I’m reading a book about the history of science in art, which makes me realise that my knowledge of how we perceive colours is pretty bloody basic.

I “know” that we have three types of “cone” receptors at the back of the eye, that respond to different wavelengths. Wikipedia has the diagram below:

http://en.wikipedia.org/wiki/File:Cones_SMJ2_E.svg

My questions (for a start) are:
- How is this response curve measured?

- When did this theory first achieve theory status?

- How did they measure the response curve way back then?

- How much variation is there in the response of cones of nominally the same type in one individual?

- Are there some cones which cannot be allocated a definite type; mid-way between M and L for instance?

- How much variation between individuals? Do some people have a red curve further to the left than other people’s green curve?

- How does the brain know what type of cone the signal is coming from? Do they connect to different areas of the brain, or are the signals different, or what?

- How come green and red look so different when there is such a large overlap in the response curves?

Oh my goodness…..that is a whole area of research! I’ve got to go to dog training. I’ll have a quick flick around for some background stuff now, but I’ll try to come back this afternoon.

Here is one paper:

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2121835/

Has a history section somewhere.

Try this one. I trained with Mike Kalloniatis (the author). Looks like he’s done a pretty thorough overview

http://webvision.med.utah.edu/book/part-viii-gabac-receptors/color-perception/

Note – I skimmed very quickly.

buffy said:

Oh my goodness…..that is a whole area of research! I’ve got to go to dog training. I’ll have a quick flick around for some background stuff now, but I’ll try to come back this afternoon.

Thanks buffy. I just jotted down questions as they came to me.

It’s called the Trichromatic Theory of Colour Vision. If you want to do some more searching. It always helps to be in the know about the terminology for more accurate searching.

:)

Hmm, I probably shouldn’t have started this when I’ve got a pile of work that needs doing by the end of the week :)

(Not to mention the dreaded BAS :()

Thanks again buffy, look like good links.

I’ll see if I can answer my questions :)

One final question:

Considering the large overlap between the red and green cones, how come there is no such colour as “reddy-green”?

The Rev Dodgson said:

One final question:

Considering the large overlap between the red and green cones, how come there is no such colour as “reddy-green”?

And if there was, what colour would it be in the dark?

The Rev Dodgson said:

The Rev Dodgson said:

One final question:

Considering the large overlap between the red and green cones, how come there is no such colour as “reddy-green”?

And if there was, what colour would it be in the dark?

Perhaps because green is not a primary colour?

jjjust moi said:

The Rev Dodgson said:

The Rev Dodgson said:

One final question:

Considering the large overlap between the red and green cones, how come there is no such colour as “reddy-green”?

And if there was, what colour would it be in the dark?

Perhaps because green is not a primary colour?

But it is. Our cone receptors have peak response to Red, Green and Blue, and these are the three colours used to make projected colour images.

The Rev Dodgson said:

Considering the large overlap between the red and green cones, how come there is no such colour as “reddy-green”?

There is, actually. It’s called “yellow”.

Another person to follow up would be Professor Barry L. Cole. Lots of colour vision research, including stuff relating it to real life.

The Rev Dodgson said:

- How come green and red look so different when there is such a large overlap in the response curves?

A complete answer to this question would involve the nature of consciousness. But if we restrict ourselves to the purely physical, then one can view colour as being a direction in a three-dimensional colour space. Mathematically, this colour space isn’t a vector space, but it is useful to visualise it as a three-dimensional vector space. The overlap between the response curves would suggest that the directions represented by the cones themselves are not orthogonal in this colour space. But the differences between the response curves means that the directions are linearly independent, and thus one can diagonalise the matrix of basis vectors so that they are orthogonal. Nevertheless, we are unable to distinguish between spectral yellow (a pure wavelength) and RGB yellow (a mixture of red and green wavelengths).

KJW said:

The Rev Dodgson said:

Considering the large overlap between the red and green cones, how come there is no such colour as “reddy-green”?

There is, actually. It’s called “yellow”.

But it’s called “yellow” because it doesn’t look like red or green, so calling it “reddy-green”? wouldn’t make sense.

This is different to “bluey-green” which looks like it is halfway between blue and green.

KJW said:

The Rev Dodgson said:

- How come green and red look so different when there is such a large overlap in the response curves?

But the differences between the response curves means that the directions are linearly independent, and thus one can diagonalise the matrix of basis vectors so that they are orthogonal.

Could you explain that further please.

The Rev Dodgson said:

KJW said:

The Rev Dodgson said:

Considering the large overlap between the red and green cones, how come there is no such colour as “reddy-green”?

There is, actually. It’s called “yellow”.

But it’s called “yellow” because it doesn’t look like red or green.

Doesn’t it? Smoothly transition from RGB to RGB and one will see that the transition is perceptually smooth.

The Rev Dodgson said:

KJW said:

There is, actually. It’s called “yellow”.

But it’s called “yellow” because it doesn’t look like red or green.

Doesn’t it? Smoothly transition from RGB(255,0,0) to RGB(0,255,0) and one will see that the transition is perceptually smooth.

(I don’t know why the RGB arguments didn’t display)

>This is different to “bluey-green” which looks like it is halfway between blue and green.

There are combinations of red and green which can be described as “reddy green” or “greeny red”, but such terms aren’t often used. The blue-green designation is more familiar possibly because such colour blends are more popular.

Bubblecar said:

>This is different to “bluey-green” which looks like it is halfway between blue and green.

There are combinations of red and green which can be described as “reddy green” or “greeny red”, but such terms aren’t often used. The blue-green designation is more familiar possibly because such colour blends are more popular.

Could you post an example?

Another way of putting this question is why when we go from blue to green there is no point on the spectrum where the colour does not have a certain bluishness, or a certain greenishness, or both.

Yellow on the other hand is neither greenish or reddish, it is a different colour between yellowish-red and yellowish-green.

That’s how I see it anyway.

The Rev Dodgson said:

Another way of putting this question is why when we go from blue to green there is no point on the spectrum where the colour does not have a certain bluishness, or a certain greenishness, or both.

Yellow on the other hand is neither greenish or reddish, it is a different colour between yellowish-red and yellowish-green.

That’s how I see it anyway.

I suppose the pertinent question is what you would expect red-green to look like if not yellow.

KJW said:

The Rev Dodgson said:

Another way of putting this question is why when we go from blue to green there is no point on the spectrum where the colour does not have a certain bluishness, or a certain greenishness, or both.

Yellow on the other hand is neither greenish or reddish, it is a different colour between yellowish-red and yellowish-green.

That’s how I see it anyway.

I suppose the pertinent question is what you would expect red-green to look like if not yellow.

I’d expect it to look like a combination of green and red, in the same way that cyan looks like a combination of blue and green, and magenta looks like a combination of red and blue.

It may be significant that yellow was the only one of the subtractive primaries that was recognised in traditional artists colour theory (and still is for that matter). Magenta was sufficiently red to be called red, and cyan was sufficiently blue to be called blue, but there is no way you could call yellow either red or green.

Between the edges here you get various greeny reds or reddish greens:

I’m particularly interested in whether this is entirely inside our (or at least my) brains, or whether there is some physical basis in the eye for it.

(Actually you keep getting them from the second gradation on the left to the other end, because I ran out of space :))

Bubblecar said:

Between the edges here you get various greeny reds or reddish greens:

OK, so what is going on there?; it doesn’t go through yellow (or orange).

The Rev Dodgson said:

I’m particularly interested in whether this is entirely inside our (or at least my) brains, or whether there is some physical basis in the eye for it.

I’ve only popped in – did any of Buffy’s links answer this. There is a lot (A LOT) of neurological processing involved in perceiving colour. Contrast and context and even physiological state will affect colour perception.

There are also tetrachromat humans- a few women have two opsin types for red colour (sex linked trait). But they need training to actually perceive the difference, and some artist men trained in colour theory will also perceive a difference albeit without the red opsin variant.

worth a podcast and listen

http://www.radiolab.org/2012/may/21/

To what extent is color a physical thing in the physical world, and to what extent is it created in our minds? We start with Sir Isaac Newton, who was so eager to solve this very mystery, he stuck a knife in his eye to pinpoint the answer. Then, we meet a sea creature that sees a rainbow way beyond anything humans can experience, and we track down a woman who we’re pretty sure can see thousands (maybe even millions) more colors than the rest of us. And we end with an age-old question, that, it turns out, never even occurred to most humans until very recently: why is the sky blue?

neomyrtus_ said:

And we end with an age-old question, that, it turns out, never even occurred to most humans until very recently: why is the sky blue?

I thought that was a question that the vast majority of people asked at a certain stage when they were young. And it is considered in that statement to be a recent question? I don’t see why it would be.

>OK, so what is going on there?; it doesn’t go through yellow (or orange).

It’s just a red selected from the Photoshop colour picker on one 600 × 800 frame, and a green selected on another, with the green added as a 10% opaque new layer 11 times onto the original.

http://discovermagazine.com/2012/jul-aug/06-humans-with-super-human-vision#.UXOn6kogb5w

http://nexthumanproject.com/references/Tetrachromats.pdf

hypnosis and loss / addition of colour perception

http://www.news.harvard.edu/gazette/2000/08.21/hypnosis.html

http://ajp.psychiatryonline.org/article.aspx?articleID=174269

http://www.ncbi.nlm.nih.gov/pubmed/10910791

http://mindhacks.com/2008/10/26/synaesthesia-induced-by-hypnosis/

Skunkworks said:

neomyrtus_ said:

And we end with an age-old question, that, it turns out, never even occurred to most humans until very recently: why is the sky blue?

I thought that was a question that the vast majority of people asked at a certain stage when they were young. And it is considered in that statement to be a recent question? I don’t see why it would be.

listen to the podcast – there are cultural groups of people who don’t register blue. Also historical (ancient Greek) references to a red sky.

http://webvision.med.utah.edu/book/part-vii-color-vision/color-vision/

neomyrtus_ said:

listen to the podcast – there are cultural groups of people who don’t register blue. Also historical (ancient Greek) references to a red sky.

But kiddies would still have asked why is they sky x colour?

Skunkworks said:

neomyrtus_ said:

listen to the podcast – there are cultural groups of people who don’t register blue. Also historical (ancient Greek) references to a red sky.

But kiddies would still have asked why is they sky x colour?

a naive child in the program called the sky white (IIRC).

neomyrtus_ said:

Skunkworks said:

neomyrtus_ said:

listen to the podcast – there are cultural groups of people who don’t register blue. Also historical (ancient Greek) references to a red sky.

But kiddies would still have asked why is they sky x colour?

a naive child in the program called the sky white (IIRC).

Maybe, I am disagreeing that it is a rare and modern thing to consider.

the greeks referred to a bronze sky

dawn sat on her golden throne

http://www.bbc.co.uk/blogs/tv/2011/08/horizon.shtml

OK, so what is going on there?; it doesn’t go through yellow (or orange).
——————————————-

You canna display RGB on a computer screen…

Skunkworks said:

Maybe, I am disagreeing that it is a rare and modern thing to consider.

do what you like – I just drew attention to a program which discusses that what may be obvious to us ain’t always obvious to others.

Mr Ironic said:

OK, so what is going on there?; it doesn’t go through yellow (or orange).
——————————————-

You canna display RGB on a computer screen…

Not sure what you mean.

How detailed do you want to get. This is getting a bit complicated for me (and I’ve got a fair grasp) but you might want to go into the finer stuff. And maybe follow up some of the references.

http://onlinelibrary.wiley.com/doi/10.1113/jphysiol.2010.192658/full

There are quite a lot of rather old papers around, but this one seems to be a fair representation of the state of play. It comes down to “Bless you! We don’t really know!”

Mr buffy here…..have you done that BAS yet?

;)

buffy said:

Mr buffy here…..have you done that BAS yet?

;)

Standard practice here is to submit at about 5 mins to midnight on the last day :)

buffy said:

How detailed do you want to get. This is getting a bit complicated for me (and I’ve got a fair grasp) but you might want to go into the finer stuff. And maybe follow up some of the references.

http://onlinelibrary.wiley.com/doi/10.1113/jphysiol.2010.192658/full

There are quite a lot of rather old papers around, but this one seems to be a fair representation of the state of play. It comes down to “Bless you! We don’t really know!”

Not too detailed!

I will follow up the references, but it looks like there’s a pretty big gap between the academic stuff and the dumbed down stuff for the rest of us.

Not sure what you mean.
———————————————

Well all human manufactured colours are wrong…

CMYK is a close apoximation of the theory, on paper, however we cannot produce the colours exacty. Hence the K.

With computer screens we end up with vinyetts, graduated screens, that give a poor example of the colours aimed for.

Never mind everyones screen is set to different setting settings on differing grades.

Mr Ironic said:

Not sure what you mean.
———————————————

Well all human manufactured colours are wrong…

CMYK is a close apoximation of the theory, on paper, however we cannot produce the colours exacty. Hence the K.

With computer screens we end up with vinyetts, graduated screens, that give a poor example of the colours aimed for.

Never mind everyones screen is set to different setting settings on differing grades.

OK, but I don’t think that answers the question.

I mean computer screens are quite capable of adding red and green and showing yellow.

CMYK is a close apoximation of the theory, on paper, however we cannot produce the colours exacty. Hence the K.

yep, the shadows were never “black” enough. more a really dark brown.

Might be a signal processing thing like musical chords.

http://upload.wikimedia.org/wikipedia/commons/d/d1/Major_triad.svg

OK, so what is going on there?; it doesn’t go through yellow (or orange).
————————————————————-

OK, I think the example colour pallet chosen is incorrect.

Anyway to add to the debate…

There are no ‘Primary Colurs’…

Yellow, purple and (can’t remember) maybe puce… would do similiar work.

And then women and blokes receptors differ.

And if you live with an artist…

Then give up.

Bit of perception of colour on now at SBS.

>OK, I think the example colour pallet chosen is incorrect.

?

He wanted blends of red & green, he got them. 100% correct, right, AOK & check.

He wanted blends of red & green, he got them. 100% correct, right, AOK & check.
————————————————————
100 in the cent… No way any day any way.

But (I believe Rev wanted) more like…. red toooo Green.

>But (I believe Rev wanted) more like…. red toooo Green.

The Rev hasn’t complained about my reddish greens etc, and he’d be a fool to try.

The Rev hasn’t complained about my reddish greens etc, and he’d be a fool to try.
————————————————

Well he did question where the yellow was…

Formed from red and green squares at three different scales. The lower-right is an actual yellow RGB(127,127,0)

The lower-right is an actual yellow RGB
———————————————————

Looks like baby shit brown diluted with asparagus urine to me…

Mr Ironic said:

Anyway to add to the debate…

There are no ‘Primary Colurs’…

Yellow, purple and (can’t remember) maybe puce… would do similiar work.

Yes, you can choose other sets of colours to act as the basis vectors of a colour space, but some choices are better than others. A poor choice only covers a small subset of all the colours that are visible to normal human vision.

See http://en.wikipedia.org/wiki/Gamut

I think something went wrong there, KJW

Mr Ironic said:

The lower-right is an actual yellow RGB (127,127,0)
———————————————————

Looks like baby shit brown diluted with asparagus urine to me…

That’s because it’s got a value of 50%. If we boost it to RGB (125, 125, 0), then we get a 100% yellow.

Mr Ironic said:

The lower-right is an actual yellow RGB
———————————————————

Looks like baby shit brown diluted with asparagus urine to me…

Assuming linear response, RGB(127,127,0) is the correct average of RGB(255,0,0) and RGB(0,255,0). However, non-linearity is evident. Also, on LED displays, the viewing angle is also significant.

Oops! Sorry about the typo. Let’s try that again…

That’s because it’s got a value of 50%. If we boost it to RGB (255, 255, 0), then we get a 100% yellow.

but some choices are better than others. A poor choice only covers a small subset of all the colours that are visible to normal human vision.
—————————————————

Yes, the gamut range is the important factor.

Ease of repoducing it still leaves us with, what, 85% of the visual spectrum?

Metallics and sheen and depth being a by product loss.

I was watching a car/manufacturing show where they showed a Rolls tester faulting a car for a paint defect that was invisible? I expect there was some hyper-bowl but it is still astonishing, checking for errors than an owner could not detect.

That’s because it’s got a value of 50%.
—————————————————-
If it is a solid colour then where are the components?

Mr Ironic said:

That’s because it’s got a value of 50%.
—————————————————-
If it is a solid colour then where are the components?

As KJW said “The lower-right is an actual yellow RGB (127,127,0)”
Those colour components range from 0 to 255, so it’s a yellow with a value of (slightly less than) 50%, as I said previously.

KJW said:

Mr Ironic said:

The lower-right is an actual yellow RGB(127,127,0)
———————————————————

Looks like baby shit brown diluted with asparagus urine to me…

dv said:

I think something went wrong there, KJW

Assuming linear response, RGB(127,127,0) is the correct average of RGB(255,0,0) and RGB(0,255,0). However, non-linearity is evident. Also, on LED displays, the viewing angle is also significant.

Actually, if one considers an RMS-average, the average of RGB(255,0,0) and RGB(0,255,0) is RGB(181,181,0). I initially tried to match the colour by eye, but decided that a mathematical average would be more appropriate, even if it didn’t match the other examples.

“The lower-right is an actual yellow RGB (127,127,0)”
———————————————————
No it’s not.

It is an approximation.

Set by a dude, that gets paid to do that stuff.

KJW said:

Actually, if one considers an RMS-average, the average of RGB(255,0,0) and RGB(0,255,0) is RGB(181,181,0). I initially tried to match the colour by eye, but decided that a mathematical average would be more appropriate, even if it didn’t match the other examples.

Yeah, an RMS average isn’t particularly applicable here. A better approach is to convert the colours to a colour space that’s closer to the human colour perceptual space, eg http://en.wikipedia.org/wiki/CIE_1931_color_space

decided that a mathematical average would be more appropriate,
——————————————————————————

Hearin lies the problem…

We cannot produce colour with any real sense of accuracy.

Artist’s come the closest.

Mr Ironic said:

“The lower-right is an actual yellow RGB (127,127,0)”
———————————————————
No it’s not.

It is an approximation.

Set by a dude, that gets paid to do that stuff.

There is no definitive RGB colour space. So the actual appearance of RGB (127,127,0) depends on the construction of your monitor, and possibly also on settings in your OS / graphics card that you may be able to alter.

So the actual appearance of RGB (127,127,0) depends on the construction of your monitor, and possibly also on settings in your OS / graphics card that you may be able to alter.
———————————————————
Yes, so it is open for opinion.

All those dilutions of colours are manufactured into the programs. What represents what is beholden to the limitations of the computer and the voyeur…

But yes, yellow is yellow as long as it hits a daffodil that is purple…

Mr Ironic said:

“The lower-right is an actual yellow RGB (127,127,0)”
———————————————————
No it’s not.

It is an approximation.

Set by a dude, that gets paid to do that stuff.

Huh?

RGB (127,127,0) was the setting used to create the lower-right corner. The other three corners are actually a mixture of tiny red RGB (255,0,0) and green RGB (0,255,0) squares, not yellow. The point was to show The Rev Dodgson what red-green actually looks like.

It is an approximation.

Set by a dude, that gets paid to do that stuff.

Huh?
—————
OK, the math is fine and the theory is OK.

But in practise it is rubbish. We just don’t have the gear.

So we guess.

I’m an artist. I just chose a red and a green and blended them, and got reddish green and greenish red. I could do the same with blue & green and get blends that the Rev would readily recognise as blue-green etc.

It’s just that reddish green tends to be perceived as a less distinctive, “murkier” blend. Quite why that should be so is an interesting question.

…and unlike KJW I was specifically looking for colours that could be described, in the way Rev meant, as “reddish green” etc rather than “yellow”.

Quite why that should be so is an interesting question.
———————————————————————————
Because adding colours, from the pallet, tends them toward black.

Bubblecar said:

I’m an artist. I just chose a red and a green and blended them, and got reddish green and greenish red. I could do the same with blue & green and get blends that the Rev would readily recognise as blue-green etc.

It’s just that reddish green tends to be perceived as a less distinctive, “murkier” blend. Quite why that should be so is an interesting question.

When you say you’re an artist, you mean with paints, right? If so, then that’s subtractive mixing, whereas the mixing of red and green on a monitor is additive. Printers mix ink subtractively using cyan, yellow, and magenta, the complements of the additive primary colours red, green, and blue respectively

Mr Ironic said:

Quite why that should be so is an interesting question.
———————————————————————————
Because adding colours, from the pallet, tends them toward black.

Yes, but blue-greens do tend to be perceived as more vivid than red-greens (unless we’re talking yellow).

KJW said:

Printers mix ink subtractively using cyan, yellow, and magenta, the complements of the additive primary colours red, green, and blue respectively

Sorry, I meant:

Printers mix ink subtractively using cyan, magenta, and yellow, the complements of the additive primary colours red, green, and blue respectively

>If so, then that’s subtractive mixing, whereas the mixing of red and green on a monitor is additive.

Yes, but in subtractive mixing, blue-greens are still generally poerceived as more distinctive than red-greens, which I suspect is what the Rev may have been getting at.

Printers mix ink subtractively using cyan, yellow, and magenta, the complements of the additive primary colours red, green, and blue respectively
————————————————————————————-

Actinic and sub-actic light, being the difference.

Yes, but blue-greens do tend to be perceived as more vivid than red-greens (unless we’re talking yellow).
————————————————
The further apart on the colour wheel…the closer to black.

Try adding yellow to purple, is it lighter or darker?

On the topic of subtractive mixing, I was taught in primary school that mixing blue with yellow forms green (and I’ve seen that to be the case). Yet, in the theory of subtractive mixing, mixing blue with yellow forms black… ???

Yet, in the theory of subtractive mixing, mixing blue with yellow forms black… ???
————————————————————————

What was the light force?

Bubblecar said:

…and unlike KJW I was specifically looking for colours that could be described, in the way Rev meant, as “reddish green” etc rather than “yellow”.

Actually, my image was inspired by your attempt to produce a red-green.

Fair enough. I daresay if I’d chosen a slightly different red and green I might have achieved a more yellow result somewhere in the picture, but the blends I got seemed to satisfy a “reddish green” description :)

Although in real life it’s not a colour description I’d often use, except for example when using a semi-transparent green over a red background to produce a green with a reddish inner glow etc.

a semi-transparent green over a red background to produce a green with a reddish inner glow etc.
—————————————-

Yeap… Who said what about artists’….

Bubblecar said:

>If so, then that’s subtractive mixing, whereas the mixing of red and green on a monitor is additive.

Yes, but in subtractive mixing, blue-greens are still generally perceived as more distinctive than red-greens, which I suspect is what the Rev may have been getting at.

Probably.

FWIW, one of my colour-blind friends will often refer to certain shades of brown as reddish-green.

Our visual system can’t directly detect yellow. Light in the yellow region of the spectrum sets off both our red detectors and our green detectors (but not our blue detectors), so we can get the same perceived colour as a spectrally pure yellow light by making an appropriate mixture of red light and green light. It doesn’t matter whether you do this directly with light sources, or indirectly with pigments, as long as you get the proportions correct.

For some reason, our visual system isn’t that bothered with making major distinctions in the bluey-green region of the spectrum. But it’s quite keen on discriminating between red, yellow and green. Maybe it’s something our ancestors evolved due to the advantage it gives in spotting ripe fruit.

So our brains generally tend to register colours between red & green as combinations with yellow (or brown, which is just a dark reddish yellow), rather than as reddish-green blends. But this can be affected to an extent by context – e.g., the red-green transition picture posted earlier doesn’t contain any bright yellows, and the transition “tells” the brain that the intermediate colours are somehow related to red and green.

I think the closest to what The Rev Dodgson would describe as red-green comes from having a red optical filter covering one eye and a green optical filter covering the other eye.

PM 2Ring said:

Bubblecar said:

>If so, then that’s subtractive mixing, whereas the mixing of red and green on a monitor is additive.

Yes, but in subtractive mixing, blue-greens are still generally perceived as more distinctive than red-greens, which I suspect is what the Rev may have been getting at.

Probably.

FWIW, one of my colour-blind friends will often refer to certain shades of brown as reddish-green.

Our visual system can’t directly detect yellow. Light in the yellow region of the spectrum sets off both our red detectors and our green detectors (but not our blue detectors), so we can get the same perceived colour as a spectrally pure yellow light by making an appropriate mixture of red light and green light. It doesn’t matter whether you do this directly with light sources, or indirectly with pigments, as long as you get the proportions correct.

For some reason, our visual system isn’t that bothered with making major distinctions in the bluey-green region of the spectrum. But it’s quite keen on discriminating between red, yellow and green. Maybe it’s something our ancestors evolved due to the advantage it gives in spotting ripe fruit.

So our brains generally tend to register colours between red & green as combinations with yellow (or brown, which is just a dark reddish yellow), rather than as reddish-green blends. But this can be affected to an extent by context – e.g., the red-green transition picture posted earlier doesn’t contain any bright yellows, and the transition “tells” the brain that the intermediate colours are somehow related to red and green.

I have said at times that of the three primary colours red, green, and blue, red appears to be the odd one out. I suspect that the qualia of red is more than just visual, and that there are non-visual aspects to it as well. For example, red seems to be associated with danger and I believe this is a part of the distinction between the qualia of red and that of green and blue.

PM 2Ring said:

KJW said:

Actually, if one considers an RMS-average, the average of RGB(255,0,0) and RGB(0,255,0) is RGB(181,181,0). I initially tried to match the colour by eye, but decided that a mathematical average would be more appropriate, even if it didn’t match the other examples.

Yeah, an RMS average isn’t particularly applicable here. A better approach is to convert the colours to a colour space that’s closer to the human colour perceptual space, eg http://en.wikipedia.org/wiki/CIE_1931_color_space

Judging by eye, RGB(181,181,0) would have been closer in brightness to the other quadrants, but still less bright.

KJW said:

I have said at times that of the three primary colours red, green, and blue, red appears to be the odd one out. I suspect that the qualia of red is more than just visual, and that there are non-visual aspects to it as well. For example, red seems to be associated with danger and I believe this is a part of the distinction between the qualia of red and that of green and blue.

FWIW, the genes for the pigments in the red & green cones are on the X chromosome, but the blue pigment gene’s on chromosome 7. So you’d think that’d make blue the odd one out.

For example, red seems to be associated with danger and I believe this is a part of the distinction between the qualia of red and that of green and blue.
——————————————————————————————————————-

Why do you think that?

What dangerous animal is red?

More to the point how many animals are red?

It is distictive by its absence…

I think buffy has mentioned in the Past that women who are heterozygous can see more colours than normal.

PM 2Ring said:

KJW said:

I have said at times that of the three primary colours red, green, and blue, red appears to be the odd one out. I suspect that the qualia of red is more than just visual, and that there are non-visual aspects to it as well. For example, red seems to be associated with danger and I believe this is a part of the distinction between the qualia of red and that of green and blue.

FWIW, the genes for the pigments in the red & green cones are on the X chromosome, but the blue pigment gene’s on chromosome 7. So you’d think that’d make blue the odd one out.

Not from an informational perspective, because the brain doesn’t “know” anything about the origin of the signals, only what it does with them, which would be different in the case of red if red is associated with danger.

KJW said:

PM 2Ring said:

Yeah, an RMS average isn’t particularly applicable here. A better approach is to convert the colours to a colour space that’s closer to the human colour perceptual space, eg http://en.wikipedia.org/wiki/CIE_1931_color_space

Judging by eye, RGB(181,181,0) would have been closer in brightness to the other quadrants, but still less bright.

To compare the brightness of RGB colours (without full conversion to a CIE space), you can use this formula below for luminance:

From http://en.wikipedia.org/wiki/Luminance_%28relative%29

For RGB color spaces that use the ITU-R BT.709 primaries (or sRGB, which defines the same primaries), relative luminance can be calculated from linear RGB components:

Y = 0.2126 R + 0.7152 G + 0.0722 B

The formula reflects the luminosity function: green light contributes the most to the intensity perceived by humans, and blue light the least.

I use that formula to scale RGB colour space in a simple colour matching program I wrote several years ago (in C), and it works rather well.

Mr Ironic said:

For example, red seems to be associated with danger and I believe this is a part of the distinction between the qualia of red and that of green and blue.
——————————————————————————————————————-

Why do you think that?

What dangerous animal is red?

What about blood?

What about blood?
————————————

Maybe…

But it is a bit late in the altercation to start fretting…

Mr Ironic said:

It (red) is distictive by its absence…

You think so? Amongst organic chemicals, green and blue compounds are rare, while colours at the red end of the spectrum (yellow, brown, etc) are much more common. In fact, it is very rare to do a chemical reaction in organic chemistry where the crude product isn’t at least slightly yellow.

Yellow is common because a large proportion of organic compounds absorb in the UV range, and if the absorption spectrum has a tail into the blue, then it will appear yellow. If the electronic transition is of lower energy than typical, the compound will absorb significantly in the blue and green regions, producing orange or even red compounds. But to produce blue or green compounds requires greater absorption of red than blue or green, and this will involve rather special chromophores than in the case of yellow, orange. or red compounds.

Red + Green = Brown.

If you have just 3 colours, Yellow, Blue, Red, then you can mix them all together to get brown.

Yellow + Purple = Brown

Blue + Orange = Brown

If you have equal amounts of Red, Blue and Yellow in the above, then you will end up with the same colour Brown.

tauto said:

Red + Green = Brown.

If you have just 3 colours, Yellow, Blue, Red, then you can mix them all together to get brown.

Yellow + Purple = Brown

Blue + Orange = Brown

If you have equal amounts of Red, Blue and Yellow in the above, then you will end up with the same colour Brown.

KJW said:

while colours at the red end of the spectrum (yellow, brown, etc) are much more common.

I should remark that pure compounds are rarely brown. Brown colour is usually due to a mixture of different coloured compounds.

>Red + Green = Brown.

Additively speaking, red + green = yellow (or reddish green, depending on ratios etc :)). But when mixing pigments, various browns can be achieved in various combinations, as you say.

PM 2Ring said:

KJW said:

PM 2Ring said:

Yeah, an RMS average isn’t particularly applicable here. A better approach is to convert the colours to a colour space that’s closer to the human colour perceptual space, eg http://en.wikipedia.org/wiki/CIE_1931_color_space

Judging by eye, RGB(181,181,0) would have been closer in brightness to the other quadrants, but still less bright.

To compare the brightness of RGB colours (without full conversion to a CIE space), you can use this formula below for luminance:

From http://en.wikipedia.org/wiki/Luminance_%28relative%29

For RGB color spaces that use the ITU-R BT.709 primaries (or sRGB, which defines the same primaries), relative luminance can be calculated from linear RGB components:

Y = 0.2126 R + 0.7152 G + 0.0722 B

The formula reflects the luminosity function: green light contributes the most to the intensity perceived by humans, and blue light the least.

I use that formula to scale RGB colour space in a simple colour matching program I wrote several years ago (in C), and it works rather well.

It occurred to me that the formula you gave is not relevant or only partly relevant to the problem at hand. This is because the formula you gave is linear, whereas the brightness effect that occurred in my image is non-linear. That is, RGB(127,127,0) is significantly less than half as bright as RGB(255,255,0). By construction, the tiny squares of RGB(255,0,0) and RGB(0,255,0) must be half as bright as RGB(255,255,0). Yet, RGB(127,127,0) is much less bright than the tiny squares and therefore much less than half the brightness of RGB(255,255,0). This is not about the relative brightness of the different colours, but the non-linearity of the transfer function between the RGB values of a single colour and the brightness of that colour.

KJW said:

It occurred to me that the formula you gave is not relevant or only partly relevant to the problem at hand. This is because the formula you gave is linear, whereas the brightness effect that occurred in my image is non-linear. That is, RGB(127,127,0) is significantly less than half as bright as RGB(255,255,0). By construction, the tiny squares of RGB(255,0,0) and RGB(0,255,0) must be half as bright as RGB(255,255,0). Yet, RGB(127,127,0) is much less bright than the tiny squares and therefore much less than half the brightness of RGB(255,255,0). This is not about the relative brightness of the different colours, but the non-linearity of the transfer function between the RGB values of a single colour and the brightness of that colour.

However, my suggestion to apply an RMS-average was based on incorrect reasoning (which actually makes your formula not relevant at all because that would only apply to the metric used to perform a weighted RMS-average).

The above is a comparison between a 50% mixture of squares of the corresponding primary colours RGB(255,0,0), RGB(0,255,0), or RGB(0,0,255) with black squares RGB(0,0,0) on the left with the corresponding solid primary colours RGB(191,0,0), RGB(0,191,0), or RGB(0,0,191) on the right. On my monitor, the left and right are approximately equal in brightness as judged by my eyes. However, how they compare to you will depend on your monitor such as the contrast, etc. The left is always 50% of the brightness of the maximum brightness for the respective colour, but the right will depend on the comparison between the RGB values of 191 and 255 with regards to settings that produce non-linear response, such as contrast.

For those interested, this (and the other) image was produced using a spreadsheet with small cell size and small zoom setting.

BTW, the left side of the image should appear quite uniform (apart from the fine black dots). Any bands that appear are a sign of aliasing which can probably be corrected by adjusting the monitor.

I’d just like to say that due to reading this thread last night, I dreamed I was living in a house filled with cyan furniture and cyan walls and I wasn’t allowed to paint over it.

Thanks a lot!

Divine Angel said:

I’d just like to say that due to reading this thread last night, I dreamed I was living in a house filled with cyan furniture and cyan walls and I wasn’t allowed to paint over it.

Thanks a lot!

Cyan is a nice colour :-)

KJW said:

For those interested, this (and the other) image was produced using a spreadsheet with small cell size and small zoom setting.

Right, I’m going off to have a play. :)

FWIW, I do see Bubblecar’s image as containing reddy-greens and greeny-reds, which is what I asked for. The bottom-right in KJW’s image I see as a yellowy green, but PM’s bright yellow I see as just yellow.

It’s interesting that the same colours in the same ratio have these different effects.

The history of the theory of all this (which I’m half way through) is interesting too.

http://www.amazon.com/Science-Art-Optical-Western-Brunelleschi/dp/0300052413

It’s an interesting book, with great illustrations, but it’s not an easy read, and is written first from the perspective of a historian, then the artists, with the science not getting much of a look in.

The Rev Dodgson said:

KJW said:

For those interested, this (and the other) image was produced using a spreadsheet with small cell size and small zoom setting.

Right, I’m going off to have a play. :)

FWIW, I do see Bubblecar’s image as containing reddy-greens and greeny-reds, which is what I asked for. The bottom-right in KJW’s image I see as a yellowy green, but PM’s bright yellow I see as just yellow.

It’s interesting that the same colours in the same ratio have these different effects.

The history of the theory of all this (which I’m half way through) is interesting too.

http://www.amazon.com/Science-Art-Optical-Western-Brunelleschi/dp/0300052413

It’s an interesting book, with great illustrations, but it’s not an easy read, and is written first from the perspective of a historian, then the artists, with the science not getting much of a look in.

Also it’s about 2/3 on perspective in art and 1/3 on colour.

3 1/2 stars, recommended with reservations.

“I have said at times that of the three primary colours red, green, and blue”

No one is ever going to believe me, but I still think it is important to point out that there is no such thing as ‘primary colours’ fullstop, there are primary colours for a specified system. You can use any three colours you like as primaries but you will get a different gamut.

R,G,B are used in monitors and tvs because of the agreeably large gamut achievable with readily producible phosphors (actually the red wasn’t so easy) but otherwise there is nothing magic about these colours.

>>I think the closest to what The Rev Dodgson would describe as red-green comes from having a red optical filter covering one eye and a green optical filter covering the other eye.,,

No, you will probably just get optical rivalry that way as one eye is sending a green picture and one is sending a red one (in the simplest terms) and the brain toggles between them, doesn’t meld them. Although exactly which wavelengths are chosen for the transmission might have a small effect because of the overlap of the two wavelength sensitivity curves.

MartinB said:

“I have said at times that of the three primary colours red, green, and blue”

No one is ever going to believe me, but I still think it is important to point out that there is no such thing as ‘primary colours’ fullstop, there are primary colours for a specified system. You can use any three colours you like as primaries but you will get a different gamut.

R,G,B are used in monitors and tvs because of the agreeably large gamut achievable with readily producible phosphors (actually the red wasn’t so easy) but otherwise there is nothing magic about these colours.

Are you saying it’s entirely coincidental that the three primary colours used in monitors and tvs (and film and every other additive colour image system) happen to be the same three colours that our cone receptors are divided into?

MartinB said:

“I have said at times that of the three primary colours red, green, and blue”

No one is ever going to believe me, but I still think it is important to point out that there is no such thing as ‘primary colours’ fullstop, there are primary colours for a specified system. You can use any three colours you like as primaries but you will get a different gamut.

R,G,B are used in monitors and tvs because of the agreeably large gamut achievable with readily producible phosphors (actually the red wasn’t so easy) but otherwise there is nothing magic about these colours.

I was under the impression that primary colours can’t be made by mixing two other colours. So while red and blue are primary colours green is not.

buffy said:

>>I think the closest to what The Rev Dodgson would describe as red-green comes from having a red optical filter covering one eye and a green optical filter covering the other eye.,,

No, you will probably just get optical rivalry that way as one eye is sending a green picture and one is sending a red one (in the simplest terms) and the brain toggles between them, doesn’t meld them. Although exactly which wavelengths are chosen for the transmission might have a small effect because of the overlap of the two wavelength sensitivity curves.

But is that how it works?

Aren’t there 3D colour systems using glasses with a different colour for each eye, which produce reasonably good colour images? (I may be misremembering here though)

The Rev Dodgson said:

MartinB said:

“I have said at times that of the three primary colours red, green, and blue”

No one is ever going to believe me, but I still think it is important to point out that there is no such thing as ‘primary colours’ fullstop, there are primary colours for a specified system. You can use any three colours you like as primaries but you will get a different gamut.

R,G,B are used in monitors and tvs because of the agreeably large gamut achievable with readily producible phosphors (actually the red wasn’t so easy) but otherwise there is nothing magic about these colours.

Are you saying it’s entirely coincidental that the three primary colours used in monitors and tvs (and film and every other additive colour image system) happen to be the same three colours that our cone receptors are divided into?

But maybe it’s just a case of needing a colour from each end and one from the middle to get a wide range of colours, so you end up with about the same colours in both designed and evolved systems.

Yes, but the 3D systems have two different pictures, one for the red filtered eye and one for the green filtered eye. The only way to get 3D is to have a slightly different picture presented to each eye. This is why (?) our eyes are horizontally displaced in our heads. Slightly different picture from each eye, brain works out distances. I have to teach people who lose the sight of one eye how to become less clumsy about pouring drinks and stuff. You just need two bits of info, so they need to learn to put their hand on the glass/cup they are pouring into so they get the two bits……one from the eye that works, and the other from the proprioceptive messages through the brainstem. Try it yourself – over the sink. Try pouring from a jug into a glass with one eye shut.

>>so you end up with about the same colours in both designed and evolved systems.<<

Maybe it’s just the way it is in this galaxy.

“Are you saying it’s entirely coincidental that the three primary colours used in monitors and tvs (and film and every other additive colour image system) happen to be the same three colours that our cone receptors are divided into?”

It’s not entirely a coincidence, no: to get a decent gamut you are going to have to get somewhere close to the end ranges of our colour vision so something that is close to blue and something that is close to red. But it’s also not entirely the case that the division is the same. There are significant differences between the colours of the phosphors and the colours our eyes respond to. For example the red phosphor in monitors peaks at more like 580nm.

MartinB said:

It’s not entirely a coincidence, no: to get a decent gamut you are going to have to get somewhere close to the end ranges of our colour vision so something that is close to blue and something that is close to red.

I just said that :)

“I was under the impression that primary colours can’t be made by mixing two other colours. So while red and blue are primary colours green is not.”

And it is precisely this misunderstanding that makes me repeat this point of pedantry. No colour is ‘special’ in this way*, every colour can be made by mixing together a combination of other colours.

*Spectrally pure colours will in general require an unphysical mixture of other colours to be created, but again this is a property of all of these colours, not just some special magical ones.

I think there has been some confusion in this thread between pigments and light. Although it has been explained, I think it is still a bit of a problem.

>an unphysical mixture of other colours to be created

Can you expand on that please, Martin. I don’t get it.

buffy said:

I think there has been some confusion in this thread between pigments and light. Although it has been explained, I think it is still a bit of a problem.

It doesn’t help that artists are still taught that the three subtractive primaries are red blue and yellow.

The Rev Dodgson said:

buffy said:

I think there has been some confusion in this thread between pigments and light. Although it has been explained, I think it is still a bit of a problem.

It doesn’t help that artists are still taught that the three subtractive primaries are red blue and yellow.

Good artists ignore most of what they’re taught, and forget the rest.

My personal “wow” for colour vision is the peak sensitivity at 555nm……a lovely grassy green colour. I like to speculate why this is so. Also interesting that the peak shifts in scotopic vision to 507, cutting out the red receptors.

http://hyperphysics.phy-astr.gsu.edu/hbase/vision/bright.html

sibeen said:

>an unphysical mixture of other colours to be created

Can you expand on that please, Martin. I don’t get it.

He’s talking mixing light (additive mixing) vs mixing pigments (subtractive mixing).

Wiki on the biological basis of “primary” colours:

Biological basis

Primary colors are not a fundamental property of light but are related to the physiological response of the eye to light. Fundamentally, light is a continuous spectrum of the wavelengths that can be detected by the human eye, an infinite-dimensional stimulus space. However, the human eye normally contains only three types of color receptors, called cone cells. Each color receptor responds to different ranges of the color spectrum. Humans and other species with three such types of color receptors are known as trichromats. These species respond to the light stimulus via a three-dimensional sensation, which generally can be modeled as a mixture of three primary colors.

Before the nature of colorimetry and visual physiology were well understood, scientists such as Thomas Young, James Clark Maxwell, and Hermann von Helmholtz expressed various opinions about what should be the three primary colors to describe the three primary color sensations of the eye. Young originally proposed red, green, and violet, and Maxwell changed violet to blue; Helmholtz proposed “a slightly purplish red, a vegetation-green, slightly yellowish (wave-length about 5600 tenth-metres), and an ultramarine-blue (about 4820)”. In modern understanding, the human cone cells do not correspond to any real primary colors.

Species with different numbers of receptor cell types would have color vision requiring a different number of primaries. For example, for species known as tetrachromats, with four different color receptors, one would use four primary colors. Since humans can only see to 380 nanometers (violet), but tetrachromats can see into the ultraviolet to about 300 nanometers, this fourth primary color for tetrachromats is located in the shorter-wavelength range.

Many birds and marsupials are tetrachromats, and it has been suggested that some human females are tetrachromats as well, having an extra variant version of the long-wave (L) cone type. The peak response of human color receptors varies, even among individuals with “normal” color vision; in non-human species this polymorphic variation is even greater, and it may well be adaptive. Most placental mammals other than primates have only two types of color receptors and are therefore dichromats; to them, there are only two primary colors.

It would be incorrect to assume that the world “looks tinted” to an animal (or human) with anything other than the human standard of three color receptors. To an animal (or human) born that way, the world would look normal to it, but the animal’s ability to detect and discriminate colors would be different from that of a human with normal color vision. If a human and an animal both look at a natural color, they see it as natural; however, if both look at a color reproduced via primary colors, such as on a color television screen, the human may see it as matching the natural color, while the animal does not, since the primary colors have been chosen to suit human capabilities.

http://en.wikipedia.org/wiki/Primary_color

buffy said:

My personal “wow” for colour vision is the peak sensitivity at 555nm……a lovely grassy green colour. I like to speculate why this is so. Also interesting that the peak shifts in scotopic vision to 507, cutting out the red receptors.

http://hyperphysics.phy-astr.gsu.edu/hbase/vision/bright.html

I looked that up, and I’m a little disappointed to find that Scottish people do not in fact have “special” vision.

Sometimes you need to add a negative amount of a colour, which is unphysical :-)

MartinB said:

Sometimes you need to add a negative amount of a colour, which is unphysical :-)

Sounds like cheating :)

swim with me into your blackest eyes.

♫Blackest Eyes | Porcupine Tree | In Absentia EU disc 1♪♩

What is the wavelength of magenta?

http://en.wikipedia.org/wiki/File:Linear_visible_spectrum.svg

What is the wavelength of magenta?

which or whose one?

Glance Fleeting said:

What is the wavelength of magenta?

http://en.wikipedia.org/wiki/File:Linear_visible_spectrum.svg

About 450 and 650 I suppose.

MartinB said:

“I have said at times that of the three primary colours red, green, and blue”

No one is ever going to believe me, but I still think it is important to point out that there is no such thing as ‘primary colours’ fullstop, there are primary colours for a specified system. You can use any three colours you like as primaries but you will get a different gamut.

R,G,B are used in monitors and tvs because of the agreeably large gamut achievable with readily producible phosphors (actually the red wasn’t so easy) but otherwise there is nothing magic about these colours.

I largely disagree with this. While it is true that the colours red, green, and blue are not unique colours and there is some flexibility in choosing the basis colours of the colour space, in the end, the basis colours are still going to be regarded as red, green, and blue. Thus, while there may be some variation in the colours defined as red, green, and blue, red, green, and blue are still the primary colours.

KJW said:

MartinB said:

“I have said at times that of the three primary colours red, green, and blue”

No one is ever going to believe me, but I still think it is important to point out that there is no such thing as ‘primary colours’ fullstop, there are primary colours for a specified system. You can use any three colours you like as primaries but you will get a different gamut.

R,G,B are used in monitors and tvs because of the agreeably large gamut achievable with readily producible phosphors (actually the red wasn’t so easy) but otherwise there is nothing magic about these colours.

I largely disagree with this. While it is true that the colours red, green, and blue are not unique colours and there is some flexibility in choosing the basis colours of the colour space, in the end, the basis colours are still going to be regarded as red, green, and blue. Thus, while there may be some variation in the colours defined as red, green, and blue, red, green, and blue are still the primary colours.

How would it work if we replaced green with yellow (in an additive system)?

The Rev Dodgson said:

How would it work if we replaced green with yellow (in an additive system)?

I don’t think it would. That was my point. But it could still work if the basis colours were changed to something that is close enough to their original colours that we would still regard them as more-or-less the same colour.

Well the stupid ipad won’t quote properly so I’ll ignore that.

Just say we had a requirement to reproduce high-fidelity greens but didn’t need reds. What would you do? You would choose a couple of green primaries at least. This is a contrived example of course but there is nothing fundamentally wrong with it.

The conventional choices make sense because of their large gamut, and the more they are used the more they are conventionalised. But it still is the case that the primaries are relative to a system.

I also find it completely unsatisfactory to say that blue is “the” primary colour and then admit that that could be referring to any of a bazillion different colours. Once we start pretending that these convenient boxes have some objective meaning it’s all over.

Of course it would “work”. You would get higher resolution in your yellows and oranges, but not be able to get much by the way of greens. For most purposes that would be no good but for some it might well be. As I keep saying the choice of primaries depends on what it will be used for.

MartinB said:

Of course it would “work”. You would get higher resolution in your yellows and oranges, but not be able to get much by the way of greens. For most purposes that would be no good but for some it might well be. As I keep saying the choice of primaries depends on what it will be used for.

Yellow + blue = green.

For subtractives but not for additives.

MartinB said:

For subtractives but not for additives.

So?

jjjust moi said:

MartinB said:

For subtractives but not for additives.

So?

The difference between subtractive and additive is huge. For additive, as you keep adding more colour, you get closer to white. For subtractive, as you keep adding more colour, you get closer to black.

KJW said:

jjjust moi said:

MartinB said:

For subtractives but not for additives.

So?

The difference between subtractive and additive is huge. For additive, as you keep adding more colour, you get closer to white. For subtractive, as you keep adding more colour, you get closer to black.

So if I keep adding yellow to blue I get black…………..sorry.

No discussion on colour perception is complete without a look at Edwin Land’s work.

I don’t pretend to have an in depth understanding but he did claim to have invented a two colour projection system that could reproduce all visible colours

jjjust moi said:

KJW said:

jjjust moi said:

So?

The difference between subtractive and additive is huge. For additive, as you keep adding more colour, you get closer to white. For subtractive, as you keep adding more colour, you get closer to black.

So if I keep adding yellow to blue I get black…………..sorry.

But if you add red to it?

Well, you don’t get white, either, which is what you get when you additively combine yellow with blue.

But if you add red to it?

Why, when you are making green?

jjjust moi said:

But if you add red to it?

Why, when you are making green?

I’m talking about the difference between additive and subtractive.

jjjust moi said:

MartinB said:

For subtractives but not for additives.

So?

The question was specifically about additives.

MartinB said:

Of course it would “work”. You would get higher resolution in your yellows and oranges, but not be able to get much by the way of greens. For most purposes that would be no good but for some it might well be. As I keep saying the choice of primaries depends on what it will be used for.

I guess I should have said: would it work in the context of reproducing the full range of colours visible to humans?

No, but no finite gamut does.

MartinB said:

No, but no finite gamut does.

Make it: : would it work in the context of reproducing the full range of colours visible to humans as far as is possible? then.

This thread needs a reference to that book white rabbit used to flog, but my googling is fruitless.

MartinB said:

Well the stupid ipad won’t quote properly so I’ll ignore that.

Just say we had a requirement to reproduce high-fidelity greens but didn’t need reds. What would you do? You would choose a couple of green primaries at least. This is a contrived example of course but there is nothing fundamentally wrong with it.

The conventional choices make sense because of their large gamut, and the more they are used the more they are conventionalised. But it still is the case that the primaries are relative to a system.

I also find it completely unsatisfactory to say that blue is “the” primary colour and then admit that that could be referring to any of a bazillion different colours. Once we start pretending that these convenient boxes have some objective meaning it’s all over.

MartinB said:

Of course it would “work”. You would get higher resolution in your yellows and oranges, but not be able to get much by the way of greens. For most purposes that would be no good but for some it might well be. As I keep saying the choice of primaries depends on what it will be used for.

Ultimately, the notion of the primary colours, ie the bases of the colour space, is not arbitrary and is constrained by the response functions of the retinal receptors.

OCDC said:

This thread needs a reference to that book white rabbit used to flog, but my googling is fruitless.

http://www.goodreads.com/book/show/58565.The_Physics_and_Chemistry_of_Color

now – anyone been looking at them retinal ganglia – them’s fascinating.

http://www.livescience.com/10813-neural-diagrams-explain-color-vision.html

I agree with the specific factual points that Martin is making yet don’t have a problem with people saying there are three primary colours.

Proper artists are always happy to say: FUCK THE COLOUR POLICE

Methods of measuring colour differ and the difference can be important. Kaolin clay is basically white, but it has very slight variations which are crucial to its use in paper coating applications. A company I once worked for did an entire deposit colour evaluation using a method that was not the industry standard. What a complete stuff up that was. It took much manipulation of the data to try to map it into the industry standard method, but it was only ever an approximation.

Mathematically, it is possible to define the notion of primary colours in such a way that the primary colours are uniquely specified, at least for a given individual or rather for a specified set of retinal response functions. To do this, one considers the infinite-dimensional space of functions that are everywhere greater than or equal to zero over all the wavelengths of light that are visible. In this space, the three retinal response functions are normalised, and as such are unit vectors (the space is not actually a vector space, but I’ll use the term “vector” anyway) in this space. Let call these unit vectors R (l), G (l), and B (l), corresponding to the L (l), M (l), and S (l) mentioned in the opening post. One can check that these span a three-dimensional space by checking that they are mutually linearly independent. That is:

R (l) G (l’) B (l’‘) – R (l’) G (l) B (l’‘) + R (l’) G (l’‘) B (l) – R (l’‘) G (l’) B (l) + R (l’‘) G (l) B (l’) – R (l) G (l’‘) B (l’) =/= 0

the products of the three functions being outer (tensor) products. Then, an arbitrary function ƒ (l) (everywhere greater than or equal to zero) can be mapped to the three-dimensional space spanned by the above unit vector by forming the inner product with each of the above unit vectors:

R = ò_l ƒ (l) R (l) dl
G = ò_l ƒ (l) G (l) dl
B = ò_l ƒ (l) B (l) dl

Although this notion of the primary colours is somewhat different to the usual notion, it does uniquely define the primary colours because the definition is entirely in terms of the retinal response functions, which are empirically determined.

The R, G, B values of my previous post represents the entire space of colour vision (in trichromats). Suppose one wants to consider the gamut of a display. Let r (l), g (l), and b (l) be the normalised spectra of the light-emitting or light-reflecting elements of the display. Then the sum:

r ƒ_r (l) + g ƒ_g (l) + b ƒ_b (l)

taken over the space of (r,g,b) values represent the entire displayable spectra. Then:

R (r,g,b) = ò_l (r ƒ_r (l) + g ƒ_g (l) + b ƒ_b (l)) R (l) dl
G (r,g,b) = ò_l (r ƒ_r (l) + g ƒ_g (l) + b ƒ_b (l)) G (l) dl
B (r,g,b) = ò_l (r ƒ_r (l) + g ƒ_g (l) + b ƒ_b (l)) B (l) dl

The mapping of the (r,g,b) space into the (R,G,B) space is the ability of the display to reproduce the range of colours of vision. That is, voids in the (R,G,B) space are colours that cannot be reproduced by the display. Note that a display only needs to reproduce the colours of vision, not the colours of the real world. A display can do no better than cover the entire (R,G,B) space, even if it is capable of producing a larger space of spectra.

KJW said:

Let r (l), g (l), and b (l) be…

That should be:

Let ƒ_r (l), ƒ_g (l), and ƒ_b (l) be…

KJW said:

A display can do no better than cover the entire (R,G,B) space, even if it is capable of producing a larger space of spectra.

Unless you wish to show the budgie something on your computer display. :-P

KJW: since you explicitly are exhibiting primary colours for a specific individual you are doing exactly what I was suggesting.

DV: Clearly there is a cline of usage. I don’t really expect it to be made explicit all the time, but if we don’t remind people every now and again that these are choices then people end up thinking that there are specific magical colours with unique properties.

>>now – anyone been looking at them retinal ganglia – them’s fascinating.

http://www.livescience.com/10813-neural-diagrams-explain-color-vision.html<<

Oh no!! Now you’ve done it! Stop confusing things with complications! They were getting along nicely with the idea of threes………now they have to combine in the early processing steps. Next you will be wanting to go on to the higher processing parts of the visual pathway. And we know next to nix about them.

:)

Well I nearly started answering TRD’s question about why yellow is perceptually so distinct in terms of how the three retinal responses are processed in opponent pairs, but I decided against it…

MartinB said:

Well I nearly started answering TRD’s question about why yellow is perceptually so distinct in terms of how the three retinal responses are processed in opponent pairs, but I decided against it…

Awww …. go on!

I didn’t because I don’t really know this stuff properly, and you can read wikipedia as well as I can :-)

But basically, although the retina receives colour stimulation in three channels R, G, B, they are not the signals that get sent to the brain. Instead the visual pathway produces various combinations of these signals. One of them is something like (and remember I’m not trying to say this is exactly correct) B – (R+G). So although the eye doesn’t ‘see’ yellow directly, what the brain ‘sees’ is Blue<->Yellow in this channel. (The other channel is Red <->Green) and the third is a brightness signal from the rods.)

MartinB said:

I didn’t because I don’t really know this stuff properly, and you can read wikipedia as well as I can :-)

But basically, although the retina receives colour stimulation in three channels R, G, B, they are not the signals that get sent to the brain. Instead the visual pathway produces various combinations of these signals. One of them is something like (and remember I’m not trying to say this is exactly correct) B – (R+G). So although the eye doesn’t ‘see’ yellow directly, what the brain ‘sees’ is Blue<->Yellow in this channel. (The other channel is Red <->Green) and the third is a brightness signal from the rods.)

Thanks, that’s just the sort of thing I realised I didn’t have a clue about, which prompted this thread.

The history of all this (which I’m half way through) is also very interesting.

Aristotle seems to have caused a lot of confusion.

buffy said:

……..now they have to combine in the early processing steps. Next you will be wanting to go on to the higher processing parts of the visual pathway. And we know next to nix about them.

:)

only because a crappy, inverted, reversed, pixellated image without colour in the peripheral regions isn’t much to write home about..

awesome stuff, especially with colour processing being evident in different cortical regions to contrast processing.

http://brain.oxfordjournals.org/content/121/9/1669.full.pdf

and when easily hyponotised subjects are told to not see colour – the brain region for colour processing doesn’t light up in MRI scans on the subjects’ branes.
(previous link) -

>>awesome stuff, especially with colour processing being evident in different cortical regions to contrast processing<<

So, then do we start a thread on movement detection (peripheral retina is better at it than central retina), and edge detection and……..

MartinB said:

KJW: since you explicitly are exhibiting primary colours for a specific individual you are doing exactly what I was suggesting.

Actually, it is specific to a set of retinal response functions. It just happens that there is variation of this amongst different individuals of the population. If there wasn’t, then the definition would apply to the entire population. The point is that the definition isn’t determined by the application, but by the receiving individual over which nobody has any control. The point also is that primary colours do exist as distinct from other colours, even if the specific colours are dependent on the receiving individual.

The point is that the definition isn’t determined by the application, but by the receiving individual over which nobody has any control.

:-/ In this case allowing the receiving individual to see is the application.

The point also is that primary colours do exist as distinct from other colours, even if the specific colours are dependent on the receiving individual.

I’m not sure what the point is. Of course within a given system, primary colours are unique. My point is that different systems have different primaries, and you are agreeing with this.

Obviously all practical systems wanting to cover a large range of the uman visual gamut will have something reddish something blueish and something greenish, and I’ve never disputed that either.

MartinB said:

I didn’t because I don’t really know this stuff properly, and you can read wikipedia as well as I can :-)

But basically, although the retina receives colour stimulation in three channels R, G, B, they are not the signals that get sent to the brain. Instead the visual pathway produces various combinations of these signals. One of them is something like (and remember I’m not trying to say this is exactly correct) B – (R+G). So although the eye doesn’t ‘see’ yellow directly, what the brain ‘sees’ is Blue<->Yellow in this channel. (The other channel is Red <->Green) and the third is a brightness signal from the rods.)

In analogue colour television (PAL system), the three primary colours are not handled symmetrically. This is because colour television had to be backwards compatible with the original B&W standard. The main signal is the brightness signal as it was in the B&W standard, but a new signal is superimposed onto this using a 4.43361875 MHz subcarrier. The red and blue signals are represented by the sine and cosine components of this signal (I don’t recall which colour is sine), the phase being relative to a subcarrier burst on the horizontal sync pulse. The green signal is then extracted as:

green = brightness – red – blue

I always thought this was a rather ingenious way to overcome both the limitation of backwards compatibility with B&W and the limitation of a (synchronised) carrier signal to two dimensions of analogue information.