What is meant by monochrome?

[TonyW]

When I use photoshop to convert an image to black and white, at any point in the image the levels of red, green and blue channels on a scale from 0 to 255 are the same. This is still true when I convert from AdobeRGB to sRGB. But when I introduce a tint like a sepia, the relationship between the levels is not at all obvious.

Are the colours in a monochrome image adjusted according to some subjective idea of points of different luminosity being of the same colour or is there a well defined arithmetical relationship?

I have had a picture of mine called monochrome when I have done nothing to make it so and conversely I have looked at a print of a supposedly monochrome picture and seen some colorisation.

[IzzieK]

From what I understand, "monochrome" means "one colour" so in my mind, black and white is a monochrome and sepia is another. Some just have a tint of blue or green and nothing else in between...So an image that is black and white but selectively chosing another part with another colour, e.g., Trevor's The Deer Departed will no longer be considered monochrome.

[Mike Buckley]

Are the colours in a monochrome image adjusted according to some subjective idea of points of different luminosity being of the same colour or is there a well defined arithmetical relationship?

When you are tinting a black-and-white image, it's definitely subjective. As an example, you can apply the tint differently to the bright tones than the dark tones depending on the look you want to achieve. I don't know if this can be done in Lightroom but it can be done in Google/Nik Silver Efex Pro 2.

[xpatUSA]

When I use photoshop to convert an image to black and white, at any point in the image the levels of red, green and blue channels on a scale from 0 to 255 are the same. This is still true when I convert from AdobeRGB to sRGB. But when I introduce a tint like a sepia, the relationship between the levels is not at all obvious.

Are the colours in a monochrome image adjusted according to some subjective idea of points of different luminosity being of the same colour or is there a well defined arithmetical relationship?

I have had a picture of mine called monochrome when I have done nothing to make it so and conversely I have looked at a print of a supposedly monochrome picture and seen some colorisation.

Tony,

Firstly, we all know that "monochrome" means one color (from the Greek).

However, while 'mono' is easy enough, 'color' is a different matter. I find it easier to think of a 'color' as being a particular angle around the additive color wheel - rather than trying to figure out the RGB numbers. From that point of view, it is best to think of the HSB color space, shown as a cylinder. Lets say we choose an angle around the cylinder edge of 265 degrees, which I might call violet and you might call mauve. If we keep that angle constant for every pixel in an image, we can say it is mono-hued i.e. mono-colored i.e. monochromatic. Therefore, any variation in the distance from the center (Saturation) or any variation up-and-down (Brightness) does not affect the Hue. I just now made this monochrome image on that basis:



Did it in Elements - drew some black lines, then kept changing the foreground HSB saturation and brightness (but not the hue) followed by the filler tool each time.

The semantic trap here is that the definition of 'a color' is complicated and varied, leading us into CIE stuff and various color spaces.

But just 'color' can be defined simply by hue: the sensation of color.

I think sepia is a subset of monochrome. By which I mean that, in sepia, the hue is held constant at some sort of brown; the saturation is perhaps held at some value pleasing to the eye and the brightness is varied to form the image. Might be wrong there, apart from the constant hue. Sepia experts please comment.

I can't think of any way to easily consider monochrome channel values in RGB space mathematically, other than the trivial example of gray-scale. That is to say, if you put two sets of channel values side-by-side, it's hard to know intuitively whether they are the same hue or not.

In CIELAB space, monochrome could be considered as a constant angle between a* and b* with the magnitude of the vector (chroma) and the Lightness L* being variable.

Setting the above pedantry to one side, in the Real Word an image could have some hue variation but still appear to be monochrome. That is because distinguishing small changes of hue is quite difficult for the eye; or small changes of anything for that matter. Google "Just Noticeable Difference" . . .

[davidedric]

I think many of use our software to manage printing colours, I do with Lightroom, and turn off colour management in the printer. I understand that the printer may use the coloured inks to make a b&w print, and that can lead to a colour cast. I have been told that with b&w it is better to lot the printer do the colour management. Can any printing experts comment?

Dave

[xpatUSA]

. . . or is there a well defined arithmetical relationship?

There's an expression for hue which, if you've accepted my earlier definition of 'monochromatic' may be of help:

<image deleted>

CiC might balk at .svg images, but your browser should be able to show it (FireFox didn't, something to do with CiC, I imagine).

Otherwise, the formula for hue from RGB is buried in here:

https://en.wikipedia.org/wiki/HSL_an...Hue_and_chroma
.

[Rent]

Dave I usually take the photo in monochrome and then print in monochrome But I have had a print turn brown then yellow over about 12 months I tried one framed behind glass and in over 5 years its not faded
Roy

[DanK]

I think many of use our software to manage printing colours, I do with Lightroom, and turn off colour management in the printer. I understand that the printer may use the coloured inks to make a b&w print, and that can lead to a colour cast. I have been told that with b&w it is better to lot the printer do the colour management. Can any printing experts comment?

Dave

This is correct, at least for some printers. It is true of my Canon Pixma Pro 100, and a few people have said the same about their Epson printers. Tim Gray, who has a very good photo Q&A online as well as his commercial products, also recommends letting the printer control B&W printing for this reason.

However, this obviously won't work for mononchromes other than grayscale.

[Manfred M]

This is correct, at least for some printers. It is true of my Canon Pixma Pro 100, and a few people have said the same about their Epson printers. Tim Gray, who has a very good photo Q&A online as well as his commercial products, also recommends letting the printer control B&W printing for this reason.

However, this obviously won't work for mononchromes other than grayscale.

Dan - that is definitely my experience as well with an Epson printer.

If I let Photoshop mange the colours for a B&W print, I seem to get a green colour cast. Letting the printer take care of that gives me a nice neutral B&W print.

[John 2]

Just to add to the above comments on printers, I also let the printer control colour/B&W but I find that the make of paper has an effect. Consequently I have a different set up say for Ilford than that for Permajet and that's even true for different surfaces to a degree. This is presumably why manufacturers offer profiles although I have never used them.

[Manfred M]

Just to add to the above comments on printers, I also let the printer control colour/B&W but I find that the make of paper has an effect. Consequently I have a different set up say for Ilford than that for Permajet and that's even true for different surfaces to a degree. This is presumably why manufacturers offer profiles although I have never used them.

That is one of the reasons manufacturers provide profiles (and an important one), the other reason is as part of a colour managed workflow, it ensures that the colours come out being consistent between various papers and printers.

[davidedric]

I mostly use Permajet papers, and have them generate custom profiles for my printer.

When I got a new printer a few months ago I inadvertently had both the printer and Lightroom doing colour management (multiple occurrences of the printer in Control Panel). I could see something was out of kilter, so I printed off their their profile generating image and posted it off only to get a message saying it was impossible to create a profile.

Just posted for interest.

Dave

[TonyW]

Thanks for all the replies. I am more than ever convinced that I don't want to try to do my own printing at this stage. It is another big subject that I would have to get on top of and probably more expensive than using a good commercial printer as I do now.

There's an expression for hue which, if you've accepted my earlier definition of 'monochromatic' may be of help:

<image deleted>

CiC might balk at .svg images, but your browser should be able to show it (FireFox didn't, something to do with CiC, I imagine).

Otherwise, the formula for hue from RGB is buried in here:



https://en.wikipedia.org/wiki/HSL_an...Hue_and_chroma
.

Ted, thanks for the link. I think the answer to my question is buried there but I will have to work on it to properly understand what is going on. At least it seems that the subject is not simple.

[xpatUSA]

Ted, thanks for the link. I think the answer to my question is buried there but I will have to work on it to properly understand what is going on. At least it seems that the subject is not simple.

I probably didn't help with my messed up post and it's .svg image. Here it is again with the mathematical formula for hue circled, which I think is what you were asking about:



How on earth they come up with that formula, I have no idea, but I'm sure it ain't simple.

The hexagon shape is the truest representation of RGBCMY but, typically, they had to dumb it down for the punters so that it looks kinda like an artists color wheel :-(

Hope this helps rather than hinders . . .

[TonyW]

Ted, I stared at that diagram for a long time and I think I finally worked it out. It seems like a reasonable approach but it would be nice to see how is agrees with a subjective judgement.

The hue H as defined here is not quite the angle but the scaled distance around the perimeter of the hexagon rather than around the perimeter of the circumscribing circle.

[xpatUSA]

Ted, I stared at that diagram for a long time and I think I finally worked it out. It seems like a reasonable approach but it would be nice to see how is agrees with a subjective judgement.

What you could do is plug those RGB fractions times 255 into the Adobe foreground dialog and a) see how it looks and b) see what hue angle it gives. In Elements 6 it gives 200 deg.

I would suggest that it might be more meaningful to see how subjective judgements agree with a particular hue angle, there being so many different names for a particular hue and so many perceptions due to variations in human vision especially them folks who have four sets of cones not the regular three . . .

The hue H as defined here is not quite the angle but the scaled distance around the perimeter of the hexagon rather than around the perimeter of the circumscribing circle.

Yes, in the Wiki link I gave it explains about the errors. The hexagon is the truer representation, being the projection of the RGB cube but the circles, cylinders or cones are easier to grasp, I guess.

[TonyW]

This is probably more detailed than most people are interested in and I apologise in advance.

I just discovered the simple arithmetic relationship that I was looking for. I converted an image to monochrome (black and white with a tint) in photoshop. At various points in the image I calculated the average value (ranging from 0 to 255) of the respective colour channels. In the monochrome image, the difference between values for each colour and the average were the same for each point. For example, one point had values of 134, 115, 90 for the red, green and blue channels respectively. The average was 113. The differences were then 21, 2, -23. At another point, the colour values were 195, 176, 151 with an average of 174. Again the differences were 21, 2 -23.

The monochrome image did look monochrome to my subjective eye.

[Mike Buckley]

I understand how that would have been true if you had applied the same tint to the entire image. However, if you applied it differently to the bright tones than the dark tones, I would expect the mathematical relationship you came upon to no longer being constant.

[TonyW]

Mike, I was a little surprised too. I expected the differences to be proportional but not the same. I did not do anything really special in the conversion, just clicked on the tint option in the photoshop dialogue box.

[xpatUSA]

It looks to me like applying a tint in PS results in a simple hue rotation by a fixed angle everywhere in the image. In the case quoted, both sets of RGB numbers give the same hue angle of 34 degs (orange-ish brown) which supports my view. Not only that but multiplying the three RGB numbers by some factor e.g. 0.8 does not change the hue either.

But surely a tint is a tint, and if "applied differently" to various parts of the image - wouldn't that result in a multi-chromatic image?

I don't use PS, so my understanding may be a little off as to what "applied differently" means . .

Meanwhile, there is a better formula (Java code) here:

http://stackoverflow.com/questions/2...b-value-to-hsv
.