Colour, Saturation and the Common Tomato (Pt 1)

Tim Tucker

No longer a newbie, moving up!
Mar 23, 2015
Reaction score
Can others edit my Photos
Photos NOT OK to edit
I'd like to discuss and show you something that I think is not well enough understood in photography. I'll start by going back to my main man in the white suit.

A man in a white suit only stands out in a sea of grey

So how do we make him stand out more? If we make him whiter we have to start removing the shading (grey) that makes up the texture and shading that gives both the impression of light and the 3D effect. Simply because you can't make the white parts whiter you are forced to make the parts that are not quite white whiter. So adding white is really about subtracting the parts that are not white, you remove tone and texture from your image.
You make him stand out more by making everybody else greyer. Think about this because we have not changed our subject at all, we make him whiter simply by altering the background. We change how he contrasts with the background and that changes how we see him, and his suit without actually changing the suit at all.
It's easy to see and understand this with a B&W photo, but I'd like to try and demonstrate this, and why it works in colour. I don't have a white suit, but I do have a tomato or two.

This is going to take a fair few images so I've kept them fairly small. For this you're going to need to download them and be able to both view them in isolation and flick between them quickly and easily. So you'll need an image viewer set to full screen with all the images at actual size.
What we're going to do is compare them in a way that hopefully will give you a partial understanding of the way our eyes work. For our first image I've chosen a favourite, a backlit subject. Or one where the main colour pointing towards the camera is in shade, and I've kept the background simple and neutral. Here are the first two images:





Open them both in your image viewer set at full screen with the images at actual size. Have
ex-1_50%_saturated.jpg showing and look at it for at least 30sec to a minute. Now flick briefly, for a second or two to ex-1_my-edit.jpg and back again. ex-1_my-edit.jpg will look dull and colourless.

Now reverse this and have ex-1_my-edit.jpg on your screen for at least 30sec to a minute. Now flick briefly, for a second or two to ex-1_50%_saturated.jpg. Now ex-1_50%_saturated.jpg looks over-saturated and luminous.

Why? Because we see relative values, (the difference in things), not absolute values and our brain compensates, (adjusts), the image to how we remember it should look. (If you don't think that's true then try to picture a blue tomato. See? It's easy to picture a red one). When we look at one image in isolation for a while it becomes the norm against which the other is judged, when we flick briefly to the other we register the difference relative to the norm. We see only if one is more or less saturated and the values of those colours change depending on the sequence in which we view them. Sorry about the bold, but it's worth drawing attention to.

So then there's no real correct way of processing because everything's valid, oh, and what has this got to do with the white suit?

No, not really, and I'll get to the suit in a minute. Let's say you're sharpening an image, you tweak up the sharpening and flick back to see how it compares to the edit before the action. Even if the edit before is optimal it will still look softer than the over-sharpened image because we only see the difference, and you can begin to see how images become over-sharpened. Also this over-processing works best if you're lazy with your vision, if you just glance at something and allow your brain to compensate and correct the inconsistencies. But not everybody is lazy with their vision and not all images are viewed in isolation.
I'm not saying that there 's a correct way of doing things, or that your image needs to look correct because there isn't and it doesn't. It only needs to look convincing. I'm only saying that it's you who has the memory of the original scene, because it's you who saw it with your own eyes. It's also you who has been looking at the image in isolation and so it's normalised to you, combine this with your memory of the scene and your brain correcting, especially if you're lazy with your vision (remember the blue tomato?). We as the viewers have just flicked to your image after looking at another and we have memories of maybe similar scenes that could look different to the one you present.
Now you can begin to see how the image you see and think you're presenting can look quite different to another viewer. "Correct" processing is not about conforming to a norm but removing as much of this difference as you can, (you will never remove it all as perception is linked to experience and our experiences vary). The image you present is seen as you wish it to be seen, and not how you think it's seen.
Just for fun I'll show you the tone-mapped with added contrast and clarity version. Now I've seen plenty of landscapes, even on this site, where the photographer has applied similar editing:



Back to the white suit and one more image:



Download this and view it against ex-1_50%_saturated.jpg in the same way as we did before and you should see that the difference in colour between the two is not as marked as it was before. Compare it against ex-1_my-edit.jpg by flicking between them and you may notice that difference in colour. Which is strange because there is none, all I've done is changed the background colour. I've made the rest greyer, or in colour terms I've made it less red. Your perception of an object can not only change depending on what you view before or after but changes depending on what you contrast it against within your image. This is known as contrasting elements. See how it removes the ambiguity when flicking between it and the 50% saturated version, it no longer looks so dull and colourless and it takes your eye far less time to adjust to it. The image is less dependant on what you view before or next to as the colour has become more stable. We do this by understanding how the eye works, by providing the balance the brain automatically tries to apply. By doing this the viewer will see more of what you intend as their brains will not be correcting it so much to how they think it should look. This is the normal world, because it's what we see every day that our eye become normalised to (notice I said normal and not natural. We become accustomed to different views including photographers getting used to seeing over-processed shots, remember non-photographers don't ;) ). It's only because we can alter things so radically and easily with digital editing that we can get so far from it so quickly. And as I said you don't need to make things correct, only convincing.

Now what about the other bit, the bit about making the suit whiter because surely I can also make the tomato redder?

First a little primer on colour and exactly what saturated colour is. Two things to note first are we are not talking about colour management here or the RGB colour space, only about how we relate to colour and the terms we use. Secondly, because our application of the terms is broad the meaning of each term is precise. Although there are some peculiarities arising from trying to program saturation within the RGB colour space the meaning of the term is exactly the same.

Colour is described by three variables, hue, saturation and value (brightness).

Hue: Colour is make up of all the visible wavelengths of light. Each wavelength is a separate colour. You can see this when you split white light by a prism, exactly the same effect as you see in a rainbow. The separate wavelengths are the hues or pure colours. When you view an object all you really see is the light reflected off it. When you shine white light (light of all wavelengths) on an object it reflects light of lots of different wavelengths and absorbs maybe only a few. So when we talk of a colour's hue we are talking about the wavelength that is dominant, the one the object reflects most of.

Saturation is a measure of how dominant the hue is in a colour, the ratio of how much light there is of the hue to how much there is of the other wavelengths. A colour is said to be fully saturated when there is only the dominant hue left, the colour comprises of only one wavelength and is a pure colour or hue (or the closest you can get to it in an RGB colorspace as it is not possible to reproduce a pure colour (hue) with RGB colour). So to saturate a colour you must steadily subtract every other wavelength until you are left with the one dominant one (or hue). Every other wavelength of light (or hues) mixed in equal quantities is grey, so to saturate a colour you subtract grey. Is it beginning to sound like the white suit yet?

Value, brightness or luminance is just how light or dark a colour is.

Now let's see this subtracting grey in action. We will need ex-1_my-edit_Background_tint.jpg and ex-1_50%_saturated.jpg open in the image viewer. We will look only at the tomatoes because we've changed the colour of the background. Looking at the tomato in my edit with background tint you will see that it's surface is reflective, it reflects not just red but a lot of other wavelengths as well. See this in both the specular reflections and the way it reflects the foreground. When you flick to the saturated version see how these disappear because by increasing saturation you've subtracted these wavelengths by subtracting grey. You are left with just the dominant hue, the pure red that's the dominant hue of the tomato. So just as with the man in the white suit, by increasing saturation you're not really adding red at all but subtracting the grey and equalising the values.

Now as I said the tomatoes were backlit so we shot into the shadow forcing us to process some of the colour back in and I used 50% saturation to highlight the differences. You can of course use less saturation and here is a version at 30% saturation so you can compare it to my edit. You can also of course use a combination of all of the techniques and you will find that you'll get far more vivid colour with far less of a boost in saturation and so will retain a lot of the other information you might normally loose. So the very last image I present is just that, my background tint with a 15% saturation boost. Which version you prefer is entirely up to you, I'm just showing you the differences which I'm not sure that everybody really notices when editing:





That's enough for now. This is Part 1 and, if there's enough interest, in part 2 I'd like to show you how this works in a less controlled environment by trying my tomatoes against a red background.
the 2nd to last shot looks pretty good to me. I dont like teh blue cast youre adding back in.

the tonemapped shot made me puke. good thing i have wet ones here at my desk.
the 2nd to last shot looks pretty good to me. I dont like teh blue cast youre adding back in.

the tonemapped shot made me puke. good thing i have wet ones here at my desk.

I would entirely agree. As a complete image it wouldn't work as it's not convincing. This was more about the effects on the red and I only include it to highlight the effect it has on the red. Pt 2 is more about the complete image, not only the red tomatoes but how the processing changes the way they relate to the background and the impression of light.

I also find the tone-mapped one hideous, and I had a hard time not toning it down. But a lot of photographers shoot landscapes either into the light and shadow or after the sun is set and so end up with a file that has heavily underexposed foreground forcing them to apply very similar processing. So I include it to show the effects the processing has on colour. :)
I got it. :) Still trying to read through the entire post as well... :p

On the last two, i dont see much of a difference between the two in regards to the tomatoes, but the WB stands out to me. I always prefer a warmer image.
Because of the exact reasons you mentioned, I was told to edit to where you think it looks good, then dial it back by 10%.
Because of the exact reasons you mentioned, I was told to edit to where you think it looks good, then dial it back by 10%.

That is very correct, but I'm also trying to explore something else. Just as with the white suit you do not add white but really subtract the grey, the parts that define texture and impression of light, so it is with saturation. You do not add colour at all but subtract it with the saturation slider and with it some of the texture and impression of light. It's partly about being able to see and understand the effects and partly to introduce the idea that making something stand out is about exaggerating or maintaining the differences. The editing that's commonly referred to as "adding" such as sharpness, clarity, saturation is really subtractive as it removes information and equalises values (the red of the tomatoes becomes more uniform), making things stand out less. For instance if you want to make something appear sharp you "add" sharpness, but doing that globally actually subtracts (destroys) edge detail and though over-sharpened images may appear sharper when you flick between them, the less sharpened ones will have much finer and more delicate detail. You can actually make it appear sharper by maintaining a slight softness in the rest of the image, make the rest greyer.

Most reactions