Transcripts
1. The Histogram Demystified in Photoshop - Introduction: Hello, I'm Helen Bradley. Welcome to this Graphic Design for Lunch class, demystifying the histogram in Adobe Photoshop. Now I know that this is a point of confusion for a lot of people. So what I'm going to do is answer some of the questions that you probably already have about the histogram, such as what is it anyway? Should I even care about it? If you can't do math and that looks awfully like math, then is it really for you? Well, we're going to look and see what the image histogram even is. We're going to answer the question of what sort even telling us about our image, we're going to see if there's a shape that it should be. We're going to see how you can use it to adjust your images. At the end of this class, you're going to have a really good appreciation of what the image histogram is, what it's telling you about your image, and how you can use it to make better photos. Now as you're watching these videos, you're going to see a prompt which lets you recommend this class to others. Please, if you're enjoying the class, do two things for me. Firstly, give it a thumbs up and secondly, write in just a few words why you enjoyed this class. These recommendations help other students to say that this is a class that they too might enjoy and learn from. Now if you'd like to leave me a comment or a question, please do so. I'll read and respond to all of your comments and questions, and I look at and respond to all of your class projects. If you're ready now let's get started demystifying the histogram in Photoshop.
2. Pt 1 What is the Histogram?: For us to be able to look at the histogram in Photoshop, we're going to need two things. Firstly, we need a photograph, which I've downloaded here from unsplash.com, and we're also going to need to see the histogram itself, to see the chart that is showing us information about the photo. Now, as you're working with your photos, you will see a histogram in a couple of places. If you choose "Layer", "New Adjustment Layer", and for example, choose "Curves", and I'll just click "Okay", you will see the histogram. A histogram is a chart and we're seeing it here in the curves dialogue. It's here because it's useful to us when we're adjusting our photo. Now, in some versions of Photoshop, earlier versions of Photoshop, the histogram did not appear in the Curves window. If you're working on those earlier versions of Photoshop, you can choose "Layer", and then "New Adjustment Layer", and choose "Levels", and I'll click "Okay". The Levels adjustment also has this histogram, has this chart in it. There are two times when you will see the histogram when you're adjusting your image, but you can also see it if you ask to see it. I'm going to choose "Window" and then "Histogram", and this is the default histogram that you're going to see in Photoshop. Now, this one is a color histogram, but because we want to match it up with the curves and levels histogram, I'm going to click this dropdown list and choose "RGB", and this is the same histogram as we saw in those other dialogues. Here, I have a little warning sign, and that's just telling me that Photoshop's using what it's calling a cached version of this histogram. It had a brief look at the image and it's drawn out a histogram. But it's saying to us, "If you want a better version, just click this icon and I'll go and give you a proper version," so this is that proper version of the histogram. Now, the histogram is a chart, it is mathematics, but it's pretty simple math. What it's showing us is what we call the tonal range in the image. The spread of pixels from very dark to very light. This particular histogram is ignoring color. What Photoshop's done is it's gone and checked every single pixel in this image, and it's determined where it is on a range of 256 possible tones from black over here to white over here, and so it's had a look and see if it's got any black pixels, and it's counted up how many of those that has, and it's plotted those on the chart. Then it looked at the next two black pixels, and then the next two black pixels, and it's charted them for all the way from black to pure white. Then the middle is mid gray. What we're seeing is the spread of the tones in this image. We can see from this histogram that there is a preponderance of light tones in this image because this is big bump in the chart. There are more pixels of light tones than there are dark tones in the image. Looking at the image, we can say that, can you see any blacks in this image? There's possibly a little bit down here, and maybe some in here, but really, it's all pretty light and we can even say from the histogram, it's all lighter than mid gray. Let's have a look at another image. Here we have a relatively similar histogram. Again, everything is pushed towards the lights end of the chart, and that's true to the image. You can see there's only very small areas of dark in this image, and most of the image is taken up with these clouds. There are a preponderance of light pixels in the image, not so many dark ones. All of these images I've downloaded from unsplash.com, just so that we can use it to look at the histogram. Now, here we have a lot of color in the image, but let's have a look at the tonal range. Well, here we've got a lot of very, very light pixels that's probably here in the height of this bag. There are also a lot of pixels around the mid tones. These are the mid tones in the image, mid gray if we turn this into a black and white image. They're probably here in the chair, in the hat, in the lighter parts of the leaves, and in this blanket thing here. There are also a few blacks in the image, and that's probably in the chair as well, and perhaps part of the leaves. We've got a few blacks, a lot of mid tones, and a little bit lighter, and a lot of really light pixels in this image. Here we've got another one of those images which has a preponderance of light pixels, and that's the sky here. But you can see also that we've got three little bumps of pixels here, so we've got a little bump at the mid tones, a little bit lower than the mid tones, and a little bit darker than that. But we have no really black pixels here, very, very few black pixels. Here you might at first think that this image is really black, but it's not really black in here, it's really a dark gray. My eyes are telling me that, but so too is the histogram. You see there's nothing here at the very end of the chart, there are no blacks. In fact, the first blacks in this image are really just dark gray, and there also no whites in this image. You can see that the chart does not go all the way up to the wall here. It's well back from the end of the wall. Most of the detail in this image is in the dark grays to gray white. No whites, no blacks. This image has a lot of black. It's got a lot of dark pixels, but also black, so you can see that the chart has run into the wall here. These are pure black pixels, what we would define this as is detail lost in the shadows. These are so many black pixels that there are some details, some texture in this image that has been lost because it's all gone to black. For the rest of the image, pretty nicely expose. We've got obviously a lot of dark pixels because it's been shot at probably sunset. We have a few light pixels in the sky, but you can see even though the sky looks pretty light, we're not running up against the wall of this chart, and we don't have a really big bump here. The preponderance of pixels are dark pixels in this particular image. Here's another image with a very interesting histogram. Again, there's nothing in the black. There's no detail in probably about the first one-fifth of this histogram, and we're seeing it in the image. The image has this cloudy look. It doesn't have any pure black pixels. Everything else is pretty spread out across the tonal range. We've got a few dark pixels, while we're seeing that in the image here. They're dark, but they're not black, and everything else is pretty evenly spread across the tonal range. Now, as we're looking at each of these images, you will see that every single one of them has a unique histogram, and sometimes they have multiple bumps and sometimes they have one bump only. It doesn't really matter what shape the histogram is. What we're doing is learning what the histogram's telling us about the image, and then we can make choices about how we fix the image and the histogram is going to help us do that. Now, in this image, we do have what might be an inverted commas called problems. There are two of them. Firstly, the whites are jammed right up against the edge of this chart. That tells me that there are some blown out pixels, some pure white pixels in this image. They're blown out, we can't get detail out of them. At the other end of the chart, the same things happen. There are some black pixels, and that would be detail lost in the shadows. We've got blown out highlights and detail lost in the shadows. This is not an ideal histogram, but this is the histogram for this particular image. Here in this image, you might think that there would be blacks. There are certainly a lot of really dark action happening in the top part of this image, but in fact, there are practically no blacks in it at all. There is a slight problem on the top end of the histogram, the whites area because the detail in the chart has gone right up against the wall, so there are white pixels in this image. They might look gray, but they're in actual fact pure white pixels. It's not ideal to have pure white pixels in an image because that is blown out highlights. Although for this image, it works just fine. But if you were to see this on the back of your camera and these histograms that you're seeing in Photoshop are exactly the same ones you see on the back of your camera. If you were seeing that on the back of your camera, you might want to make some adjustments to your camera so that you didn't get these blown out highlights so that you had some detail in the highlights in the image. This landscape is really quite nicely exposed, although it does lack again some black pixels, you can see that the chart doesn't come all the way to the black end. We don't want it to run up against the wall, but we potentially may want some black pixels in this image. Again, even though there's a lot of snow and there's some white clouds in the sky, we don't have many pure white pixels that we haven't lost detail a lot. We don't have blown out highlights in this image. It's pretty well exposed if lacking in blacks. Now, the next two images are mine, and I chose them because they are significantly overexposed. Although the chart itself looks all right, it looks like I don't have any totally blown out white pixel or very few of them. The detail in the sand has been completely lost. So this is an extremely overexposed image. Most of the image content is in the really light area of the histogram, and that's telling us that it is potentially overexposed, and this one is. I've one more, again, overexposed. Again, the detail in the chart doesn't quite reach the wall, is not banging up against the wall. But we can just see from this image that there's detail loss. The sand has no detail, and all of these white buildings have no detail at all. There should be some detail on the roof. We can see that this is corrugated iron, but we're not seeing any of the bends in the corrugated iron that we would expect to see. So again, this is a highly overexposed image. You've seen that the image histogram can tell you a lot about an image. Looking at the image itself and looking at the histogram can tell you a lot. In this case, this is severely overexposed. But if we have a look at this histogram and have a look at this histogram here, in fact, it works for this image. The histogram is not the similar here, but this is a much better image. We need a combination of our eyes and the histogram to be able to tell us whether we've got everything right. But certainly, the histogram can tell us a lot about our image.
3. Pt 2 How Photographers can use the Histogram: Now, that we've had a look at and we're able to read an image histogram, let's talk about how that would affect us as photographers. If you're shooting these photographs on your saying histograms like this on the back of your camera, because this histogram in Photoshop is exactly the same histogram as you would see on the back of your camera, what can you do to fix it? Well, in this situation, the biggest concern is this blown out highlights. You can generally recover a little bit of data from shadows, but recovering blown out highlights can be very difficult. This is also a long exposure. You can see that this person is moving, so we're getting a sort of long exposure effect here. What you might do is just speed up the shot just a little bit. If you speed up the shot little bit, you would get less light into the camera and you would not get blown out highlights. That is one option as you're shooting to adjust for a histogram that doesn't look good on the back of your camera. With this image, the problem is we don't have any blacks. If you had a look at this image and if you said, I really want this to be black instead of sort of gray, if I think that this is overexposed, then what you would do is to close down the aperture on your camera. This one's been shot with a pretty wide open aperture. You can see that the background is out of focus, so we've got a shallow depth of field, which means we're probably shooting at F4 and F5, something in that sort of region. Well, if we close down the aperture a little bit or shot this a little bit faster, would get life light into the camera, and that would move the histogram across allowing us perhaps some darker pixels in the image. With this shot, you're always going to have difficulty because you've got good exposure across the whole of the shot and you wouldn't want to blow out the sky. There are a lot of dark pixels in the image, but shooting this, you would be pretty proud of getting a histogram like this on a shot like this where you have a lot of dark areas but you have potentially a really light sky. If you saw that on the back of your camera and probably go ahead and shoot it, I don't think you're going to do a whole lot better than this. With this image it's got problems at both ends of the histogram. There are no light or white pixels here, there are no black ones. The rest of the image is pretty well exposed. If you were going to speed up this shot so that you get some dark pixels, you really going to move the entire histogram all the way across. You'd probably accept a histogram like this on the back of your camera knowing that you could make adjustments to this image later on in Photoshop or Light-room or any photo editing application that gave you the ability to alter the tonal range in an image.
4. Pt 3 Correcting with Levels: Now while the photographers for these images have probably edited these images today, what they see as the creative possibility for their shot. Let's have a look at a few of these images, and if we were, [inaudible] and if they were not happy with the tonal range in the image, let's see how we would fix that. I'm going to start with levels because everybody has a levels adjustment with a histogram. I'm going to choose layer, new adjustment layer, and then levels. Levels is a nice simple histogram adjustment, and that's a good place to start so over here we've got our black pixels, and a black slider. Here's our mid- tone slider, and here's our white slider so for this image, we have already identified that it lacks blacks. There are no really dark tones so we can fix that by just dragging in here on this slide, and that's going to add some black tones to the image. Here's the before and here's the after, you can see that they fix has added some more dark pixels into the image if you wanted to, you could even push it a little bit further. You're going to lose detail here in these particular tones. They're going to go to black but again, that's giving you a much crisper color image. Now for the mid-tone slider, we may not want to adjust for this image, but let's say what it would do for us. Counter-intuitively, if we pull it to the right, we're going to darken the image. You might think that would lighten up, but it actually darkens it. We go the other way we're going to lighten it so you may already be seeing some creative possibilities in this levels adjustment that are working counter to improving the tonal range in the image. You can say that if you drag the mid- tones sliders to the left, that you're getting this ethereal foggy look to an image. It may be as simple as coming into a photograph and adjusting our mid-tones to get this foggy look into an image. But for this image, I think we can probably put this slider back to its pretty much default position. Let's have a look at another image again, another one of these slightly foggy images. Let's choose layer, new adjustment layer, and we'll do levels on this one as well. Again, we've got the histogram, again, we've got no blacks in the image this time we've got no whites either. Let's drag in on the white slider, and typically what we'll do is we'll bring the white slider to just the position where the data in the chart starts to tape out. We're just going to make that a new white point. What we're doing is we're saying that whatever color this was, it's now going to be white. We're going to stretch the tones in this image, let's go back here and let's adjust the black. We're saying whatever tone this was and it was a dark gray we're saying to photoshop call that black now, and here's the before and after on this image. Now, I'm going to go back here to the layers palette because what I want to look at now that we've applied this levels adjustment to the image, is I wanted to go and see what the histogram now looks like. Let's choose window and histogram so you can see that the histogram has this telling lines through it that tells us that it's been stretched and it's been stretched by photoshop. That's a typically and if you like fixed histogram but you can see now that the tonal range in this image goes all the way from black to white. We've stretched the tonal range, It used to be from here to here, and now it's all the way across and the image is looking a lot blacker and a lot whiter. There's a lot more contrast in this image, as a result, this is the low contrast, everything in the mid-tones version of the image, and this is the much higher contrast, more crisp image. This is an aesthetic consideration, it's not a hard and fast rule. If you like images that look like this, then your histogram is just perfect because you like this image. If you prefer things to be more contrast, to be a little bit crunchier, a little bit sharper, then this is what your histogram would want to look like, and this is a nice fix for this image.
5. Pt 4 Correcting with Curves: Now in the previous video we had a look at a levels adjustment, let's talk this levels adjustment and let's go and say how curves is a little bit different. I'm going to choose layer, new adjustment layer and then curves. Now with the curves adjustment, I still get the histogram, particularly in these later versions of Photoshop, so I can still make my adjustments. I don't have a mid tone gray slider here, but I do have a black and white. I can bring these sliders in and I'm going to get the same effective histogram result if you can say that it's stretched a little bit different in shape because this adjustment is going to be marginally different to the one I did with the levels. But the curves dialogue has some other fixes up at slave. At the moment we've got a straight line curve and what happened because it used to look like this line had even stateness to it. Let's see how much stable it is, now that we have adjusted the slide as at the bottom. This line is a little steeper, and when it's a little steeper, the entire image is a little bit more contrast. What we're doing is we're adding some contrasts across these tones in the image. But the curved line is just that. It's a curve so we can bend it. I'm just going to pull up here on the top part of the curve line and I'm going to pull down here, just click and drag on the bottom part of the curved line. Let's see what we've done. Well, if this area of the chart is the lights, then we have just lightened the lights, so we get the ability and the curve dialogue to impact the very lightest pixels in the image a little bit more than we could in levels. We have lightened the lights by dragging the curve above that line. Down here, I drag the curved line below the straight line. I've curved it down to make a s-shape in it, because I pulled down on the darkest end of the curve, I'm darkening the darkest pixels. Now, that's been a very enthusiastic fix, you wouldn't want to fix it that much, but it does give you the idea that the curve can be manipulated. You can pull it up and you can push it down to change the way that the histogram is going to look. You can craft an image in terms of the tonal range in the image using this curves dialogue. Now, in earlier versions of Photoshop, even if you didn't have the histogram here, you can still use the curve dialog and it's still going to work exactly the same. You just don't have the benefit of knowing where the tones are in your image. Let's go back to another one of these images that we've fixed and let's have a look at the curves on this now. Since got a whole lot of things hanging around on it, let's just get rid of them. Let's go into a curves adjustment layer, new adjustment layer, curves. Here we're pretty good in the light and white area of the image. We could maybe drag just in a little bit, but we're lacking blacks so we're going to drag inwards on the slider to just pick up the very end of the chart. This is what a histogram now looks like. Well, let's go and do the same thing, let's lighten the lightest pixels just a little bit, dragging up on this end of our curve and let's darken the darker pixels. This time, let's throw quite a bit of darkening into it. You can see that we're able to craft this image to be our version of this image by just adjusting these curves. Now, this curve point can be taken to any way you like. We could actually use it a little bit more in the darks in the image, not so much in these areas, so you say that moving this around can have a different effect on the image. But this shallow S curve where you drag upwards on the lighter area of the image and down on the darker tones is a classic fixed for an image. It gives the image a lot more contrast and you can see the contrast enhancement is also boosted the color in the image as a result. A lot of images will respond really well to a shallow S curve adjustment like this. Before we leave this curves, let's go and have a look at that landscape as well. With this landscape, we didn't have a lot of blacks, and it will probably respond pretty well to a curves adjustment. Let's try that, new adjustment layer, curves. Let's drag in to say if we can get some blacks in the image. Well, we don't want to over darken it, so I'm thinking probably not coming in all away might benefit the image. We could pull in a little bit on the whites as well. Now, when you are adjusting the curve, you don't have to use the S curve. You could drag down to darken the overall tones in the image, or you could drag up to lighten them. If we drag up, we might be able to aggregate some black pixels in the image, but we're dragging up, so we're lightning everything else just a bit. Let's have a look at the before and after on this one. This is the before and this is the after. It's got a different look, the colors are a little bit richer, particularly in the grass, and we haven't got that quite so foggy look in the image. You can see that we've crisp it up a lot because we bought blacks into the image. We've expanded the tonal range in this image significantly. Now it goes all the way through the blacks and all the way to the white. You can see here that the original image did not and the result is quite significantly different. As I said, the histogram that you are saying in Photoshop when you open any photo in Photoshop is exactly the same histogram as you would have seen on the back of your camera as you take your shot. The histogram is a really awesome tool for helping you improve your photos. By learning to raid the histogram, you can make adjustments to your camera so that you can shoot better photos. When you're working inside a program like Lightroom, Adobe Camera, Raw Photoshop, anything that has a tone curve adjustment. You can now read the histogram, understand what it's telling you about your photo and use the details in the histogram to improve your image using the tone curve or using levels so that you get a better result.
6. Project and wrapup: Your project for this class will be to take a photo. Go and download something from unsplash.com. I'm going to give you a link to some of the images that I've used in this video, particularly the ones that we adjusted using levels and curves. Go and download a photo or use a photo of your own. Have a look at the image histogram. Ask it what it's telling you about your photo, and then go and make adjustments to your photo using either curves or levels. Post an image of your fixed image as your class project. I hope that you've learned things about photograph, photography, histograms, and Photoshop that you didn't know before you started this class. As you were watching this class, you will have seen a prompt which asked if you would recommend this class to others. Please, if you enjoyed the class and learned something from it, do two things for me. Firstly, answer yes to recommending this class to others, and secondly, write just a few words about why you enjoyed the class. These recommendations help other students to see that this is a class that they too might enjoy and learn from. If you'd like to leave me a comment or a question, please do so. I read and respond to all of your comments and questions, and I look at and respond to all of your class projects. My name is Helen Bradley. Thank you so much for joining me for this episode of graphic design for lunch, and I look forward to seeing you in an upcoming episode soon.
Helen Bradley, Graphic Design for Lunch™
