The Science Behind The Ishihara Test
Vložit
- čas přidán 20. 07. 2024
- Pseudoishochromatic Plates - including the famous Ishihara Plates - are the most common type of test used for detected colorblindness. This video will explain how they work, why they look like that and how to make your own.
Footnotes and Transcript:
chromaphobe.com/the-science-b...
Glossary, including most of the technical terms in this video:
chromaphobe.com/words/
PIP Generator:
franciscouzo.github.io/ishihara/
I really appreciate your channel. Thanks for breaking things down in a way I can explain to other people what being a "severe" deutan really means. Saying "pink and gray are identical" confuses then even more. lol
I saw the 6 before you mentioned that it changed to a 6 and was hoping for a color-perception explanation of why a 5 looked like a 6
Disappointed with your "better" vision? haha. Could also be your screen color balance is a bit out of whack.
@@Chromaphobe I saw the 6 as well. I suppose this is just more proof that taking these tests outside of controlled environments leads to inaccuracy.
Precisely!
Same here. Just right before being revealed, my brain yells "Wow, those camouflage effects are so cool that it can disguise 5 as 6!" 😅😅
At 6:20 I saw a 6 but my brain changed it into a weird 5 because that's what it expected.
And I am interested in the topic why 'the ishihara test is insanly susceptible to cheating'.
heh, nice stuff with the 5 at the end lol
I thought I saw a 5, but did note that the top left diagonal of the '5' looked a bit oddly angled, so good prank. even then, the extra line for the 6 is thinner
Oh my god I got totally memed when you switched the 5 to a 6
6:33 I see the "6" on the simulation (yes, very distorted, but I see difference in color) while I don't see it on the original :P (I have mild deutaranomaly).
I'm a colour normal but I stumbled upon your channel recently and I'm really enjoying your content! Very under appreciated! Just came to say I'm intrested in cheating the Ishihara test, and maybe could you show off alternatives while you're at it?
Hey, glad you found me! I've been getting a lot of appreciation in the last few days and it feels good. By alternatives, Do You mean cheating the other types of tests?
@@Chromaphobe well I don't even know about other types of tests so both showing them off and seeing the tradeoffs with them, which would probably include cheating in them if possible
6:17
I see nothing
But in big one is 6
And on top 5(if i don't see 5 i think im hecking blind)
can you please do a tutorial on how you made the color blind test and what software did you use
the middle seems easier than the bottom
The middle one is just an approximate simulation of the real deal, which is different for everyone.
❤
I wonder if the isoluminant point varies from one person to another with CVD… perhaps based on cone counts / expression somehow. I have a set of cousins, three boys, that each had their own apparent color confusion patterns. Their mom thought it was three different types of colorblindess and I had trouble explaining why that didn’t seem possible genetically. Just an odd case. Also, looking forward to a reverse ishihara vid / script… would love to know how those work
That's exactly right! The cone mosaic, that is... How many red vs. Green (vs. Blue) cones you have differs greatly. It can actually be the source of very mild CVD (and IS the source of most cases of Tritanomaly, in a way) , yet doesn't affect your red-green channel nearly as much as the peak wavelengths. Just pulling from the top of my head (so big grain of salt), normal ratios of red:green can range from 3:1 to 1:10. You can imagine how that can skew your luminant curve severely.
Is there a version of the CIE 1931 color space diagram which has been adjusted to be isoluminant?
The original is meant to be somewhat isoluminant to "the standard observer". CIELUV is a little better here, but I don't think either do it perfectly. CIELAB has some projections that are isoluminant, but CIELAB doesn't have a neat, standard "Chromaticity diagram" like the other two.
I've never seen or made a Chromaticity diagram that is isoluminant to a protanope, for example, though.
One of the problems is that the colors along the Line of Purples are by definition at the limit of human perception. In a perfect Chromaticity diagram, those would all be pure black. If those are gonna be isoluminant to anyone, they would have to be infinitely bright.
Can you give the link to the reverse plate generator?
The one in the description does do reverse PIPs but I don't find them that precise at all.
Why is it that a copunctal point exists, from which all the lines of confusion radiate? Or is that just more of a theoretical approximation? I would have thought the lines of confusion would be a little wavy or irregular.
Great question. Honestly, I've wondered this myself and I'm not sure. I need to read more about it. And obviously, it's going to depend on what color space your Chromaticity diagram is in. When you have CIELUV, the same confusion lines would not be linear like in this (CIEXYZ) Chromaticity diagram, but maybe it is just an approximation so you can sum up confusion lines with some simple coordinates.
@@Chromaphobe yes I was thinking about it and it definitely depends on how your chromaticity diagram is oriented. Obviously if the shape was transformed to a big triangle then the lines of confusion would be curved (but still not intersecting), all else being the same. But I do understand that in a way you are taking a 2D image and flattening it into a line. Interesting! There are probably some cool mathematics in play behind the scenes to determine the shape of the chromaticity diagram.
I could still see the number (and yea, I noticed it was a 6!) on my screen, both on my external monitor and laptop built in. they are both pretty bad monitors, tho.
I tested it on three screens that tend to differ a bit on color and on one both my color normal wife and I could kinda make it out. That's why screen based tests are always going to be worse than printed. A publisher can control the printing process. A programmer can't control your monitor calibration.