 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Medical Forum / General / Vision / January 2007Resolution limit of the visual system
Hi, When I look at the contrast sensitivity function (i.e. http://vision.psy.mq.edu.au/~peterw/csf.html or http://webvision.med.utah.edu/KallSpatial.html) I find the spatial frequency to peak at about 60 cycles per degree at high contrast. But on some websites I also find calculations based on the visual acuity of 1 arcminte that lead to a max spatial frequency of 6,88 lp/mm like this: 1 arcminute = 0,000291 rad; 0,000291 rad*250 mm viewing distance = 0,0727 mm smalles object size; 0,0727 mm*2 = 0,1454 mm/lp; 1 mm/0,1454 mm = 6,88 lp/mm at 25 cm distance If I calculate the lp/mm at the same distance based upon 60 cycles per degree I get the result of 13,74 lp/mm: 1/250*(180*pi) = 0,229° 60 cycles per degree*0,229°=13,74 lp/mm Can somebody please tell me which value is correct or if there is something that I do just not see? Best regards! Joseph Thomas I´m not sure but the value of 60 cycles per degree (13,75 lp/mm) might be related to contrast situations that are higher than you would expect for average situations. So 30 cycles per degree (6,88lp/mm) should be a more practical value that resembles what we encounter in nromal life. But maybe someone who really knows about this stuff can say whether this is true or not. Marc Wossner > Hi, > [quoted text clipped - 5 lines] > 1 arcminte that lead to a max spatial frequency of 6,88 lp/mm like > this: Spatial frequency is another way of specifying visual acuity; spatial frequency defines it in terms of cycles per degree whereas the arc minute specifies the size of the critical detail of the target. One arc minute is an arbitrary number that specifies the Snellen 20/20 target, it is not a measure of maximum attainable acuity. Targets that are smaller than 20/20 are visible to many humans, the maximum Snellen is around 20/10, some claim 20/8. In the contrast sensitivity studies, frequency was increased until the gratings could no longer be detected so the maximum is not 20/20 (the Snellen equivalent of 60 cycles is better than 20/20). If you want to compare 1 arc minute to spatial frequency, you need to find out what the spacial frequency of a a 20/20 grating is and compare that value to 1 arc minute. I don't know what the arc minute value of 20/8, but if you convert 60 cycles/degree to arc minute I suspect you will find a value very close to the arc minute value of 20/8. Judy > Dr Judy wrote: > > Joseph Thomas wrote: > > Hi, [quoted text clipped - 12 lines] > arc minute is an arbitrary number that specifies the Snellen 20/20 > target, it is not a measure of maximum attainable acuity. > Targets that are smaller than 20/20 are visible to many humans, the > maximum Snellen is around 20/10, some claim 20/8. In the contrast [quoted text clipped - 4 lines] > spacial frequency of a 20/20 grating is and compare that value to 1 > arc minute. > I don't know what the arc minute value of 20/8, but if you convert 60 > cycles/degree to arc minute I suspect you will find a value very close > to the arc minute value of 20/8. Well I?m a bit confused now. From what I read so far I thought the 1 arcminute would be the smallest separation between two points under which they are perceived as separate by an average human with 20/20 vision. And by those means I thought 20/20 is the max acuity of this average human. That?s why I was looking for a link between this value and the 60 cycles/degree derived by contrast sensitivity studies. But how can someone with 20/20 vision resolve better than 1 arcminute? - Even more confusion! What do I overlook? In the end I?d like to know how many cycles the average observer with 20/20 vision can resolve. Joseph Thomas >how can someone with 20/20 vision resolve better than 1 arcminute? - >Even more confusion! What do I overlook? Contrast? Consider a star. The Airy disc of a star is somewhere around 20 arcseconds in size to the human eye. That's the same for every star, because all stars are point sources from this distance. That's just a third of an arcminute, yet the contrast between the diffraction disc and surrounding sky make bright stars plainly visible. The same effect can be seen with bright lines on a dark background and dark lines on a bright background. As the contrast between the line and background increases, the width of the line necessary for detection goes down. How much of this is the result of optics, and how much the result of processing in the brain, is another question.  Signature- Mike Ignore the Python in me to send e-mail. >Mike Ruskai wrote: > >On or about 16 Jan 2007 15:56:56 -0800 did "Joseph Thomas" > > <joseph.thomas@gmx.net> dribble thusly: > >how can someone with 20/20 vision resolve better than 1 arcminute? - > >Even more confusion! What do I overlook? > Contrast? So ist it really the case that a) the contrast sensitivity function has been measured with people who have average 20/20 vision and that b) the ability to resolve cycles per degree rises as the contrast rises but that 60 cycles/degree are only possible in situations where contrast is unreasonably high, as Marc statet further up? Joseph Thomas Dear Joseph, Subject: Contrast versus two points of light. Re: Consider a star. The Airy disc of a star is somewhere around 20 arcseconds in size to the human eye. That's the same for every star, because all stars are point sources from this distance. That's just a third of an arcminute, yet the contrast between the diffraction disc and surrounding sky make bright stars plainly visible. It was ASSUMED that the AVERAGE human could resolve two points of light separated by 1 minute of arc. But this is a gaussian distribution of the ability of an eye "fully corrected" to separage two points of light. This resolution is not necessarily "contrast". This 1 minute of arc resolution was taken by Snellen as the AVEAGE visual acuity that a large percentage of eyes could have with a "corrective" lens. As Judy stated, some people have the resolution of 1/2 minute of arc, or 20/10 vision -- but that is exceptional. (Part of the "tail" of a gaussian distribution.) The Snellen letters are 5 minute-of-arc wide, with the assumption that this is verifies 1 minute of arc resolution. Best, Otis > >Mike Ruskai wrote: > > >On or about 16 Jan 2007 15:56:56 -0800 did "Joseph Thomas" [quoted text clipped - 12 lines] > > Joseph Thomas I just found a formula to convert between Snellen and spatial frequency: Snellen denominator = 600/cycles per degree or cycles per degree = 600/Snellen denominator so 600/60 cycles/degree = 10, Snellen = 20/10 or 600/20 = 30 cycles per degree So is it safe to state that the resolution ability of an average human with 20/20 vision is 30 cycles/degree or 6,88 lp/mm? If it?s so I do still not understand where the 60 cycles/degree of the contrast sensitivity function come from. Joseph Thomas > I just found a formula to convert between Snellen and spatial > frequency: [quoted text clipped - 11 lines] > still not understand where the 60 cycles/degree of the contrast > sensitivity function come from. The two measures are measuring different things. First of all, CSF only asks the observer to detect a black and white grid, while the Snellen requires recognition and identification of a letter, a task that is more difficult to do and involves different areas of the brain. Secondly, the average best corrected Snellen acuity is 20/20; however, as an "average", it means that about 75% of people with healthy eyes have a best acuity of between 20/15 and 20/25 and about 98% would be between 20/10 and 20/30. So the "average" is 20/ 20 but some people see better. The constrast sensitivity function is the maximum seen by any observer, so it's maximum value is equivalent to about 20/10. The two measures are not really comparable and are useful for different reasons. If you are designing a medication bottle label and want to make sure most people can read it, you would use a Snellen size of at least 20/30. If you are designing something with a grid and want to make sure most people don't "see" the lines, then you would look at CSF and make sure the cycles/degree is greater than 60. Dr Judy Well I think I got it by now. I was on a completely wrong track because I assumed that the csf would be for an average human with 20/20 vision. - I could have known better as this is nowhere stated on all the websites I browsed. But sometimes you make presumptions and stick with them without further thinking. That´s why all your words got lost in the dark. - Thanks for your patience! Joseph Thomas > Well I think I got it by now. I was on a completely wrong track because > I assumed that the csf would be for an average human with 20/20 vision. > - I could have known better as this is nowhere stated on all the > websites I browsed. But sometimes you make presumptions and stick with > them without further thinking. That´s why all your words got lost in > the dark. - Thanks for your patience! I think the root of your confusion lies with the common assumption that 20/20 is "normal" or "average" vision. Many people, and most people without eye disease, achieve better than 20/20, 20/15 is likely closer to what would be "normal". The CSF you looked at may have been average CSF with average Snellen of 20/20. However, do not confuse average acuity with "average person". There is no "average person" but there are average results of measuring many persons. A actual person with an actual best vision of Snellen 20/20 may have a max CSF of 60 cycles. This is because the two tests do not measure the same thing, detecting a line is a different cognitive task that identifying a letter. Dr Judy > Joseph Thomas > Dr Judy wrote: > > Well I think I got it by now. I was on a completely wrong track because > > I assumed that the csf would be for an average human with 20/20 vision. [quoted text clipped - 17 lines] > same thing, detecting a line is a different cognitive task that > identifying a letter. Well, it looks like if I would have to acknowledge that I´m still a bit off track. So I will try to get deeper into those detection and recognition tasks to gain a more fundamental understanding. Thanks for so far, Dr. Judy! Joseph Thomas > Well, it looks like if I would have to acknowledge that I´m still a > bit off track. So I will try to get deeper into those detection and > recognition tasks to gain a more fundamental understanding. Thanks for > so far, Dr. Judy! Are you trying to understand for fun or are you trying to design something that depends on human visual resolution You can check out http://webvision.med.utah.edu/KallSpatial.html http://www.psych.ndsu.nodak.edu/mccourt/Psy460/Visual%20acuity/Visual%20acuity.html for more detail And if physics and monitor resolutions are interesting or understandable to you: http://www.psych.ndsu.nodak.edu/mccourt/Psy460/Spatial%20frequency%20analysis/Sp atial%20frequency%20analysis.doc http://www.isd.mel.nist.gov/US&R_Robot_Standards/Visual_Acuity_Standards_1.pdf good luck and happy reading Dr Judy > Dr Judy wrote: > Are you trying to understand for fun or are you trying to design > something that depends on human visual resolution My interest is in photography and I´m trying to understand the perception side better in order to make better use of the medium. So it´s either not just for fun or professional but somewhere in between. > You can check out > http://webvision.med.utah.edu/KallSpatial.html [quoted text clipped - 4 lines] > http://www.psych.ndsu.nodak.edu/mccourt/Psy460/Spatial%20frequency%20analysis/Sp atial%20frequency%20analysis.doc > http://www.isd.mel.nist.gov/US&R_Robot_Standards/Visual_Acuity_Standards_1.pdf Thanks a lot for the links! That´s quite interesting material for me. Joseph Thomas |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2009 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.