turned into a discussion about the condition where the eye is capable of seeing a "MAXIMUM CR" ONLY?

Seems to me it is more informative to discuss all the variable parameters that limit CR range.

Obviously, there could be a difference in seeing high CR dark detail in a scene, if the bright area is 3" in diameter and 3' in diameter.

Glen: I think that Darin and Chris have come to realize that they cannot win the simultaneous contrast argument. Far too many references and answers from scientists do indeed state that simultaneous contrast is limited to 100-300:1. So now they are shifting to spatial and temporal contrast aspects which were never in dispute. So I think I'll abandon this thread, since I know what I need to know. If Chris wants to move into face-saving post mode, fine with me. I never had any hopes to convince him anyway.

I think you have this backwards..i.e. persistence of vision would illustrate local luminance adaption is relatively slow

http://www.handprint.com/HP/WCL/color1.html

http://micro.magnet.fsu.edu/primer/l...sionintro.html

There are two types of luminance adaptation mechanisms; contrast adaptation which takes seconds to minuter, and contrast gain which can take < 100 ms.

One thing I am not clear about is whether you consider something being shown at 60 frames per second to be 60 simultaneous frames with each limited to 200:1 without eye movement, or if just the motion in the images themselves negates the thing about simultaneous frames. For instance, the HVS is using motion as one of the clues to make out detail. As an example, if one frame had 250:1 and things outside 200:1 were crushed and the next frame had the exact same levels, except with the brightest and darkest parts shifted laterally, is it your claim that a person would not be able to perceive that motion at both ends. Again, in this case both images have the same CR with the same peak and the same valley, but just in different spots.

It also isn't clear to me if your position is that the ganglion cells could not report anything different for a 200:1 image than if it was changed to 20k:1 by lowering the darkest item. I don't see anything that indicates that the ganglion cells couldn't report something smaller than is coming off the screen. For example, output 100:1 even though the image is 200:1, but would report 200:1 if the image were 20k:1. Do you see anything that makes that clear? If there is 4% washout from a bright object to the darker part of the image, then it sure looks like they could report less than is coming off the screen and still have room to report a higher number if the CR coming off the screen increased significantly.

While it is interesting to consider different cases where we are more limited and less limited, this thread was about the upper limit at once or in one image. If a person says that people can't see anything beyond x:1 and it turns out we can under some conditions (but that there are also conditions where we can't) then they are wrong if they don't specify any conditions and make it out that it is for all conditions. And if they are teaching ISF courses and making people think that images don't need any more than 200:1 (or 300:1 or whatever it is) at once because we couldn't see improvements beyond that, then that is unfortunate.

Again, if a person wants to say that under y conditions we can't see beyond x:1, then that is fine (if it is true). But a test that uses images that limit our vision more than other images do doesn't prove an upper limit and the result should not be claimed as an upper limit. I bet I could come up with images where you couldn't tell the difference between 20:1 and 20k:1, but I wouldn't claim 20:1 as the limit.

--Darin

Uh... since the first post...?

Darin, my intrepretation of the 200:1 statement is related to the overall bright image with dark detail. I have not been able to test a ANSI CR test pattern to determine if with increased brightness (luminance) we have a change in visible CR and it's threshold.

I believe this leads to >>> In a totally light controlled, dedicated theater, you don't need 60+fL off the screen, you want better blacks. It would take too long for the eye to adjust to a dark scene after seeing very bright scenes.

Unfortunately, the statement made here about the comment in the ISF class is not a complete statement, it lacks meaning. We can only assume what the missing pieces are. There has been some agreement in this thread that "in certain situations" the eye may be limited to 200:1. That alone validates the statement, under those conditions. On the other-hand, Chris's comment that we can see more than 1000:1 is valid for his defined conditions, but is not valid for "all" conditions.

As I mentioned before, I have discussed this with somebody who told me he teaches ISF courses. We talked about what image he has used and it was clear to me that he didn't understand that way more than that is needed for simultaneous CR before we couldn't see improvement in simultaneous CR off the screen. That is, before we started talking. After we talked hopefully he understood it more (he seemed to) and went off and did some of the testing I suggested. It would have been much easier to discuss this with a projector in a dark room and some examples, but we live in different states.

Then only for those conditions. Do you agree that "in certain situations" the eye may be limited to 20:1? I could say that validated statements that we are limited to 20:1, under those conditions, but then I should have said it was for those conditions and not just said that humans are limited to 20:1.

I believe this thread started out asking about the upper limit. That is, an "up to" condition. Not that we can see that under every condition, as we are more limited in some cases than others.

--Darin

Glen it's as simple as that..look at a bright scene, your dark adaptation is reset. In fact the afterimage is sill there if you close your eyes. How is one going to appreciate that superior black level if you can't perceive it? The link below illustrates this very limitation!

Dark adaptation
Two phases of recovery: cones fast, rods slow: why?
Bleaching of visual pigment: cones regenerate fast, rods slow
But sensitivity decreases 100 fold for a 10% bleach! Why?

What is the structure that adapts?
Independence of left and right eyes
Rod-cone independence (approximate) means adaptation precedes the retinal ganglion cell (the common path for rod and cone signals)
BUT: in rod vision, adaptation is in a pool (Rushton): bleaching one rod reduces the effectiveness of stimuli for nearby rods. The rods themselves do not adapt, but they send a signal to a later neural pool to provoke sensitivity adjustments there.
Cone vision: no pool, each cone has its own sensitivity control.
Why this makes sense: few photons per rod, many per cone
Evidence for the signal from bleached photoreceptors: pupil constricted in the dark; afterimage; both effects are due to a signal that originates in the retina (pressure blinding).