Photoshop: Why don't colours with equal amount of light in RGB don't convert to the same grey

  • 1
  • Question
  • Updated 2 years ago
  • (Edited)
I thought, R140-G0-B0 and R0-G70-B70 mean the same amount of light, thus the same lightness, and should accordingly convert into the same grey. Yet they don't. Why?*, **, ***

Thanks a lot.

* Does that have to do with the famous Weber–Fechner law, or Fechner's scale (to my understanding: the different colours are perceived as being of different lightness, and maybe the conversion to grey is done perceptively, which would mean that the greys that turn out don't reflect the lightness)? I can't find anything on the matter.
** Not a merely theoretical question, One of its applications seems to be the use of the Hue/Saturation adjustments, where changing hue actually changes the grey value (or the grey conversion outcome for that matter), with obvious practical implications, but with import on the understanding of what hue and lightness really are as well. I suppose in common understanding the grey value equals lightness and nothing else. That, in a next step, has implications for the understanding of the CIELCh or CIEHLC colour space and thereby on the Lab colour space, and so on.
*** I asked the same question here, and got referred to this present site.
Photo of Nikolai Franke

Nikolai Franke

  • 2 Posts
  • 1 Reply Like

Posted 2 years ago

  • 1
Photo of christoph pfaffenbichler

christoph pfaffenbichler, Champion

  • 1208 Posts
  • 166 Reply Likes
I thought, R140-G0-B0 and R0-G70-B70 mean the same amount of light, thus the same lightness
In which Color Space is that supposed to be the case? 
Photo of Nikolai Franke

Nikolai Franke

  • 2 Posts
  • 1 Reply Like
Hello Christoph, thanks for your swift reply. I thought that to be generally the case, regardless of the colour space. Apparently, my non-comprehension roams deep. I used and tried this in Adobe RGB.

I'd totally be willing to read something on the matter, but found nothing as of yet. Maybe the relatively simple explanation is that the, let's say: perceived brightness of any RGB-described colour doesn't equal the numeric amount of light, and the conversion to grey uses the perceived lightness as described in the Lab space ... ? 

Thanks again. 
Photo of christoph pfaffenbichler

christoph pfaffenbichler, Champion

  • 1208 Posts
  • 166 Reply Likes
Try reading this: 
http://www.gballard.net/psd/cmstheory.html

An RGB value has no luminance, chroma or hue per se, only if the Color Space is known can the corresponding Lab value be determined. 
Photoshop uses the RGB Working Space as the basis of the calculation if an image is untagged. 
Photo of Mark Heaps

Mark Heaps, Champion

  • 62 Posts
  • 24 Reply Likes
Hi Nikolai,

lots of factors at play here, but even though the math both add up to 140 as a sum, they are defined in different parts of the RGB gamut. And the actual spacial differences between Red Green Blue aren't equal.

For example, most of the time when we teach about a gamut people see something like this. Although this is helpful for understanding how to get CMY out of RGB, it's not an accurate depiction of the RGB gamut.



In reality, RGB gamut is more accurately shaped like this.



So when Adobe applies a conversion to grayscale (there are MANY ways to do this which can affect results. Each applying their own approach. (Convert to grayscale image mode, Desaturation, Channel Mixer, etc...) But because the three channels aren't equal in capacity even though they are dumbed down into 8bit data, the math doesn't translate equally. You can see in the image above, RED is a much smaller area than say Green.

Bring up your color picker sometime and turn on gamut warning. Then scroll through and see how much is grayed out. You'll be surprised at the differences.

So mathematically you're creating equal sums, but not equal proportions as they are relative to space in the gamut.

Hope that helps.

Mark Heaps
@lifebypixels