Ad
Ad
Ad
Pages: « 1 2 [3]   Bottom of Page
Print
Author Topic: Dynamic range of human visual system  (Read 12591 times)
BJL
Sr. Member
****
Offline Offline

Posts: 5082


« Reply #40 on: May 29, 2013, 09:10:36 AM »
ReplyReply

Bart,
    Thanks for the explanation of how the latest glossy papers can have albedo (diffuse reflectivity) as low as 1/200, about eight times lower than than any natural substance. (This also explains why matte papers have far less impressive Dmax.)

But glossy surfaces can reach a much higher D-max, because less light is scattered (also back to the observer) and more is reflected away from the observer.

Anyway, whatever the details it is clear that no print paper come close to matching the subject brightness range [SBR] that can often occur in natural scenes when there is significant variation in incident light levels. To which I will add that no part of a displayed image (on print or screen) gets close to matching the luminosity of the brightly lit parts of the outdoor scenes that produce those examples of high SBR, so any attempt to reproduce the contrast range of such a scene in a displayed image must have the maximum brightness far lower than it was in the actual scene, so that the shadows are also far darker, and so will disappear into the murk. Also, I am fairly sure that only the static contrast range of the eye is relevant to viewing displayed images, not the far greater dynamic range available when we move our eyes from sunlit to deeply shaded parts of a scene. This partly relates back to the fact that no part of a displayed image will be bright enough to cause our irisis to close down as much as they do in bright sunlight, or to cause the chemical adjustments to "lower ocular sensitivity".
Logged
BJL
Sr. Member
****
Offline Offline

Posts: 5082


« Reply #41 on: May 29, 2013, 09:21:09 AM »
ReplyReply

Since you're using a Mac, simply paste 10 ^ (2.20 - 0.05) in the spotlight field and you'll get the result.
Or put it into the Mac OS app called "Calculator", which is where Spotlight takes you.

Going even further OT, it is strange to me that some people use complicated software that costs actual money like Excel for tasks that I am used to handling with basic tools included with the OS.

By the way, is there a reliable source for a print paper Dmin value, like the Dmin=0.05 used here?
« Last Edit: May 29, 2013, 09:24:43 AM by BJL » Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3013


« Reply #42 on: May 29, 2013, 10:36:18 AM »
ReplyReply

By the way, is there a reliable source for a print paper Dmin value, like the Dmin=0.05 used here?

Hi,

That would require a calibrated densitometer, or a spectrophotometer.

You could also use the data from e.g. Ernst Dinkla's SpectrumViz application, and take an average reflection % of paper white and calculate:
Optical Density = - Log10(reflection%), e.g. 90% reflection is - Log10(0.90) = 0.046 or 4.6%.

Cheers,
Bart
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #43 on: May 29, 2013, 11:14:22 AM »
ReplyReply

Or put it into the Mac OS app called "Calculator", which is where Spotlight takes you.

Going even further OT, it is strange to me that some people use complicated software that costs actual money like Excel for tasks that I am used to handling with basic tools included with the OS.

By the way, is there a reliable source for a print paper Dmin value, like the Dmin=0.05 used here?

I already have Excel and it is pinned to my Windows task bar for easy and rapid access. Since I have already paid for it, why not use it? BTW, if I enter 10^2.4 into google search, a calculator also pops up with the result. Windows also has a calculator with the needed function.

Bill
« Last Edit: May 29, 2013, 11:18:54 AM by bjanes » Logged
BJL
Sr. Member
****
Offline Offline

Posts: 5082


« Reply #44 on: May 29, 2013, 11:56:35 AM »
ReplyReply

Bill, my comment was not a Mac vs Windows thing; I just used the Mac OS example because that is what had already been mentioned, and what I have in front of me at the moment.

It was a matter of my preference for using a simple, free tool that starts instantly and has an interface optimized for the simple task at hand over using a complex (and for many of us, costly) "kitchen sink" tool that opens more slowly (for those who do not have it constantly open) due to all the unneeded "plumbing". Sort of like when people use bloatware like Word to produce simple notes and memos that for me are more quickly and easily handled by nice little "free" tools like Mac's Textedit or Windows' Notepad(?). Or using the currently notorious PhotoShop when alternatives like Lightroom or Aperture or whatever deal with most photographers' needs better and cheaper.
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5255


WWW
« Reply #45 on: May 29, 2013, 12:30:00 PM »
ReplyReply

Since you're using a Mac, simply paste 10 ^ (2.20 - 0.05) in the spotlight field and you'll get the result.

That is way cool!!! Many thanks.
Logged
theguywitha645d
Sr. Member
****
Offline Offline

Posts: 970


« Reply #46 on: May 29, 2013, 03:36:52 PM »
ReplyReply

You do not need a log scale to calculate contrast ratios. Density is log based. If you want perceptual values, you should use L* or at least gamma 2.2, which can easily be done with Bruce Lindbloom's Companding Calculator.

You should recheck your texts, as I think your recalled values are way off as demonstrated in my previous post using Ilford photographic paper. Your outdoor scene contrast ratio is way off. Karl Lang is a serious guy who knows his stuff.

Bill

From Photographic Materials and Processes by Les Stroebel, John Compton Ira Current, and Todd Zakia (all former teachers, BTW), on page 359 clearly state the average scene has a luminance ratio of 160:1. They supply data where the White clouds come in at 1,114 candela/sq ft or 3,500 footlambert all the way down to the base of a tree in heavy shade at 12 and 38 respectively. I would say these authors know their stuff as well. Zakia in his book Photographic Sensitometry: The Study of Tone Reproduction on page 7 list the average scene has a luminance range of 160X.

It might be worth noting that scenes don't have density.
« Last Edit: May 29, 2013, 03:41:15 PM by theguywitha645d » Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #47 on: May 29, 2013, 05:02:20 PM »
ReplyReply

From Photographic Materials and Processes by Les Stroebel, John Compton Ira Current, and Todd Zakia (all former teachers, BTW), on page 359 clearly state the average scene has a luminance ratio of 160:1. They supply data where the White clouds come in at 1,114 candela/sq ft or 3,500 footlambert all the way down to the base of a tree in heavy shade at 12 and 38 respectively. I would say these authors know their stuff as well. Zakia in his book Photographic Sensitometry: The Study of Tone Reproduction on page 7 list the average scene has a luminance range of 160X.

It might be worth noting that scenes don't have density.

I don't know about the average scene, but many scenes have a much higher luminance ratio, or else we wouldn't need to use HDR techniques with our D800s or your Pentax MF digital. 160:1 is only 7.2 stops. Look at the examples in Karl's treatise. You are correct about scenes not having density. I got side tracked from the discussion of the print.

Bill
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1615


« Reply #48 on: May 29, 2013, 05:45:17 PM »
ReplyReply

Does it make any sense to talk about "scene DR" without mentioning patch sizes? Are we talking about the brightest/darkest 1% of the scene? 0.00001%? Surely, many landscapes must contain some small birds holes in a tree off to one corner where the luminance is minimal compared to highlights or the sun?

-h
Logged
Jim Kasson
Sr. Member
****
Offline Offline

Posts: 538


WWW
« Reply #49 on: May 29, 2013, 05:55:24 PM »
ReplyReply

Does it make any sense to talk about "scene DR" without mentioning patch sizes? Are we talking about the brightest/darkest 1% of the scene? 0.00001%? Surely, many landscapes must contain some small birds holes in a tree off to one corner where the luminance is minimal compared to highlights or the sun?

Good point. How long is the coastline of England? Depends on the scale. Lens flare, diffraction, etc will fill in a lot of the small holes, but that's not a function of the scene itself. I guess the bottom scene-referred scale is on the order of the wavelength of blue light.

Jim
Logged

BJL
Sr. Member
****
Offline Offline

Posts: 5082


« Reply #50 on: May 29, 2013, 06:04:14 PM »
ReplyReply

From Photographic Materials and Processes by Les Stroebel, John Compton Ira Current, and Todd Zakia (all former teachers, BTW), on page 359 clearly state the average scene has a luminance ratio of 160:1.
That seems reasonable for a typical scene with clear sunlight and significant shadows:
- the reflectivity of natural materials varies over a range of about 20:1 (from 4% for fine charcoal and fresh asphalt to 80-90% for fresh snow) http://en.wikipedia.org/wiki/Albedo
and
- there is an EV range of about three stops from a typical scene in bright sunlight (EV15) to open shadows in a scene with clear sunlight (EV12) for another factor of 2^3 = 8  http://en.wikipedia.org/wiki/Exposure_value
together getting to 160:1.

The 100,000:1 extreme case mentioned elsewhere seems to require a variation in incident light of about 5,000:1, or 12 stops (12 EV).
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1615


« Reply #51 on: May 30, 2013, 04:19:16 AM »
ReplyReply

Good point. How long is the coastline of England? Depends on the scale. Lens flare, diffraction, etc will fill in a lot of the small holes, but that's not a function of the scene itself. I guess the bottom scene-referred scale is on the order of the wavelength of blue light.

Jim
If you want to be really nerdy, what is the probability that a really dark and small spot will emit zero photons in our way within a given time-window, and what is the DR of X/0 ?:-)

But perhaps the discrete nature of light should be seen as "noise" in this case, what we really are interested in is the linear reflectance*illumination?

-h
Logged
theguywitha645d
Sr. Member
****
Offline Offline

Posts: 970


« Reply #52 on: May 30, 2013, 10:18:03 AM »
ReplyReply

I don't know about the average scene, but many scenes have a much higher luminance ratio, or else we wouldn't need to use HDR techniques with our D800s or your Pentax MF digital.

I never use HDR techniques with either my D800 or Pentax 645D. The need for these techniques is not that common. I have found most folks that use HDR extensively either have trouble controlling their process or simply like the look.
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #53 on: May 30, 2013, 10:27:13 AM »
ReplyReply

I never use HDR techniques with either my D800 or Pentax 645D. The need for these techniques is not that common. I have found most folks that use HDR extensively either have trouble controlling their process or simply like the look.

I don't have the luxury of a 645D, but I too rarely feel the need for HDR with my D800e landscape images. Their DR is less than I had thought as shown by the preceding discussion. Thanks for your input. Erik Kaffer often states that the need for very high DR is overblown for most imaging and I would agree.

Bill
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1615


« Reply #54 on: May 30, 2013, 01:07:20 PM »
ReplyReply

I never use HDR techniques with either my D800 or Pentax 645D. The need for these techniques is not that common. I have found most folks that use HDR extensively either have trouble controlling their process or simply like the look.
I like to talk about these two separately:
1) High dynamic-range capture. Possible by exposure stacking or single shots from very good sensors
2) Tone-mapping to map HDR to LDR displays/print

HDR capture does not have a "look" to it. It is about recording the scene more accurately. If anything, the highlight clipping and shadow noise of regular sensors is a "look".

Tonemapping certainly can have a distinct "look", but that look should also be possible when processing good sensors in modern raw developers.

-h
Logged
mouse
Full Member
***
Offline Offline

Posts: 136


« Reply #55 on: May 30, 2013, 04:19:57 PM »
ReplyReply

Age of the human factors into the equation. I've read that human eyesight decreases in brightness by one f/stop every ten years.

My guess is that not a result of diminished sensitivity of the "sensor" but rather increased attenuation of light by the intervening elements.
Logged
BJL
Sr. Member
****
Offline Offline

Posts: 5082


« Reply #56 on: May 30, 2013, 07:44:45 PM »
ReplyReply

My guess is that not a result of diminished sensitivity of the "sensor" but rather increased attenuation of light by the intervening elements.
Also, our irises lose its ability to open as far in low light, so we lose some of the low end of our dynamic range. (I use the first person, because I am experiencing this process, as I suspect are many of this forum's participants.)
Logged
250swb
Full Member
***
Offline Offline

Posts: 210


« Reply #57 on: June 02, 2013, 03:57:29 PM »
ReplyReply

When it comes to 'reality' there is a divergence between scientific tests of what an eye, or a camera, records. And the difference is in what the brain remembers.

The accuracy of a photograph to represent the moment it was made has little to do with dynamic range of the eye, or the DR of the camera. While the eye has a vast DR, the memory for a scene is much less discerning. The memory remembers the bright light that made you squint, it remembers dark corners that made you peer, but it doesn't do it all in one go, because the eye is never adapted across a wide DR scene all in one go. So an accurate photograph, one that 'remembers' the scene as the eye saw it has a smaller dynamic range. Which is where modern digital cameras start to render reality as it appears to the camera, not the photographer. In a scene, a sunset perhaps, we shield the eyes, we look down, we look up, we don't take all the detail of a wide DR scene in the one hit, unlike the camera, so while we remember the glare, or the deep shadows as the light falls, it is done on separate levels. The camera renders them on one level, especially if some mild HDR is done. So it is a moot point if a very high DR camera is any good at representing a human connection to a scene because just like the high pitched sounds a dog can hear, we can't see what the camera can see even if the overall DR of the eye is as wide as the camera's.

The danger of course is that photographers chase the things technology can give and forget that their photographs are starting to bear no resemblance to what the viewer might expect. An extreme HDR photograph may elicit the question 'that can't be right can it?', while a clearly artistic rendering will be accepted for what it is, as will B&W, they are clearly not reality. But even on a lower subliminal level high DR of modern cameras makes the brain distrust the image if the intent of the image is the accurate recording of reality, and reality is entirely what the brain remembers or expects, not how a camera records it. But that is a bigger philosophical discussion.

Steve
Logged

Iliah
Sr. Member
****
Offline Offline

Posts: 369


« Reply #58 on: June 03, 2013, 12:04:51 PM »
ReplyReply


> 5 orders of magnitude static (16,6EV) and 9 orders adaptive (29,9EV)

1000 and 101 technically are of the same order of magnitude. If the result of division is less than 10, numbers are of the same order of magnitude. This may cause quite a difference in EV, like subtracting up to 3 stops from the result.

> For my monitor (I shot a black and white pattern) I found a difference between pure white and pure black of about 6,7EV

Have you tried to check the number using exposure meter, and measuring from the screen off to have reference, filled with just black, and filled with just white, to avoid additional flare?
Logged
joneil
Full Member
***
Offline Offline

Posts: 114


« Reply #59 on: June 04, 2013, 08:59:19 AM »
ReplyReply

  An article form the UK Guardian I was interviewed for - over ten years ago now, that might be of interest to this topic:
http://www.guardian.co.uk/science/2001/feb/01/technology2

  Deals more with the range of human colour vision than the actual dynamic range of human vision, but I think it ties in, especailly since colour can influence how the "brain" sees brightness.

  The range of human vision varies greatly.
Logged
Pages: « 1 2 [3]   Top of Page
Print
Jump to:  

Ad
Ad
Ad