Spectrum Locus vs. Colour Spaces


I was reading a post here entitled Comparing RGB Working Spaces and I watched the video recommended in that post by Mr. Andrew Rodney. I have watched that video after reading that post once before. But, this time as the video got to the section about Spectrum Locus, two questions came to mind for which I can find no clearly specific answers, as yet. They are as follows:

Since the spectrum Locus which describes all of the colours visible to the average human observer can be plotted using numerical values why do we need additional subset Colour Spaces such as sRGB, Adobe RGB, ProPhoto RBG and many others? Why do we not simply relate all colour gamuts to their positions within the Spectrum Locus?

Well, I suppose this has been widely considered to be either a very stupid query or it has everyone stumped. If the former is the case, I apologize for not knowing everything. If the latter is the case, I look forward to seeing any thoughts that may lead to an answer. Thanks in advance.

I have not watched Andrew’s video you mention, but I’ll wade in and give it a shot. As I understand it, the spectrum locus is a 2-dimensional representation of all the colors of a human observer. So in a sense, it’s accurate to say you can represent those colors on the spectrum locus. But’s its like the difference between locating color on a 2D graph vs. a 3D graph in ColorThink. To locate a color numerically using not only color information but luminance also, we use 3D CIE Lab space.

I appreciate this Pat. Based on this post, I think I’m missing something about all this. So, Ill have to explore this more deeply.

I hope I can help here. Sorry about the late reply.

The spectrum locus is the path that the pure-color spectrum traces over a graph.

So in the case of the Yxy chromaticity diagram (horseshoe diagram) the Spectrum Locus would be the path from purple, in the lower left corner, around the outer edge clockwise, to red in the lower right corner.

So not only does it not contain all the colors visible to humans but it doesn’t even contain all the hues. The mixing hues between purple and red (magenta and friends) doesn’t appear in the rainbow.

Does this help?


Thank you for this input, Steve. I knew, when I first posted this, that I was entering a realm of colour science that was well beyond my knowledge. I am working on improving that. In the past, I have been successful at learning various concepts on other topics by reading as much as I can find, wrestling with the available information, testing, experimenting and then jumping into a discussion with others who are in the know for better or worse to test what I have learned. This subject, beginning with my original question, as at the top of this thread, has been a real challenge to explore.

After posting the question here, and reading Pats reply, I asked it again in a forum at Cambridge in Colour.com. A great deal of interesting input developed there on this topic and I eventually realised that I had found an answer to my original question. While that answer still seems appropriate, your post here provides a differing description of the nature of the so called Spectrum Locus or, as some members of the Cambridge forum vehemently preferred, Spectral Locus.

As I mentioned in my original post, I had watched, several times, a video made by Mr. Andrew Rodney called Everything you thought you wanted to know about Colour Gamut and RGB Working Spaces My original understanding of the term Spectrum Locus came from that video in which Mr. Rodney referred several times to the horseshoe shaped plot as the Spectrum Locus and the Spectrum Locus as The gamut of human vision and added that Anything that falls outside this particular horseshoe shape is not visible to us and When we talk about gamut this is the big gamut that everything else is contained within. But, if I understand you correctly, here you are saying that the "Spectrum Locus does not, in fact, contain all the colours or hues in human vision.

Im sure there must be something I am missing or about which I’m confused. So, I would appreciate it if you would help me with this, when you have a moment.

Thanks again for your time.

By the way, Id like to take this opportunity to say that ColorThink is an extraordinary tool. As Pat may recall and be able to tell you, I have raved about it to him several times since we first purchased it a few years back. Its amazing. I am so thankful that you developed it. We love it. Cant imagine being without it.


So, it looks like I have found an answer to my question and have expressed it as follows:

The Spectrum-Locus is a 2 dimensional plot of the the results of certain visual stimuli to the brain of the standard human observer which are defined as colours. It is not a mathematical grid or matrix designed to suit the functionality of machines. It is rather a map or record of the average human’s response to a range of frequencies of light which, as visual stimuli, create responses in the brain which are subsequently defined as colours.

Conversely, RGB working spaces are mathematical matrices of RGB numerical values which are derived from and bounded by the three primaries, include a white point and Tone Reproduction Curve/Gamma, are plotted in 3 dimensions and are designed to encompass portions of the Spectrum-Locus. As such, RGB working spaces accommodate the functionality of RGB capture and output devices.

I wrote and posted this in the other forum which I mentioned above and it was approved. Nevertheless, I invite anyone who feels that any portion of it may be incorrect to correct it. I’m sure it could be said better.

I’d like to thank everyone who helped me to arrive at this.