Beta now open



Tutorial: Beginning Color Correction

9 Comments

In our never-ending quest to understand all things Photoshop, one of the most-requested tutorials has been for me to break down the basics of color correction for those of you who’ve never had the opportunity to learn any part of the world of color. A better understanding of color will take your ‘shops to a completely new level, as you’ll be much better equipped to fool the eye. As we discussed in the “Light & Shadow” tutorial, your eye is much better than your brain at spotting something that doesn’t look quite right. Balancing the color between the sources allows you to make much more believable ‘shops.

One reason I haven’t written this tutorial up until now is that this may be the most difficult thing I’ve ever attempted to teach you via the internet. Usually, when I’m teaching color correction, I’m standing right behind someone and can point right at what I’m talking about. We’re going to cover a lot of ground here just going through the basics.

So the first barrier to understanding color is something that I won’t be able to teach you online: You have to learn how to see.  What do I mean by that? You will hopefully notice, after we go through this tutorial, that black sometimes isn’t really black and white isn’t really white, and once you start noticing it, you’ll wonder how it ever got by you. Most people that I teach color correction usually develop this strange disease where they can’t quit messing with the color on their friends’ TVs. Don’t worry, it goes away by itself eventually without treatment, unlike that mole you have. You should really get that looked at.

Yes, Love Sees No Color, I’m aware.
This is the primary reason that love sucks at photoshopping.

First off, how does a computer or digital camera even store color data? Well, you may remember from kindergarten that all color is made up of three primary colors, from which you can make all other colors, and black and white, to make darker or lighter variations of those colors (To be technically accurate, black & white are themselves not colors, but the complete absence or presence of color ). If you REALLY remember fingerpainting (Dude, move on…. let it go!) You’ll remember that those colors are red, blue and yellow. That’s not what we’ll be using for web images.

The reason is this: There are actually two forms of light that make up all the images we can see in this world: Transmissive and reflective. What’s the difference?

 

Transmissive v. Reflective

Well, fingerpainting, as well as most things we see on this planet, is reflective. In reflective light, the primary colors are indeed red, blue and yellow. Pure white represents equal amounts of red, blue & yellow light waves combined. We see a red ball, and we understand it to be red, because the white light source is hitting it and absorbing the blue and yellow waves, reflecting back only the red part of the spectrum. If something is pure white, all light is reflected back. If something is pure black, all light is absorbed. This also explains why you got heat exhaustion trying to run that Ironman marathon in your Ninja costume.

But this doesn’t square with another phenomenon that most of us are old enough to remember: When we used to press our noses right up on the TV screen. Remember the first time you figured out that the TV made pictures from red, blue and GREEN dots? What’s going on here?

Well, that’s the other form of light, transmissive. As its name implies, in this form of light, no external source of light is needed – it’s transmitting its own. And when the light comes from an internal light source, it operates under different rules. So the difference between the two is that if you need light to see it, it’s reflective. If you can see it in a dark room, then it’s transmissive. In transmissive light, it’s red, blue and green. 

The graphics industry, in an attempt to try to keep the two straight, refer to the transmissive form of light as RGB (red, green, blue). The reflective form is called CMYK, and they’ve used slight variation on the names of color so that the two models don’t share any initials. CMYK means cyan (light blue), magenta (a red variation), yellow, and “k” for blacK. Black needs to be added for darker colors in reflective light. In transmissive light, we don’t need that…. we can just not transmit as much light. CMYK is a slightly more advanced color model, and not the one we will need for web images, so for the purposes of this tutorial, we’ll stick with RGB.

♪ YOU DOWN WITH RGB? ♫
♫ Yeah, you know me! ♪

The RGB Color Space

Before we even really begin, understand one thing: Color correction is like chess. I can teach you how all the pieces move in less than a half hour, but it will take a lot of experience before you’ll be really proficient at it.  The more you experiment with it, the better you’ll get.

The normal modern computer monitor operates on what is known as “24-bit color”. Those of you who played old-school Nintendo will also know the term “8-bit”, the color model that old game consoles used. What’s with these “bits”, and how do they relate to color?

A “byte” can be thought of as a box, with 8 cubbyholes in it for storing information, or “bits”., thus “8-bit” color. Each box contains the equivalent of a switch that can either be on or off. So the maximum number of colors that can be represented in such a system would be 28, or 256 colors. This possible number of colors is called “color depth”.

Boy, did someone lie to you.

Out of the estimated ten million colors that an average eye can see, 256 is an extremely small portion, and that’s why 8-bit graphics were notoriously poor by today’s standards.

In contrast, the color depth most common these days, known as “24-bit” or “TrueColor”, is capable of displaying 224 colors, or 16,777,216 possible colors. Those of you scoring along at home will notice that that number has actually surpassed the ten million colors an eye can see. That’s what makes it “True Color”, as least as far as our puny human eyes are concerned.

With 24 bits to store color information in, and three primary color channels, you should be able to see that in 24-bit color, we have 8 bits in which to store EACH color’s information (8 + 8 + 8 = 24). So the effect of that is that we can store a value of zero to 255 in each color channel to indicate how much of that color should be present (since we start at zero, we end at 255.  This throws some people. Since we’re counting zero, we actually get 256 colors ending at 255).

And that’s exactly how Photoshop stores color information, with a 0-255 value for each of the red, blue and green channels. Let’s open the color picker by clicking on our foreground color swatch.

The row on the right are approximate equivalents for the selected color in two other color models, LAB color or CMYK, which we’ve already discussed. As this is a basics tutorial, let’s ignore those for the time being and stay in RGB. On the left column, we have (in order) Hue, Saturation and Brightness. Hue represents where this color would sit on a color wheel, divided like most circles into 360°.

Cap’n! There’s Klingons approaching from the PureBlue sector!

Saturation represents how “pure” the color is. The higher the saturation, the more colorful an image is. And Brightness, obviously, is how light or dark a variation of this color you have selected.

All three of these numbers can be helpful if you’re trying to slightly alter the Hue, Saturation or Brightness of a color, but these numbers are not what truly defines the color. The color is defined by the last remaining set of boxes, R, G & B, or our old friends Red, Green and Blue. As you can see in the example, pure blue consists of no red, no green, and maximum (255) blue. Makes sense, right?

All colors are made by varying amounts of each of those three primary colors. If you become comfortable with the color wheel, you can begin to see the relationships involved. For example, looking at the wheel, blue is directly opposite yellow. So how would you make “pure” yellow? Simple. Invert the numbers we used for blue. Red 255, Green 255, Blue 0.

In practice, most people won’t type the numbers in to find a color, they’ll just use the very handy color picker as Adobe intended. But you need to be aware of the relationships between the colors or you’ll have approximately ZERO chance of ever learning how to color correct properly. Blue is the opposite of yellow. Red is the opposite of cyan. Green is the opposite of magenta.

Let’s Talk About Hex, Baby…

One last thing before we leave the color picker: The number below the RGB boxes, marked with a # sign. That’s the equivalent of all three color channel numbers in what is known as hexidecimal notation. Since computers aren’t hung up on the number 10 like we are, because they lack fingers and toes, we can store the number in a smaller area by reducing red, green and blue to 2-digit numbers, or a 6-digit code where the first two places are red, the second two are green, and the last two are blue. Hexadecimal (16-based) numbering uses the numbers 0,1,2,3,4,5,6,7,8,9,A,B,C,D,E and F. In this numbering system, A=10, B=11 and so on. Where, for example, in our normal decimal system the number “24″ really means (2 * 10) + 4, in hexadecimal the number 24 would represent (2 * 16) + 4, or 36. This allows us to store the number 255, for example, as the two-digit number FF. Hex numbering is a little out of scope for a beginner’s tutorial, so if you don’t get it, don’t worry: You don’t have to. Just be aware of what a hex number is. It customarily begins with a # sign to tell us that it’s hex, and if you want to know the hex code for an RGB value, you can always just plug your RGB values into the RGB boxes, and Photoshop will automatically tell you what the hex code is, as in the above example. Max red, max green, no blue, or FF – FF – 00, giving us #FFFF00 as a hex code for pure yellow.

The Levels Palette

Our primary tool of choice for altering color levels will be the Levels palette, Which can be located via the menu at Image->Adjustments->Levels, or much easier by hitting Ctrl-L (CMD-L on a Mac). Let’s take a look at that now:

 

At the top, we have Presets. Generally, I have little use for presets,as almost every picture you’re working on will have different needs. The factory presets like Darker Contrast are quite simple to duplicate with practice anyway, but as you begin to learn, you may find some value in applying the factory presets to see what happens to the sliders I’ve numbered above to familiarize yourself with what “Darker Contrast” actually means as far as the sliders. The one use I usually have for presets is if you have a large batch of pictures that were all taken in one sitting, under the same lighting conditions. Since they generally will need the same (or very close) adjustments, saving a preset when you get one color corrected properly allows you to pull up the same settings on the next picture as well. You can Save or Load these presets from the small flyout located just to the right of the Preset name, next to the OK button.

The next dropdown is labeled Channel. If you pull that down, you’ll see that our options are RGB, Red, Green or Blue. This means that we can select just one of the three channels to work on (Red, Green, Blue) or work on the picture as a whole (RGB).

The next box shows what looks like a black hill silhouetted against a horizon. This is called a histogram, and allows you to see at a glance the levels of color, ranging from darkest on the left to brightest on the right, according to which channel you have selected in the dropdown.

Lastly are the five sliders and their associated value boxes. You may change these values either by typing new numbers into the boxes, or by pulling the sliders closest to the boxes. In order, they are: (1) The shadow slider, (2) the midpoint slider, (3) the highlight slider, (4) The black output slider, and (5) the white output slider.

These sliders are, as of yet, unaccounted for.
And, HOW has Jerry O’Connell NOT picked up a Sliders endorsement deal somewhere?

When you move the shadow slider to the right, you will make the shadows darker. This is because you’re moving up the spectrum to a lighter gray and telling Photoshop that you would like it to “make this” the new black, and adjust the entire picture accordingly. The midpoint slider tells Photoshop to make the midrange colors darker or lighter depending on where you pull it to. With the top row of sliders, darker is always right, brighter is always left. Move the midpoint left, the lighter side becomes bigger, thus the image gets lighter. Move the midpoint right and the darker end gets bigger, making the whole image darker. The highlight slider works like the shadow slider in reverse. If you pull it to the left, you’re picking a slightly off-white color to be “the new white”, and adjusting the rest of the picture based on that.

On a perfectly exposed, entirely theoretical picture, you would find that the black slider would be exactly below the beginning of the histogram’s “hill”, and that the highlight slider would be directly below where that hill ends. In actual practice, you will never come across a perfectly balanced picture like that, as few things in nature are actually “pure black” or “pure white”, but that’s generally where things look best. You will be able to see that you’re able to pull the shadow and highlight sliders wherever you like, but be aware that if you pull the shadow or highlight slider into the area under which there are portions of the “hill” on the histogram, you will be throwing that data away. More on that in a minute.

The last two sliders, the black and white output sliders, can be thought of as the same as being able to pull the shadow slider to the left, or the highlight slider to the right, where they can’t go. They work in reverse to the shadow slider and highlight slider. By pulling the black output slider to the right, you can reset black to a lower level, effectively turning what is now black into a very dark grey. The white output slider works much the same. Pull it to the left, and you make pure white turn into an off-white, darker the farther left you go. It’s easier to see the results by looking at some examples. We’ll use a very simple and familiar image to make things easier while we get some of these concepts under our belts:

Now, since this a vector image that I drew myself, this one of the few instance where you actually WILL have a pure black and pure white to start with. That’s helpful for this example, because it clearly shows us what’s happening. In the shadow slider box, we can see that the blacks have become “blacker”. Also not that dark colors have shifted, and become harder to differentiate. That’s the nature of the beast. We are, in essence, “squishing” all the black into a tighter area, color-wise. In the highlight example, we are “squishing” the white into a tighter color area. You can see that more of the image appears whiter, and that the whites are getting indistinct and “blown out”.

In the lower row, increasing the black output makes all the grays in the image go a little lighter, although they maintain their relative brightness. This gives the image a washed-out look. On the other side, we have the white output, making the whitest areas a light grey, darkening the entire image to an underexposed or “overcast” look. In the middle, I pulled the midpoint slider to the left to 1.5, which (remember?) makes the entire image lighter. This is because by pulling the MIDPOINT slider to the left, you have made the lighter half of the image the bigger portion.

Now we get to the hard part: Seeing what’s wrong in the first place. First, let’s look at a simple color cast. A cast is when the entire image has one or more colors tinting everything. Here’s an example of some color casts:

It will ABSOLUTELY take you some time before you become proficient between spotting a green cast vs. a yellow cast, or a red cast vs. a magenta one. As I’ve said, most people need to “learn to see” them. However, once you know what’s wrong with them, fixing them is usually quite easy. If it’s a primary color like red, green or blue, you simply go to that channel and reduce the midpoint slider until it looks right to you. If it’s not a primary color, you’ll need to add its inverse. For example, we know that blue is the opposite of yellow. So for a yellow cast, you would go to the blue channel and increase the blue level by pulling the midpoint level to the left.

If all of this pulling the sliders left and right seems intimidating to you, realize that in practice, you will self-correct constantly as you learn. If you know an image has a yellow cast, you should be able to see that since we don’t have a yellow channel, you will need its opposite, the blue one. So as long as you’re on the correct channel, you should be able to work out which way to pull the slider. If you pull it the WRONG way, it will make the image even more yellow. So any time you get the exact opposite results from what you expect with the midpoint slider, just pull it in the opposite direction.

Here’s a photo of nearby Portland, with beautiful Mt. Hood in the background:

I’m fairly sure if you hang out here much that you probably have a good enough eye to recognize that SOMETHING is up with the color in the picture. What is it? Any time you’re evaluating a picture, you should look at the three corresponding areas of the picture: shadow, midpoint and highlight. The highlights look pretty close to white. The midpoints are pretty close as well, maybe a very slight magenta tinge. It’s in the shadows that we have a problem. If you look at the shadows, you will see that they have a distinct magenta tint, or “cast”, as we call it. That tells us both what slider to move and in which channel. We want to reverse the damage. Looking at our color wheel again, we see that….

…directly across from magenta would be green. We want to remove a magenta cast from the shadows, which means we want to use the black output slider slider on the green channel. So set your dropdown to green, and pull your shadow output slider up to about 15-20. Wala!

The shadows look a lot closer to black now. The midpoints, by virtue of us resetting the scale by limiting the output, are also back to straight. The snozzberries taste like snozzberries.

Again, you may become confused as you learn as to whether you need the shadow slider or the black output slider below it. And again, you will self-correct as you learn. Suppose you had worked out that we needed the green channel due to the magenta in the shadows (magenta being the opposite of green). You know we need a shadow-end slider, because that’s the problem area. If you had tried to correct the problem using the shadow slider, the shadows would have become even more magenta. Just like pulling the midpoint slider the other way, when you pull a shadow-end or highlight-end slider and it does the exact opposite of what your expect, use the one above or below it.

So we just guess until we get good at it?

Not quite. There are a few methods you can use while you’re learning. First, and probably most useful as you learn, is to get familiar with how different cast problems look. Take a picture that the color looks right to you. Open up the levels dialogue and start playing with sliders. Move the blue midpoint to the right, and see what a yellow cast looks like for yourself. The more exposure you have to all the different ways color can go haywire, the better you’ll be at instinctively knowing what is wrong and how to fix it.

There’s also a few handy tools that you can use to tell you what’s going on. Chief among these is the Color Sampler tool.

This bad boy here, located under the Eyedropper flyout.

When you click on an image with the color sampler tool, two things should happen. One, it will put a numbered marker where you clicked at, and two, it should open the Info window for you (I’ve seen instances where the Info box didn’t open up. If that should happen, you can open it manually under Windows–>Info). You can have a maximum of four samplers at any given time, although you can move them around if need be by simply holding your cursor over one (you’ll notice the cursor change) and then clicking and dragging it to a new location. To eliminate a sampler entirely, click and drag it until it’s completely out of the image area.

As you can see from this example, the info box will now display the color values for whatever points you have sampled. In order, Sample #1 is on a pure black area, and it displays 0-0-0, the RGB value for black. Sample #2 displays 248-248-248, or nearly pure white. The third sample, in the red part of Rocky’s eye, displays 206-0-4, or about 80% red, no green, and just a drop of blue. And finally, the last sample, the light grey on Rocky’s eyebrow, shows 220-220-220. Besides illustrating how the color sampler works, it should also show you something about black, white, and shades of grey. Remember when I mentioned that black and white aren’t technically colors, but the complete presence or absence of color? Any time all three channels match perfectly, there is no overriding color present. True black is always 0-0-0. True white is always 255-255-255. Any other combo like 100-100-100 or 72-72-72 will always be a shade of grey. As a matter of fact, you generally have to get about 10 points away from pure balance before most people can see a difference, so colors like 140-135-145 are “close enough” that most people would consider them grey.

So how does that help us? It can tell us where the isue might be if we’re unable to see it. Let’s look at that uncorrected picture of Mt. Hood again, only this time I’ve laid samplers on what I think should be white, black, and a medium-neutral shade of grey:

Looking at the numbers, we can see that my off-the-cuff analysis was pretty close. Sampler #1, up on the top of Mt. Hood, shows 223-221-222, or pretty close to an even off-white (Again, you will rarely find pure black or pure white in nature). Sampler two shows 121-113-123. That’s a slight deficit in the green channel. And finally, sampler three shows 15-0-17, which is a pretty large lack of green. Why is 15-17 points in the shadows “pretty large”, but 8-10 points in the midtones is a “slight deficit”? Think in proportions. The midtones are missing 8-10 points out of about 122, or around 7%. The shadows are missing all green out of 15-17, or 100%.

So how did we fix the midtones by simply adjusting the black output slider? We changed the scale. Any time you adjust from one of the end sliders, you will naturally affect the middle of that scale. Say for example, on a scale from 1 to 10 in attractiveness, I rate a 5, exactly average. Yeah, I know I’m uglier than that, but play along here. If we change the scale to be from 5 to 10, suddenly I’m a 7.5. Conversely, if we change the scale to 1 to 5, I rate a 2.5 (probably a lot closer to reality). For this reason, it is vitally important that you try to get your shadows and highlights in line before you try to mess with your midpoint. Otherwise, you’ll be chasing a moving target.

Your image may not have a good neutral grey. Don’t use it then. If you have a pure red, green, or blue in your image, use that instead. Like in the image of Rocky above, his eye was mostly red, no green, just a drop of blue. Had we seen a large amount of either green or blue, we could have used that to determine if there were an overall color cast that way. Any color will work, as long as you know what to expect… if you were to give me a photograph of a piece of abstract art that I’d never seen before, with no reference whatsoever, I would be unable to color correct it properly. Fortunately, most images will have black, white, skin tones… something we can all agree on what ballpark it SHOULD be in.

Color Matching

So how would you go about, for example, matching two different skin tones for a Photoshop? Let’s say you’re going to use this head on this body:

Because why WOULDN’T you want to?

If you just use both sources as they stand, you end up with this:

Okay, I’ve seen worse… but you’re not fooling anyone with this shop. It would look a lot better had you matched the skin tones. Let’s drop a couple of color samplers on each of these images. By the way, you CAN put color samplers on two different images, I have them in one file for my convenience while explaining this.

The first thing I’d like you to notice is that even on a single face, skin tone varies wildly. We’re not Barbies, as it turns out. Check out the differences in values between #2, the highlight on Woody’s nose, versus the values on #3, located on a midpoint in his cheek. So try to find comparable skin/body parts that are lit roughly the same to sample. Comparing the highlight on one face to the shadow by the chin on another face is going to be of absolutely no value to you.

Looking at Woody’s numbers on his nose highlight (#2) vs. John Cena’s (#1), we see some differences right away. I’m hoping that now that you’re really looking at color, you can see just by looking at it that Woody’s head is too blue. But let’s check out the numbers. The red values are pretty close, there’s a tiny bit more green on Woody, but the number that’s REALLY out of whack is the blue number. As we have a black identified (#4), we see that the black’s pretty close to neutral, and the midpoint is slightly off, but not by as much as in the highlight. So since the problem seems concentrated in the highlights, and it’s too blue, we need to use a highlight-end slider on the blue channel.

If you arrange your windows so that you can see them at the same time, you can actually watch the adjustments you make “live”. Open up your levels dialogue, and look at your info panel again. Now, if you have samplers on two different layers, as I do here, the samplers for other layers will disappear when the levels dialogue is open. This is of course because you can’t affect the color on the other layer at the same time. So it helps sometimes to write down the RGB value you’re trying to match if it’s on another layer. The numbers on John Cena we were going for are 243-201-179.

Notice that in the info panel, the RGB values now have a slash, then a second set of numbers. The numbers to the left of the slash represent the original values, the numbers to the right of the slash are the numbers as they would be if you applied your current levels correction. So let’s pick the blue channel, and adjust a highlight-end slider. The highlight slider will dump more blue in the highlights if we move it left, so that’s not the one we want. We want the white output slider below it, which will limit the amount of blue in the highlights. As we pull it to the left, the numbers will begin to change…

…and around 225 the numbers for sample #2 (Woody’s highlight, remember?) get close enough for rock and roll. If you just look at the images, seems like that’ll work. You’re only going for “close”. You’ll drive yourselves nuts trying to match colors EXACTLY, and besides, as we already covered, skin tones vary wildly from one area of the face to another anyway. We’re just going for the general ballpark. So let’s go ahead and apply this correction. By the way, if you should make corrections you don’t like, and would like to reset to how the levels were when you first opened the levels dialogue, you could cancel and re-open the levels, or you can simply hold down your ALT key and the CANCEL button will magically turn into a RESET button. Let’s take a look at what we ended up with:

Now THAT will fly. Let’s put them side-by-side so we can compare the realism.

I hope it’s pretty obvious to you how much better the Woody on the right looks.

Auto Correction & Point Samplers

In the interest of completeness, I’m also going to run very quickly through two sets of tools that I’m actually not a very big fan of. The first is on the levels dialogue, and it’s the button marked AUTO. As you might guess, clicking that button will make Photoshop try to automatically color correct the layer for you. When it works, it’s great, but it only works about half the time. Let’s look at this picture of Anne Hathaway:

Because why WOULDN’T you want to?

Obviously, there’s too much magenta in this picture, specifically in the highlights. So let’s look at the green channel:

See the histogram, the “hilly” representation of color? See how it ends far before the white end of the graph? That’s an indication that there is a deficit of green in the highlights, which is exactly what we predicted. What the auto button does is look at the histogram and try to set the shadow and highlight sliders in each channel to wherever it finds the beginning and ending of the hill. Let’s try it now. Click the AUTO button.

Okay, not bad. I could nitpick it a little, the highlights look a little green/yellow for my tastes… but for most purposes, it would be close enough. But this is a professional shot, lit pretty evenly and shot with a decent camera. Let’s look at a shot that isn’t so pretty:

…And the photo is worse, too.

Glare, a magenta-blue cast, a little too contrasty… Let’s try the AUTO function out on this one.

Yes, I swear to God those are different pictures. Why did it work for Anne but nor for Gaga? The AUTO button will only help you if you’re starting with a decent picture. If the channels were screwed to begin with, either through weird lighting, poor camera quality or bad exposure, all it’s going to do is reset it to the crap that was originally shot. So like I said, when it works, it’s great. When it doesn’t, it’s useless. Your mileage may vary.

The other set of tools that I don’t recommend are the point samplers, located right below the Auto and Options buttons. There are three, in order from left to right, shadow, midpoint and highlight. The idea behind these tools is that you pick a new point to be either the new black, midpoint or white by using the appropriate dropper. The huge flaw in these tools is, as I’ve said before, you will rarely come across a picture that has pure black or pure white. Finding a perfectly balanced midpoint is also pretty challenging until you have a very firm grasp on these concepts.

The real danger with using these tools is this: You pick a new black. But let’s say that the color you pick to be the new black is actually in the ballpark of 50-50-50 instead of 0-0-0. By picking a dark grey to be the new black, you’re telling Photoshop to simply throw away any values it finds that are BELOW 50-50-50. Same with white, if you pick 200-200-200 for your new white, it will throw away any values above that. The result of that is that you’ll lose all distinction between shades of color in those ranges. This makes your shadows indistinct and your whites “blow out”. Look at the example below:

Again, nothing wrong with the tool, it’s doing what it’s supposed to. It’s in how you use it that the danger lies.

Hue/Saturation Manipulation & Affecting Individual Colors

The last basic way to affect color that I’d like to discuss in this article is the Hue/Saturation dialogue. This dialogue can be found by hitting Ctrl-U (CMD-U on a Mac), or under the menu at Image–>Adjustments–>Hue/Saturation. There are also adjustment layers that provide a Hue/Sat correction, but let’s just use it directly on the layer like we did the levels dialogue just for the purposes of this training.

The top slider, Hue, will shift the entire image’s color map as you pull it left or right. This can be useful if you’re trying to “nudge” the entire image a little one way or the other. Generally speaking, it will go greener as you go right and redder as you go left, from its original setting. If you pull it far enough, you’ll eventually invert all the colors in the image (Colors only. It doesn’t affect black or white, so you’ll never end up with an image that looks like a negative).

The next slider controls saturation. Left is less colorful, right is more colorful. This is the easiest way to “pump up” color if an image appears washed out, and pulling the saturation down to zero is also a very easy way to turn an image black & white.

The last slider is lightness, and it does exactly what you’d think it does, make the entire image lighter or darker. I find this slider pretty useless, as you can use the levels adjustments we already covered to make things lighter and darker where you actually want them to be, not across-the-board.

The real power of this dialogue lies in an ability we just can’t get out of levels: Hue/Sat can be used to adjust just ONE color in an image, instead of whatever colors fall in the same range of dark/light. This can be used to recolor just a portion of the image, or to fix flaws found in only one area of the image.

Continuing on with our picture of Anne (because why wouldn’t you want to?), let’s say you want to change the color of her gown. Open the Hue/Sat dialogue and change the dropdown from Master to Reds. If you looked at the other options as you pulled the dropdown, you’ll see that besides Master, you can also pick any of the three primary colors, as well as their opposites, the three secondary colors. Now look down at the bottom of this dialogue, you’ll see a color bar with some new sliders down there.

Those sliders show you what sections of the spectrum that Photoshop plans to use for your adjustment, and where they default to will depend on what you pulled from the dropdown. However, you can reset those to wherever you like. The triangle points to the outside represent where the selected colors will gradually begin and end, ramping up and down from either end towards the middle. The portions between the rectangle marks represent the colors that will be COMPLETELY included. So right now, you can see that, as our default stated, we basically have the reds selected. If you move the Hue or Sat sliders now, what you’re changing will ONLY affect the reds. Let’s do that now. Pull the Hue slider over to around -45.

Using a red dress illustrates one problem with this method that you should keep in mind: Notice how it shifted ALL the red in the image over to the magenta side… her lips, even the blush in her cheeks. So if you don’t want to affect those areas, you may want to use a selection or a mask to confine your effect to just the area you would like included.

This process is VERY handy when dealing with some common problems that occur on the photographer’s end. My former business, school & school sports photography, has one problem area that occurs over & over, and it seems most labs either lack the knowledge to fix it or they just plain don’t care. When you photograph teams outside, in bright light, even grass that is actually green comes out day-glo yellow sometimes.

This looks sick to me. The largest school photo lab in the US, Lifetouch, puts out images like this regularly. And it’s such an easy fix.

Open the Hue/Sat dialogue. Now, you might think we’re headed for the green channel, fixing grass and all… but remember, that’s what we WANT the grass to be, not what it actually is. The problem area is actually closer to yellow, and that’s the channel we’ll work on. Select yellow from the dropdown. As you pull the Hue slider to the right, you’ll see the grass begin to become green again. Because it’s, as I said, “day-glo”, I usually like to pull both the Saturation and the Lightness down as well. For this image, I settled on Hue +20, Sat -19 and Lightness -14. Let’s look at what that gets us:

Amazing. The grass is actually green. Who’d have thunk it?

Ya get all that?

I know I’ve thrown a ton of information at you here, and yes, these are only the basics of color correction. Getting really good at it requires practice, and I’ve never even tried to teach this sort of thing without being in the same room before. Color correction training is usually a very interactive process. So if you have any questions, or you have specific color issues that you’d like me to go over with you, feel free to bring it up in the forums. I’ll be more than happy to assist you with level corrections, or let you borrow my eyes to see what’s wrong with your image in the first place, until you’re able to start spotting color issues on your own.


 
Tags: , , , , , , ,


5 Comments


  • #1
    HexxxySexxxy
    10/11/2012 6:17 pm

    That’s some good stuff there. I’m an amateur so I almost never mess with color, but I’m definitely going to spend more time on it now.


  • #2
    maluba
    10/11/2012 10:05 pm

    mad…. FANTASTIC tutorial. there was plenty of info here that was helpful and i kept finding myself copying your sample images, pasting them into PS, and trying what you said to make sure i had a grasp on the techniques you were describing. bravo.

    madmann replied:

    That’s why I put ‘em there. :)


  • #3
    BlondeFury
    10/12/2012 11:15 am

    This was excellent!…you amaze me…

    One thing though…can I still use Tide to make my whites even whiter?

    madmann replied:

    YOU can. No one else, but I’ll make an exception for you.

    aerodynamite replied:

    I’ve found that dancing can make whites even whiter…


  • #4
    AuntieMeme
    10/13/2012 12:22 pm

    This is outstanding. It’s really cool to have something I was doing through trial and error explained so elegantly and mathematically.

    mightyzamfir replied:

    Same here. I knew what the stuff DID, but I didn’t know how the stuff WORKED.


  • #5
    aerodynamite
    10/25/2012 12:07 pm

    This was a fantastically thorough tutorial, Mad. Honestly, I usually use a ham-fisted approach to color correction, but now I think I might actually delve into the levels. But I’ll probably tie a rope around my waist so I don’t get lost…


Leave a comment

You must be logged in to post a comment.