Where exactly is pure blue (in nanometers)?

[hl_world]

Which wavelength of light will look blue without any bias toward cyan or violet? I know it's somewhere around 460nm but I don't know exactly where it is.

And if you happen to know the same for green (not in any way cyanish or yellowish), please let me know.

 

[Danger]

Not only do I not know the answer to that, but I'm not sure that anyone does. My optical receptors are not identical to yours, nor to anyone else's, so who can say whether or not we perceive the same thing?
My best recommendation is to obtain a Pantone colour palette and use some instrumentation such as a spectrometer to measure the wavelengths of various colours which are defined in the palette (booklet).

 

[K^2]

The only objective answer is the reception peak of the "blue" cone cell, which is around 430nm. But it might not appear "pure blue", and is probably going to depend on individual, as Danger suggested.

 

[hl_world]

Thanks for the responses.


@Danger
I accept that there are differences in eye rods and visual cortices but there must be some reasonable degree of accuracy in where a group of people will say they see the purest blue.

The problem with the palette/spectrometer thing is that it would reflect wideband light making it hard to determine from that the dominant wavelength. Even what appears to be the same colour can have very different radiation spectra. I suppose a better idea would be to deal with pure colours prism-dispersed from sun light with refraction index tests to determine wavelengths.

@K^2
Yeah, 430nm is violet blue like the blue band from flourescent light (~435nm) but even closer to violet.


There must be some sort of gaussian curve if you imagine a graph where wavelength (in increments of 1nm) is on the x-axis and the number of people who nominated the respective wavelength on the y axis. But we are also so similar which I guess would make the curve narrower. I just wondered if this had been done with a group of human subjects.

off topic: are you the K^2 from gtaforums?

 

[JDługosz]

See diagrams http://en.wikipedia.org/wiki/CIE_19...maticity_diagram_and_the_CIE_xyY_color_space". Looks to be around 475, with lower values being "purplish blue", etc.

 

[hl_world]

475nm is way too cyanish. In fact, it's probably on the border between azure and sky blue. The problem with referring to the chromaticity diagram or any color on a screen is that it's all presented in RGB values and the accuracy of how it's coloured takes second place to the accuracy of its mathematical values (curves). It can only be good for rough approximations.

 

[Dr Lots-o'watts]

Here's what NASA says:

http://eosweb.larc.nasa.gov/EDDOCS/Wavelengths_for_Colors.html#blue

 

[Danger]

That's a great link, Doc. Thanks.

 

[hl_world]

475nm is bluish like this [●] but it isn't pure blue like this [●]. The NASA source you linked only gives rough values. I'm looking for the values of unbiased pure colours give or take 2 or 3 nanometers.

● 590nm: not orange like it says but amber (way closer to yellow than orange). In fact, it's what traffic signals use for yellow.
● 570nm: very close to yellow [●], but easily biased to green.
● 510nm: actually that bluish green that traffic lights use.