APOD: Elephant's Trunk and Caravan (2021 Aug 27)

Comments and questions about the APOD on the main view screen.
User avatar
Chris Peterson
Abominable Snowman
Posts: 16128
Joined: Wed Jan 31, 2007 11:13 pm
Location: Guffey, Colorado, USA

Re: APOD: Elephant's Trunk and Caravan (2021 Aug 27)

Post by Chris Peterson » Sat Aug 28, 2021 1:24 pm

Ann wrote:
Sat Aug 28, 2021 5:08 am
Filter channels should not normally be mapped as orange, violet or cyan, and absolutely not as brown. The reason for that is that our eyes have no receptors for orange, violet, cyan or brown. We have receptors for red, green and blue light, and all the colors that we see are a combination of stimulating our red-, green- and blue-sensitive receptors to various degrees.
Again, I feel it necessary to correct a lot of misunderstandings.

Color mapped images in general are of little scientific value. They might be simply aesthetic, or they might help us see interesting structure that warrants further examination in the individual data channels. The colors chosen are arbitrary (other than aesthetics). There is no reason they should correspond to the order of the wavelengths in the original data. The goal is to present information, and the careful choice of colors can make more structure apparent.

Our eyes do not have receptors for red, green, and blue light. They have what are called long, medium, and short wavelength receptors, which peak, respectively, in the yellow, cyan, and violet. We see color because of the large overlap in the sensitivity curves of these receptors. They do not match at all the red, green, and blue primaries commonly used for our display devices.

The only advantage to mapping three different data channels to red, green, and blue is that those channels can be separated again into their original data. If we map to any other color, it means we are mixing the data into two or more primary channels, and that's not a reversible operation. And, of course, if there are more than three channels of data, which is often the case, there's no choice but to mix them in a way that cannot be separated again.

On a display, orange and brown are not different hues. They are the same, with just a different intensity or saturation. In either case, you're just mapping your data into a mix of red and green.
Chris

*****************************************
Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

User avatar
Ann
4725 Å
Posts: 11513
Joined: Sat May 29, 2010 5:33 am

Re: APOD: Elephant's Trunk and Caravan (2021 Aug 27)

Post by Ann » Sat Aug 28, 2021 7:02 pm

Chris Peterson wrote:
Sat Aug 28, 2021 1:24 pm
Ann wrote:
Sat Aug 28, 2021 5:08 am
Filter channels should not normally be mapped as orange, violet or cyan, and absolutely not as brown. The reason for that is that our eyes have no receptors for orange, violet, cyan or brown. We have receptors for red, green and blue light, and all the colors that we see are a combination of stimulating our red-, green- and blue-sensitive receptors to various degrees.
Again, I feel it necessary to correct a lot of misunderstandings.

Color mapped images in general are of little scientific value. They might be simply aesthetic, or they might help us see interesting structure that warrants further examination in the individual data channels. The colors chosen are arbitrary (other than aesthetics). There is no reason they should correspond to the order of the wavelengths in the original data. The goal is to present information, and the careful choice of colors can make more structure apparent.

Our eyes do not have receptors for red, green, and blue light. They have what are called long, medium, and short wavelength receptors, which peak, respectively, in the yellow, cyan, and violet. We see color because of the large overlap in the sensitivity curves of these receptors. They do not match at all the red, green, and blue primaries commonly used for our display devices.

The only advantage to mapping three different data channels to red, green, and blue is that those channels can be separated again into their original data. If we map to any other color, it means we are mixing the data into two or more primary channels, and that's not a reversible operation. And, of course, if there are more than three channels of data, which is often the case, there's no choice but to mix them in a way that cannot be separated again.

On a display, orange and brown are not different hues. They are the same, with just a different intensity or saturation. In either case, you're just mapping your data into a mix of red and green.

Wow. Thanks for the lecture, Chris. I feel as if I was getting bowled over by the power of your words. By the way, in the illustration I posted, our sexes are obviously reversed.

As for orange and brown being identical hues, just different degrees of saturation: Yes, I actually knew that, at least!

Ann
Color Commentator