View previous topic | View next topic

Believe in Pink!

Page 1 of 1

Leith
1106084.  Tue Dec 16, 2014 8:06 pm Reply with quote

I really enjoyed the 'Lying' episode. Nice to see Sarah Pascoe and Adam Hills back, and the panel seemed to work well together.

It did bring up one of my pet annoyances, though, in the form of the 'pink isn't a real colour' nonsense.

Pink is not a colour whose perception can be triggered with light of a single frequency. That puts it in the same category as purples, black, white and everything inbetween ('non-spectral colours'). Are they not real colours either?

A perception of cyan can be triggered with single-frequency light at around 600 THz, or by twin-frequency light at 550 THz and 630 THz. Is the first cyan real and the second cyan not real?

 
Leith
1106085.  Tue Dec 16, 2014 8:07 pm Reply with quote

From the episode's accompanying web page:
QI wrote:
‘Pink’ is desaturated red, but it approximates in common usage to magenta, a colour which is odd in that it ‘doesn’t exist’ in nature.

Fuchsias.

Colours are not a fundamental property of physics. They are a function of our visual perception that provide us information about the natural world we evolved in.
Pink things exist everywhere in nature, and the light by which we perceive them cannot sensibly be characterised as anything other than pink.
Conversely, perfect single frequencies of light are difficult to find in nature and are not the normal cause of the colours we see.

QI wrote:
For magenta light to exist, it would have to have a wavelength longer than red and shorter than blue, which is clearly impossible.

Our eyes respond to varying intensities of light across frequency spectra, not to discrete frequencies. Magenta can be easily reproduced with light of a suitable frequency spectrum. Modelling colour vision with single frequencies of light is clearly impossible.

QI wrote:
So magenta, which is sometimes loosely described as ‘pink’, is a construct of the brain (‘a pigment of the imagination’): when we see red and blue light together, the brain interprets it as magenta.

The brain isn't 'interpreting' anything here, at least not in any way that differs from how it processes spectral colours. In simplistic terms, what the brain receives is effectively a set of intensities* from each of its three cone types - something like the 'RGB value' (red,green,blue) used to specify colour on a computer monitor.

* If I remember right, it's actually relative intensity differences rather than absolute intensities, but that makes no difference for the purposes of this argument.

In RGB:
- Magenta is represented as (255, 0, 255) - full intensity red and blue, no green.
- Cyan is represented as (0,255,255) - full intensity green and blue, no red.

A perception of magenta is how the brain responds to the first pattern of intensities, while a perception of cyan is how it responds to the second.

Because of the way our cone cells overlap in their frequency responses, the cyan pattern could theoretically be triggered by 600 THz light triggering the 'inside edges' of two cones where they overlap, or by 550 THz light triggering the 'outside edge' or one cone while 630 THz light triggers the outside edge of the other.



In the latter case, the brain is not recognising the pair of light frequencies and somehow calculating an inbetween frequency from them. At this level of processing, the brain has no model of what frequencies are, or that they occur on a scale. There is no physical reason why 600 THz light should look the same as 550 THz light + 630 THz light. It's just that the physiology of our eyes doesn't allow us to tell the difference.

Staying with the simplified RGB model, our range of perceived colours occupy a three-dimensional space with a red axis, a blue axis and a green axis. It should be no surprise that a one-dimensional frequency scale can't be used to fully describe this range.


Last edited by Leith on Tue Dec 16, 2014 8:09 pm; edited 1 time in total

 
Leith
1106086.  Tue Dec 16, 2014 8:07 pm Reply with quote

For some genuinely* impossible colours, see:
http://en.wikipedia.org/wiki/Impossible_color

* for a given value of 'genuinely'

 
14-11-2014
1106097.  Tue Dec 16, 2014 11:16 pm Reply with quote

Leith wrote:
Fuchsias.


A grey natural human hair doesn't exist, so perhaps you may have to do better than just mentioning flowers which look like pink flowers. With nothing but true pink (anthocyanin) pigments, i.e. not a combination of other pigments. Please note that it's likely that pink pigments do exist, but grandpa Joe doesn't prove that a single hair can be grey.

 
CharliesDragon
1106105.  Tue Dec 16, 2014 11:54 pm Reply with quote

As for an impossible colour I have seen, I offer green-purple. Looking at a colour wheel that is surely not possible in any way, and my eyes usually struggle to accept it, too. It's sometimes seen in storm clouds.

 
Zziggy
1106160.  Wed Dec 17, 2014 7:08 am Reply with quote

My understanding was that pink (well actually I heard it as purple) is 'imaginary' because it is a combination of red and blue.

I was told that our eyes can see red, blue and green, and other colours we see are made up of combinations of these colours, so that for instance when we see something yellow it is a result of our eyes picking up red and picking up green and going for the colour in between: yellow. However when we see pink/purple, our eyes see red and blue, but they are on different ends of the light spectrum so there is no 'middle' to plonk for, so our brain 'makes up' pink/purple.

I'm not sure if this helps/is completely incorrect?

 
Leith
1106217.  Wed Dec 17, 2014 3:10 pm Reply with quote

Yes, that's the basis for claiming pink as 'imaginary', Zziggy. I'm arguing that that's based on a fundamental misunderstanding of how colour vision works (see second post on the thread)

Our eyes / brains don't measure light frequency in the way suggested. The signal our brain receives can roughly be described as 'type 1 light at intensity x, type 2 light at intensity y and type 3 light at intensity z'. Our visual system has no direct data on the frequency of light, doesn't know what a frequency spectrum is and can't tell if the frequency of type 1 light is higher than, lower than, close to or far away from that of type 2 or type 3 light. The idea that the brain 'picks a colour in between' in a way that has anything to do with light frequencies is a philosophical error.

A whole range of different light spectra can produce a perception of yellow, and different range of spectra can produce a perception of pink. The fact that one range includes a spectrum with only a single frequency and the other doesn't makes no different to how real each colour is.

 
Zziggy
1106218.  Wed Dec 17, 2014 3:25 pm Reply with quote

I did read that post but I err... well I didn't really understand it. Sorry :(

So it's just 'imaginary' because it's a colour which we can see but isn't in the visible light spectrum?

 
'yorz
1106227.  Wed Dec 17, 2014 3:59 pm Reply with quote

rant deleted


Last edited by 'yorz on Wed Dec 17, 2014 5:45 pm; edited 1 time in total

 
Leith
1106228.  Wed Dec 17, 2014 3:59 pm Reply with quote

Zziggy wrote:
I did read that post but I err... well I didn't really understand it. Sorry :(

So it's just 'imaginary' because it's a colour which we can see but isn't in the visible light spectrum?


Pink is entirely within the visible light spectrum. It just can't be created with single frequency light. It's more like a musical chord than a single note, if that makes any more sense. A limited range of other colours be can artificially reproduced with single frequency light, but that's rarely how they occur naturally.

When physicists started trying to analyse the make up of light, it was natural for them to use colours to identify isolated light frequencies. Colours describe light frequencies well, but light frequencies don't describe colours well, at least where human vision is concerned (and human vision is the only place in which our notion of colours has real meaning).

 
Zziggy
1106236.  Wed Dec 17, 2014 4:21 pm Reply with quote

Ah ok, I think I see now. Thanks!

 
Leith
1106254.  Wed Dec 17, 2014 5:22 pm Reply with quote

14-11-2014 wrote:
Leith wrote:
Fuchsias.


A grey natural human hair doesn't exist, so perhaps you may have to do better than just mentioning flowers which look like pink flowers.
...

Yes, that was a little facetious; I did back it up with a few other words, though! For those that don't want to wade back through, this was my assertion:
Leith wrote:
Pink things exist everywhere in nature, and the light by which we perceive them cannot sensibly be characterised as anything other than pink.


14-11-2014 wrote:

...
With nothing but true pink (anthocyanin) pigments, i.e. not a combination of other pigments. Please note that it's likely that pink pigments do exist, but grandpa Joe doesn't prove that a single hair can be grey.

Just as no single pixel on my monitor is yellow, yet most of my screen currently is. Is it real yellow or unreal yellow? How large or small an area of your visual field must the light be averaged over for its colour to be real?

Pigments are another candidate for something to which we can anchor our notion of 'what a colour is'. But just as with light, it's difficult to isolate simple physical properties of pigments or surfaces with which to explain their perceived colour.

 

Page 1 of 1

All times are GMT - 5 Hours


Display posts from previous:   

Search Search Forums

Powered by phpBB © 2001, 2002 phpBB Group