Ad
Ad
Ad
Pages: [1] 2 3 »   Bottom of Page
Print
Author Topic: Dynamic range of human visual system  (Read 17852 times)
Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1290



WWW
« on: April 03, 2010, 10:08:16 AM »
ReplyReply


I am looking for some reliable resource (or figures) about the dynamic range of the human vision. Both static (without allowing the eyes adapt) and dynamic (total dynamic range including pupil adaption).

Anyone has some info or figures?

Thanks
Logged

bjanes
Sr. Member
****
Offline Offline

Posts: 2795



« Reply #1 on: April 03, 2010, 02:13:21 PM »
ReplyReply

Quote from: Guillermo Luijk
I am looking for some reliable resource (or figures) about the dynamic range of the human vision. Both static (without allowing the eyes adapt) and dynamic (total dynamic range including pupil adaption).

Anyone has some info or figures?

Thanks
Guillermo,

Do a Google Scholar search with the key words "eye dynamic range" and look at the links to PDF files. There are a number of links that address the problem. Two different sources (e.g. Seetzen & Heidrich: High Dynamic Range Display Systems) state that eye can capture approximately 5 orders of magnitude of dynamic range effectively simultaneously and up to 9 orders of magnitude with adaption. Orders of magnitude are log base 10. 10^5 = 16.6 stops and 10^9  = 29.9 stops.

Regards,

Bill
Logged
Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1290



WWW
« Reply #2 on: April 03, 2010, 03:57:04 PM »
ReplyReply

Thanks, I will keep those figures: 5 orders of magnitude static (16,6EV) and 9 orders adaptive (29,9EV), although I think specially the last could be a bit optimistic depending on how demanding we are about distinguising detail and colour.

I was doing some tests to find out the EV between the max and min luminance (i.e. the maximum displayable DR) on my monitor and on a print copy, and found lower values than some of the figures found in the literature.

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



For the print copy (same pattern displayed on the monitor) I measured a difference of about 4,3EV:



Of course the figures may vary depending on many parameters such as monitor calibration, ambient lighting, type of paper, ink,... but these should be the standard order of magnitude.

In the reference: HDR images we find this table:

• 100,000:1. The ratio of an outdoor scene from shadow to a sunlit area. About 100 orders of magnitude.
• 10,000:1. The ratio for the human eye’s dynamic range is at least an eight order of magnitude luminance range with adaptation and five orders of magnitude of dynamic range without adaptation.
• 300:1 to 500:1. The ratio of an LCD monitor. The ratio is usually referred to as the contrast ratio. About 3 orders of magnitude.
• 300:1. The ratio of a display with sufficient range for medical diagnosis.
• 160:1. The ratio of a printer page from paper white to the black colorant.
• 150:1. The ratio for the HVS local contrast at high spatial frequencies and at small regions where very high contrast cannot be perceived. It is this ratio that allows the HVS to view the dynamic range of the printed page as acceptable.
• 90:1. The ratio of encoding for sRGB is less than two orders of magnitude.

I don't understand very well the matching between the contrast figures on the left, and the orders of magnitude they talk about. E.g. 100,000:1 for 100 orders of magnitude???
For the LCD monitor they say 300:1 to 500:1, i.e. 8,2EV and 9EV, I wonder if they include changing monitor adjustments from its purest black to its brightest achievable white.
They give 160:1 (7,3EV) for the printer page, that seems too much for me.

I was wondering how ambient lighting could affect DR of devices.
  • For instance in a projector, ambient light definitively reduces contrast (that is why lights are to be switched off during projections).
  • The same could apply to a monitor, where blacks could easily become 'washed' because of ambient light while whites do not benefit so much from it.
  • On the contrary I think ambient light shouldn't affect contrast on a printed copy, or it could even benefit of a higher contrast the higher the ambient light is, I don't know.
Just to think a bit about. Regards
« Last Edit: May 30, 2013, 06:33:20 PM by Guillermo Luijk » Logged

bjanes
Sr. Member
****
Offline Offline

Posts: 2795



« Reply #3 on: April 03, 2010, 05:10:11 PM »
ReplyReply

Quote from: Guillermo Luijk
In the reference: HDR images we find this table:

• 100,000:1. The ratio of an outdoor scene from shadow to a sunlit area. About 100 orders of magnitude.
• 10,000:1. The ratio for the human eye’s dynamic range is at least an eight order of magnitude luminance range with adaptation and five orders of magnitude of dynamic range without adaptation.
• 300:1 to 500:1. The ratio of an LCD monitor. The ratio is usually referred to as the contrast ratio. About 3 orders of magnitude.
• 300:1. The ratio of a display with sufficient range for medical diagnosis.
• 160:1. The ratio of a printer page from paper white to the black colorant.
• 150:1. The ratio for the HVS local contrast at high spatial frequencies and at small regions where very high contrast cannot be perceived. It is this ratio that allows the HVS to view the dynamic range of the printed page as acceptable.
• 90:1. The ratio of encoding for sRGB is less than two orders of magnitude.

I don't understand very well the matching between the contrast figures on the left, and the orders of magnitude they talk about. E.g. 100,000:1 for 100 orders of magnitude???
For the LCD monitor they say 300:1 to 500:1, i.e. 8,2EV and 9EV, I wonder if they include changing monitor adjustments from its purest black to its brightest achievable white.
They give 160:1 (7,3EV) for the printer page, that seems too much for me.

Guillermo,

Some interesting observations. Your reference does not use orders of magnitude in the sense they are used in the scientific community, where it is a log 10 scale. 100,000:1 = 5 orders of magnitude. See this Wikipedia link.

Regards,

Bill
Logged
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7503


WWW
« Reply #4 on: April 03, 2010, 06:02:16 PM »
ReplyReply

Hi,

Density rang on paper may be like 2.2 (D-MAX) - 0.05 (DMin), something like 140. Slightly higher DMAX than 2.2 is possible. My interpretation of 1:100000 is also five orders of magnitude, or 16.6 stops.

BR
Erik



Quote from: bjanes
Guillermo,

Some interesting observations. Your reference does not use orders of magnitude in the sense they are used in the scientific community, where it is a log 10 scale. 100,000:1 = 5 orders of magnitude. See this Wikipedia link.

Regards,

Bill
« Last Edit: April 03, 2010, 06:05:15 PM by ErikKaffehr » Logged

Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1290



WWW
« Reply #5 on: April 03, 2010, 06:05:15 PM »
ReplyReply


Yes, from school an 'order of magnitude' has always been x10 for me. But never in a very strict way, just a commonly accepted criteria.
Logged

ErikKaffehr
Sr. Member
****
Online Online

Posts: 7503


WWW
« Reply #6 on: April 03, 2010, 06:09:21 PM »
ReplyReply

Hi,

I agree with what you say. Just want to mention that perceived contrast increases with illumination level. So a well done print can really shine under good (strong) illumination but get murky under less than optimal viewing conditions.

BR
Erik


Quote from: Guillermo Luijk
I was wondering how ambient lighting could affect DR of devices.
  • For instance in a projector, ambient light definitively reduces contrast (that is why lights are to be switched off during projections).
  • The same could apply to a monitor, where blacks could easily become 'washed' because of ambient light while whites do not benefit so much from it.
  • On the contrary I think ambient light shouldn't affect contrast on a printed copy, or it could even benefit of a higher contrast the higher the ambient light is, I don't know.
Just to think a bit about. Regards
Logged

ErikKaffehr
Sr. Member
****
Online Online

Posts: 7503


WWW
« Reply #7 on: April 03, 2010, 06:11:30 PM »
ReplyReply

Hi,

Well I guess that we could also have a "binary order of magnitude", meaning times 2.

Have a good night!

BR
Erik


Quote from: Guillermo Luijk
Yes, from school an 'order of magnitude' has always been x10 for me. But never in a very strict way, just a commonly accepted criteria.
Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8887


« Reply #8 on: April 03, 2010, 08:35:25 PM »
ReplyReply

Quote from: Guillermo Luijk
I was wondering how ambient lighting could affect DR of devices.
  • For instance in a projector, ambient light definitively reduces contrast (that is why lights are to be switched off during projections).
  • The same could apply to a monitor, where blacks could easily become 'washed' because of ambient light while whites do not benefit so much from it.
  • On the contrary I think ambient light shouldn't affect contrast on a printed copy, or it could even benefit of a higher contrast the higher the ambient light is, I don't know.
Just to think a bit about. Regards


Interesting topic. The methods used to calculate contrast ratios for displays is not clear to me. For plasma displays one often sees two sets of figures, one described as the 'native' contrast ratio, and the other the 'dynamic' contrast ratio.

I imagine that the 'native' CR figure, which is the much lower figure, refers to the ratio of the blackest black to the whitest white possible within a single scene, as measured in a dark room with no ambient lighting. The dynamic CR of my Panasonic plasma TV is claimed as being '2 million to 1'. I'd really like to know how such a figure is derived. However, it seems clear that whatever the CR claimed by the manufacturer, the ambient lighting conditions play havoc with the perceived contrast.

I would actually have preferred to have bought an HD projector instead of my 65" Plasma because 65" is a bit small in a large room when viewing 'true' HD material, if you want to see all the detail. Fortunately for the viewer, the average 'so-called' HD broadcast is mush, hardly better than good quality standard definition or an upscaled DVD, and sometimes not even as good as a high quality upscaled DVD. When viewing such material, the further away you are, the better.

An advantage of the projector is, one can use the zoom to change the size of the image according to quality. If the source is true HD, then one can zoom out (or is that in) to fill the entire screen. If the source is crap (technically) one can make the image smaller to hide the defects, without the need in either case to change one's seating position.

The disadvantage of the projector is its low contrast ratio and the consequently greater requirement to view images in a darkened room. This was the main reason I chose a plasma screen in preference to a projector.

However, I now realise that in order to appreciate the full dynamic range of a high quality HD source, viewing either a Blu-ray video of a high quality TV series such as the HBO 'Rome', or one of my own jpeg images, I still need a darkened room, even with the Panasonic plasma with its amazing contrast ratios.

I think perhaps an analogy with audio is relevant here. One can have the most expensive hi fi system that money can buy, however, if one's room acoustics are poor, one is wasting one's money and not deriving the benefit of such exotic equipment.

Room acoustics in relation to hi fi is analagous to ambient lighting in relation to images on the wall, whether transmissive or reflective.
Logged
feppe
Sr. Member
****
Offline Offline

Posts: 2909

Oh this shows up in here!


WWW
« Reply #9 on: April 04, 2010, 03:58:50 AM »
ReplyReply

Quote from: Ray
Interesting topic. The methods used to calculate contrast ratios for displays is not clear to me. For plasma displays one often sees two sets of figures, one described as the 'native' contrast ratio, and the other the 'dynamic' contrast ratio.

Veering off-topic, but dynamic contrast ratio is done by turning off the backlight and measuring a black, then turning it on to full and measuring a white. Current displays don't work this way in real life, in other words dynamic contrast ratio is entirely useless for measuring the real-world contrast ratio of a screen. This is the one usually used by marketing people because it's a more impressive, if implausible, number. More info.

That's only one way you get mislead with numbers. And it doesn't stop there: I was with my parents buying a HDTV for them last Christmas, and they ran HD content on only the most expensive models at the shops. The ones costing half those ran SD content although they were 1080p TVs! So it was impossible to compare the quality side-by-side. And don't get me started on factory presets which crank up the backlight, brightness, contrast and saturation to 11 to make them pop more than the display next to it.

And yes, projector for movie viewing is fantastic. The current generation has enough firepower to view in a dim (not fully dark) room with good contrast ratio. And you can get a 100+ inch screen for much less than the price of a 65" plasma. I have one from the previous gen (Panasonic PT-AX200E) and it looks gorgeous in a darkened room.
Logged

elf
Full Member
***
Offline Offline

Posts: 232


« Reply #10 on: April 04, 2010, 04:02:05 AM »
ReplyReply

Quote from: Guillermo Luijk
Thanks, I will keep those figures: 5 orders of magnitude static (16,6EV) and 9 orders adaptive (29,9EV), although I think specially the last could be a bit optimistic depending on how demanding we are about distinguising detail and colour.

I was doing some tests to find out the EV between the max and min luminance (i.e. the maximum displayable DR) on my monitor and on a print copy, and found lower values than some of the figures found in the literature.

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

   
Logged
Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1290



WWW
« Reply #11 on: April 04, 2010, 06:12:54 AM »
ReplyReply

Quote from: ErikKaffehr
I agree with what you say. Just want to mention that perceived contrast increases with illumination level. So a well done print can really shine under good (strong) illumination but get murky under less than optimal viewing conditions.

Quote from: Ray
The disadvantage of the projector is its low contrast ratio and the consequently greater requirement to view images in a darkened room. This was the main reason I chose a plasma screen in preference to a projector.

However, I now realise that in order to appreciate the full dynamic range of a high quality HD source, viewing either a Blu-ray video of a high quality TV series such as the HBO 'Rome', or one of my own jpeg images, I still need a darkened room, even with the Panasonic plasma with its amazing contrast ratios.
I will put these two comments together just to agree.
  • It is clear that projectors need low ambient light to get a good contrast, everyone has experienced that.
  • It seems the same applies to screen devices such as LCD/plasma monitors. Ambient light reduces contrast in them.
  • For paper print, increasing ilumination level can increase contrast however.
One possible rule to explain the facts would be that self illuminated devices (projector, monitor) get worse with added ambient light. On the contrary devices  that need to reflect an external source of illumination in order to be viewed (print paper in this case) can maximise contrast with an optimised source of light (I guess both the intensity and angle of incidence are important here).


Quote from: feppe
Veering off-topic, but dynamic contrast ratio is done by turning off the backlight and measuring a black, then turning it on to full and measuring a white.
I guess many measures are done that way (specially those from manufacturers interested in giving huge figures), but I think that is not the way to get useful DR figures. In a real situation, we won't change monitor/projector/paper settings to visualize a single picture. That is why I think measuring reflected/projected levels of light from the final image we are really viewing without any change in the settings is the right way, because this will be real the contrast entering our eyes, and hence the contrast we can expect to perceive regularly using that device.
« Last Edit: April 04, 2010, 06:23:20 AM by Guillermo Luijk » Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8887


« Reply #12 on: April 04, 2010, 07:30:09 PM »
ReplyReply

Quote from: Guillermo Luijk
One possible rule to explain the facts would be that self illuminated devices (projector, monitor) get worse with added ambient light. On the contrary devices  that need to reflect an external source of illumination in order to be viewed (print paper in this case) can maximise contrast with an optimised source of light (I guess both the intensity and angle of incidence are important here).

Yes. That's true. In a completely dark room you won't see the print at all. Zero contrast! Yet the monitor will look close to its best   .

But perhaps we should distinguish between contrast and dynamic range. I don't get any sense that shadow detail is lost on a monitor when ambient light levels are increased. The contrast of the scene is simply reduced. One can compensate for this, at least to some extent, by increasing the contrast of the display.

However, when viewing a print, a lowering of ambient light levels actually reduces the perception of dynamic range in the print because shadow detail becomes less discernible.


Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8887


« Reply #13 on: April 04, 2010, 08:05:52 PM »
ReplyReply

Quote from: feppe
Veering off-topic, but dynamic contrast ratio is done by turning off the backlight and measuring a black, then turning it on to full and measuring a white.

I must admit I've never noticed the 'dynamic' CR claims for any LCD. Turning off the backlight certainly does not seem relevant to any real-world viewing experience.

However, the plasma display does not have a backlight. Each of the 2 million pixels on a plasma HD display can be switched off in real time during the display of any scene to indicate total black. That's why plasma displays tend to have much better contrast ratios than LCDs; no backlight. They are a bit like CRTs in this respect.

If one switches off a plasma screen to measure the blackest black, then the result will surely depend on the level of ambient lighting in the room where the measurements are made. If the room is turned into a 'black body' for the purpose of measurement, and such measurements are compared with the brightest level the display is capable of (with contrast controls at their maximum) then the results would surely be astronomical; far greater than 2 million to 1.
 
The 'native' contrast ratio of my Panasonic plasma, which I presume refers to the maximum CR within a single scene at a given instant, is claimed as 40,000:1.
Logged
davidKane
Newbie
*
Offline Offline

Posts: 1


« Reply #14 on: May 27, 2013, 09:02:47 AM »
ReplyReply

Take a look at this paper.

It argues the stead-state dynamic range of the visual system is 3.7 log units.

http://www.cs.bris.ac.uk/~reinhard/papers/kunkel_apgv2010.pdf

DAve
Logged
Jim Kasson
Sr. Member
****
Offline Offline

Posts: 857


WWW
« Reply #15 on: May 27, 2013, 09:53:18 AM »
ReplyReply

...this paper...argues the stead-state dynamic range of the visual system is 3.7 log units.

Thanks, Dave. It is an excellent paper. I like the control of the spatial frequencies of the target, and the use of 1/f noise in the adapting field. For photographers, maybe we should restate the conclusions in log(2) terms, rather than log(10): the steady-state dynamic range of the visual system is 12.3 stops.

Jim
Logged

theguywitha645d
Sr. Member
****
Offline Offline

Posts: 970


« Reply #16 on: May 28, 2013, 10:24:51 AM »
ReplyReply

CCD Arrays, Cameras, and Displays by Holst, 2nd edition, defines display contrast as:

Contrast = Lwhite + Lambient * reflectivity / Lblack + Lambient * reflectivity

where L is luminance.

I am not sure how you are defining contrast, but a standard outdoor scene in daylight was always defined by Kodak as about 160:1. A silver print topped out at about 30:1. Contrast should be logarithmic.
« Last Edit: May 28, 2013, 10:27:52 AM by theguywitha645d » Logged
thierrylegros396
Sr. Member
****
Offline Offline

Posts: 653


« Reply #17 on: May 28, 2013, 10:31:16 AM »
ReplyReply

My personal vision is about 14 Ev static and 28 Ev Dynamic.

But my left eye is 10/10 and my right only 6/10 Wink

Thierry
Logged
BobDavid
Sr. Member
****
Offline Offline

Posts: 1077


« Reply #18 on: May 28, 2013, 01:55:59 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.
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 9170



WWW
« Reply #19 on: May 28, 2013, 02:24:46 PM »
ReplyReply

You might ping Karl Lang, the author of this paper which discusses the subject a bit.

http://wwwimages.adobe.com/www.adobe.com/products/photoshop/family/prophotographer/pdfs/pscs3_renderprint.pdf

http://www.lumita.com
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Pages: [1] 2 3 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad