I can honestly say I invented this. Now I’m not saying there hasn’t been anyone else doing the same thing, but unlike some of my other tutorials where I’ve simply clarified or amplified what others have taught me, this I worked out on my own. As a result, the process has gone through a few iterations before I felt like it was working just right for me. So now I’m ready to share it with you! At the end you are quite welcome to ask – even in the comments section, if you’d like – “I don’t get it, what’s the big whup?”

[click "continue reading" below for more…]

The original concept was triggered by someone else’s comment of course. I attended a workshop at my local PPA (Channel Islands Professional Photographers Association, to be specific) about a year ago. Someone from Fujifilm was talking about getting exposure in camera just right in portraiture. He suggested zooming the lens to fill the viewfinder with your subject’s face, and then set your exposure there while zooming back out. His philosophy was that the face should contain the full range of contrast (i.e. light and dark), and the rest of the image could go to heck in a hand basket as long as you had that right. Or roughly that, anyway.

But I got to thinking: this could also be a good Photoshop technique, so that faces pop and don’t look flat and improperly exposed. You want the emphasis to be on your subject’s face, so you want the face to have the most local contrast, (the most range of dark to light). The viewer’s eye will naturally gravitate there if you do it right.

But how to do it right? My first attempts looked a bit weird, and I was always dialing the amount of the processing back, or viewing it later and cringing at what I’d done. But I’ve slowly refined the technique, and here’s how I do it. What’s more, I’ve been able to speed up the process AND make it look better, by trying different things (and doing it hundreds of times!).

So the lead image of this post is the final result. Certainly not a fully-retouched image, as there are a few things I didn’t tweak that I normally would. Like that big faint pink spot on the right side: that’s flare coming back from my shoot-through umbrella, just out of frame. Hopefully that’s the first thing you noticed. Did you notice the perfect facial contrast on Christina? No, you didn’t…it just looks good and natural.

Before I get to the instructions, here are a couple of shots I’ve shown recently, with comparisons. The one on the left is the finished version, with face contrast and other retouching applied. The image on the right is the one right out of Lightroom. The example I’ve used to demonstrate the technique might be a little subtle, unless you also see the difference in other images.

Note that all the following descriptions refer to the image below that description.

Below is the original image, straight out of Lightroom. No local adjustments done, just basic tone/contrast/color adjustments. I picked a decently-lit image to begin with, so the adjustments are going to be subtle. And really, that’s how they should be. An image looks really bad if you can tell something’s been done to the face contrast.

One hint for judging if you’ve gone overboard with the face contrast adjustments: view your finished image in thumbnail size. What isn’t apparent in full size can become sickeningly obvious when reduced to a small image. This is also true for vignettes, I’ve found.

Below is a crop of the original image.

Now you see I’ve selected “the mask of the face”. I.e. the front of the face, but (mostly) left out the hair, the neck etc. I haven’t been too picky as you can see. I’ve simply drawn a rough circle. You don’t have to be precise as long as you know what to look for the in the adjustment phase later

Below I’m adding a Curves layer in the layers menu on the right.

And here is with the Curves window open.

Wait, why did her face go completely white??? That’s because I’ve clicked the “Show Clipping” option at the bottom of the Curves window. It defaults to show the clipping of the shadows. As you drag the shadows marker from left to right (it’s the black pointer located above the “Input” field near the bottom left), the white area will develop solid patched of color, to show you that you’re reducing pixel values to zero and they can’t be moved any lower, and thus are “clipping” off useful data. We want to maximize contrast by bringing that pointer/slider thingie up until we just start to see patches of color.

You can see a splash of red at the edge of her face, upper left. This is her black hair clipping, and we’re not worried about that because we won’t be adjusting that.What we’re looking for is little splotches of color clipping in her actual face. You can just about see some yellow/red dots where the pupils of her eyes would be. This tells us we’ve got the darks adjusted properly.

When you move the highlights marker (white pointer on the bottom right of the graph), the selected area goes black. Now you’re looking for highlights that get blown out, or exceed the maximum possible value, when moving the slider. Here you have to be careful, because clipped highlights look worse that clipped shadows.

I usually disregard specular reflections, such as bright shine on teeth and eye catchlights. These are going to be blown out anyway, so they will be the first to show up. I wait until I get a highlight clipping on some actual skin, such as a cheek, nose or chin. Below you can see where I’ve gone just a little too far. There’s an eye catchlight, but there’s also clipping on a cheek hot spot.

So I dial it back until there’s barely anything there. Why don’t I dial it back even further, so it’s all gone? Mostly for speed. I’ll be adjusting the opacity of the layer later, which will dial back that clipping some.

And below you see where I’ve turned “Show Clipping” off. You can see the effected area of her face, which is much brighter, more contrasty and saturated. Kinda yucky. Click OK anyway to save your Curves layer.

Below I’m adjusting the Curves layer mode to “Luminosity” in the Layers palette pop-up menu. I do this to get rid of the saturation that occurs whenever boosting contrast using Curves. The Curves layer quits messing with color and saturation, and only adjusts Luminosity.

Below you see the Curves layer effect with Luminosity mode. The color of the skin matches the rest of the model now. Still needs some fine tuning of course.

Note that sometimes, if people are looking a bit pale, I might leave the mode set to Normal. This is rare, but it’s worth considering.

The Curves layer mask ought to be selected because of what we’ve been working on. Go to Edit/Fill…, or simply just hit shift-F5. This brings up the fill dialog. Select “Black” in the pop up menu, and fill that layer mask with all black.

Which removes the effect, since it’s being masked out completely (below).

Now select your paintbrush. Pick a brush size that is roughly the size of the subject’s nose. Set your opacity to 50% (see below). Your brush hardness should be 0%…I rarely set my brush to anything besides 0% actually.

The four images below show the transition as I paint. First is with the effect completely masked out (1)

Set your paint color to white if it isn’t already. Paint over the subject’s face, avoiding areas like cheeks that have hot spots on them, so as not to increase their values too much. Below is the first pass I’ve done with the brush. (2)

And then, making sure you’ve lifted your mouse/pen, do another pass. If you don’t release and start again while painting, you will only ever get 50% maximum application of paint. We want to apply 50%, then 50% again, but in an organic double-brushing action that looks better than doing a single brush at 100%. Below is the second pass I made, and this time I painted her neck too, to smooth out the transition. (3)

I thought her skin highlights were getting a bit too hot. So I changed the brush color to black, kept the opacity at 50%, and brushed quickly over the brightest spots on her nose, cheeks and forehead. It’s subtle. (4)

Below is the zoomed out version.

And then, as I often do, I adjusted the Curves layer opacity to 80% (Layers palette on the right), just so the effect wasn’t quite so obvious.

Now here is an after and before shot, so you can see the difference. It’s hard to see anything happening when you view the steps above, but here you can see there’s definitely a change. Her face has more pop, without looking like I actually did anything to it.

And the best thing is this face-contrast retouch only takes only a minute or two, tops, after you’ve practiced.

Are there better/faster/different ways to do this? Probably…and I’d love to hear about them! Drop me a comment and point me to a link, or just describe how you do it.

I should know better than to wait so long to blog about a shoot! One vacation and several shoots later, I’m having trouble remembering all the little details. But I shall endeavor to recall what I can from this shoot (see previous post here).

So above you see the shot that I had conceived ahead of time. It’s ok I guess, but it pales in comparison to some of the other images from this shoot.

When I first scouted the location, I got excited about the creek and the non-California look it has. I immediately thought:

[click "continue reading" below for more…]

I immediately thought: “fishin’!” I even went as far as to go to a local park in Ventura the day before to find a suitable stick that would work as a homemade fishing rod. Yes I know the shoot location is a wooded area! But who knew if it would have exactly the perfect stick? So I came prepared, along with string. That about sums up my contribution to this shoot: a camera, a couple of lights, a stick and some string!

I think what bothers me about this series of images is the white sweater on Irina. In retrospect, I should have nixed it, because it seems out of place. Below you can see a set up shot and where the lights are. As you can tell I’m working off the ambient here. The lights are there but they’re not doing the heavy lifting. I switched Lynae and Irina’s positions so they could both get “in front” time, although ultimately the white sweater in the background turned out to be even more distracting. Each girl is lit separately: the girl in front gets the umbrella and a brighter key (two SB-28s in tandem on a bracket), and the girl in back is lit from the side/back with either a straight SB-28 or with a Lumiquest Softbox III attached to it.I am proud to say that I actually made sure to keep the two lights coming from the same direction, to simulate daylight through the trees. I don’t always think about such things when under the gun.

I had to use some ‘ball bungies’ to hold up some very heavy palm branches, so that they weren’t blocking the girl in the back so much. Solving that problem took most of the time in this set up.

Then a very slight breeze knocked my umbrella into the stream! Ack!! The strobes didn’t take a dunk, but the umbrella took a big hit and crumpled pretty hard. AND got wet. It actually had rust stains on the cloth when I pulled it out a few days later, so I chucked it. I’m always losing umbrellas when my light stands fall over. So after this incident I decided I would buy greater quantities of cheaper umbrellas. Under the theory that expensive umbrellas, while perhaps more durable, are no more likely to survive a fall than the cheap ones are.

So after this scene, I was quite happy with the results of the shoot day (except for the umbrella and the stress of minor incidents). But it was now 5pm, and we were in Los Angeles. No one was looking forward to sitting in traffic. So I offered to keep shooting, and we all did. I turned around and went “whoa, check that out!” Big fallen tree over the stream, and some crazy dense foliage in the background. Cool!

I wanted the girls on the tree. It was about eight feet off the ground by the bank, nothing dangerous. But the problem was how do I light them? Sunlight was a little flat, so I ran a light up the pole as high as I could. Which just ended up being about face height (camera left). I’d killed my umbrella, and didn’t want to fly it that high anyway in case of a breeze. So I just used an SB-28 or perhaps with the Softbox III (see, I can’t remember!). Something light, that wouldn’t catch a breeze and go tumbling. As it was, I had to have Jackie the wardrobe stylist hold the lightstand anyway, as the terrain was too hilly.

I can’t say for sure if these are the best shots of the day, but they are certainly contenders. It was well worth staying another hour or so.

I’m reminded of two things:

1) What you plan doesn’t always turn out like you’d hoped, and what you don’t plan can sometimes turn out better. (But if you intentionally never plan anything, then it all turns out bad.)
2) Bruce Barlow has a pdf book called “Finely Focused”, which is a combination of large-format/view-camera primer, philosophical discussion and set of creative exercises to release your whatever-it-is that makes pictures better. One tip he suggests is “when you’re done shooting, always turn around 180 degrees.” You never know what might be right behind you, waiting to be photographed.

This was a very quick experiment with an improvised ‘strip light’. For those who don’t know, a strip light is a…

[click "continue reading" below for more…]

…type of softbox that is very long and narrow. A ‘strip’ in other words. It’s a rather esoteric type of light modifier, mostly used in the studio rather than on location. It has a couple of properties differentiating it from a regular softbox. Because of its long narrow shape, it creates relatively soft light along the long axis, and relatively hard light along the thin axis. So if you used one of these in the upright position to shoot a portrait, you’d have hard shadows falling across the width of the face, but soft shadows under the chin and nose. If you suspend one over the subject’s head, you have even, soft lighting from left to right, but a hard ‘butterfly’ style nose shadow.

Another feature is that it allows you to light a long object (say the side of a person) evenly, without also throwing a lot of light around the rest of the room. Best used in close though, otherwise the long-and-thin effect is negated.

There’s a nice little video here that shows you how to build a DIY strip light using fluorescent work lights. “Build” is probably too strong a word, as it’s really more of a case of “buy it and clamp it to a light stand”. Still, pretty useful and I might try it one day. But this isn’t a light I need very often…heck, haven’t needed it ever, as far as I know! No point in spending hundreds of dollars on a ‘pro’ version that I’ll use a couple of times per year—and probably use it only that much because I felt the need to get my money’s worth!

But I got to thinking about strip lights using regular strobes. At first I thought about using some plastic gutter, and firing a strobe from one end. I’d put some diffusion material over the front, and it’d turn into a strip light. It would be brighter at the strobe end, with the light falling off toward the other end. Not a bad thing sometimes, as you might want a little fall off toward the feet of your subject. But before I went out to the hardware store to buy some gutter material, I thought a little further. What about a simple V-card, closed up so that the opening was only a few inches, and then firing a strobe down the ‘crease’? Seemed like it was worth a shot. And it had the added benefit that I didn’t have to go any further than the garage.

A V-card is simply two pieces of board (usually foam board, usually white), hinged together with tape or something else. When folded in a V, it stands on its own and can be used as a reflector or a gobo/flag, depending on the color or purpose. I have a 4 ft high one made of 1.5″ wide foam insulation. The two panels making up the ‘V’ are about 2 feet wide each. At one time I used this on the beach as a light modifier, bouncing one or more strobes into it. But the spill on the ground (not to mention the wind-catching capability) eventually made me opt for umbrellas and such.

So that’s what I tried, and it seems to work as intended. The lead shot above is the result of my experiment. Note my ‘subject’ (my almost-three-year-old son) has pretty hard shadows in the creases of his forehead and down the length of his nose. But the shadows under his neck (camera left) are barely there. There’s a little fall off toward the ground, but not as much as you’d expect. The light spill is pretty well contained. Even on the couch behind him! He was about four feet away from the opening of the V-card, by the way. This was shot during daylight in our living room, with overcast skies outside. f/8 and 1/250 shutter speed if I recall correctly. His proximity to the light source allowed me to lose almost all the ambient and see just what the flash was doing. You’ll notice a little fill on the camera-right side of his face. I don’t think this was ambient. I think it was probably a little bounce from the far wall camera-right, which is a beige color.

Below is a picture of the V-card itself, with the strobe firing. Rather than taking the time to set up a light stand (and to keep my kids busy on a ‘daddy daycare day’), I had my six year old son hold the strobe over the top. I’ve got the little ‘catchlight’ reflector card on the strobe extended, to keep direct light from the strobe firing forward. You can just barely see the writing on it.

Note that the fall off—at least viewing the ‘strip light’ like this—seems pretty significant. But in actual use it didn’t seem to make much different. More subtle than I would have expected. I had the strobe set to zoom, probably the 85mm setting. This was an attempt to make the lit area of the V-card to extend as long as possible, for more even light. I could have bumped the strobe’s zoom up to the max, but I don’t know if it would have made any difference. Also, a little bit of diffusion material (like some shear white cloth) across the top half of the V-card would help even things out even further.

My version of a strip light would be hard to extend overhead in a horizontal direction, which makes the fluorescent version more appealing. The fall off would also maybe not make as much sense in a horizontal position…I don’t know. But overhead, one could use this as an interesting ‘beauty dish’ light…subjects’ noses that weren’t perfect might tend to hide their flaws, while still having a strong shadow underneath. Might be worth investing in some work lights to try it out!

P.S. I have no idea why my youngest felt the need to model with two tooth brushes in his hands. But it was very important to him that they make the shot!

This is a skin retouching technique called “degrunge”. I think I first read about it on retouchpro.com, and am elaborating on that technique. I did not invent this, but this is how I use it.

[click "continue reading" below for more…]

First, we start at the end. I was asked by the local newspaper to provide an ‘environmental’ portrait of myself for a story they were running. I guess they figured they’d save money and not send a photographer out to shoot a photographer. So I scrambled to make a self-portrait that actually had me holding a camera! Here’s the finished product:

Now here’s the image before using the ‘degrunge’ technique. I’ve made level and saturation adjustments, cloned some leaves, and done some other things just to get the image ready.

Now zoom into the image at 100%. Here you can see all the unsightly wrinkles, blemishes and the fact that I needed a shave.

Now type shift-option-command-E (or the PC equivalent) to ‘stamp the layers’, making a new layer from combining all the other layers. I’ve renamed this “degrunge layer” below. The actual image won’t change, since all you’ve done is combined everything into one new layer.

Next, select Filter/Blur/Gaussian Blur… and start with a radius of 1 pixel. What you’re trying to do is select a blur amount that nicely obscures the details you want to remove, without being so much that all detail is lost. Note that you will NOT hit ‘OK’ after this step. You’re just looking for the proper blur radius for later. If you see below, 1 pixel is not enough. The unwanted flaws, such as wrinkles and deep pores, are still apparent.

Increasing the amount to a 3.2 pixel radius, you’ll notice that the forehead wrinkles are almost completely obscured, and so are the deep pores. This is a good number to use (and I’m sure 3.0 or even 4.0 would have worked just as well). Remember this number. NOW HIT CANCEL. Don’t hit ‘OK’!

Now multiply that number by 3. Why 3? I dunno. That’s how I learned it. Seems to work well. Don’t mess with success etc. Or experiment…but it’s a good place to start.

Select Filter/Other/Highpass… and enter that blur-value-times-three, which would be 9.6 pixels in this case. Select OK. Here’s how it looks.

Now go back to Gaussian Blur, and enter your previously-determined blur number. In this case it was 3.2 pixels. This time you should hit ‘OK’. And it looks like the image below.

Now Invert the degrunge layer by selecting the proper menu item, or just hitting command-I. Which will make the image look like this:

Now go to your layers pop up menu, and instead of ‘normal’, select ‘linear light’. See the image below, in the layers menu on the right, if you don’t know where this is.

Your image will now look horrendously frightening. Like someone has slipped something in your drink. Don’t worry, just put on a mask!

Option-click the ‘create mask’ button at the bottom of the layers palette, and all will look normal again. And you’ll have a black mask for the layer (see below).

Now you’re going to paint in the mask where you need the effect. Select white for the color (I’ve still got black selected in the screenshot below). Make a brush big enough to easily paint wide portions like the forehead. Pick an opacity of 70%…you don’t need to be timid here, because you’re going to dial back the layer later. For now, you’re spraying on the “ten years younger” as much as you can. Pick a flow amount to your taste. I use around 60 or so. Below are my settings as I am about to spraypaint my way back to youthfulness.

Just reveal the skin part of the layer. Don’t unmask eyes, hair, clothing etc. Reveal the effect on wrinkles, but don’t brush in major skin folds, such as those around the mouth. Sometimes I’ll use this on lips, but usually not. The edge of the lip can present problems. I might make a very tiny brush and just paint the center of each lip. But usually not.

Right now you might be worried you’ve created a wax dummy. That’s ok! We’ll fix it.

But first, some things to avoid that I mentioned above. Note what happens when I paint over the ‘smile’ folds in my face. It looks really fake and weird, and makes me look fatter.

Also, avoid edges of high contrast, or saturated color. You get weird ghosting if you get too close. Here’s an example of where I painted too close to the collar of my shirt. Notice the purple that bleeds over the border onto my skin. Avoid this too.

Now zoom back out to assess your image. You’ll want to adjust the opacity so it looks more natural. Somewhere around 50% usually looks best (see below).

Go have a cup of coffee, and come back later. If you’re shocked and horrified about how fake the skin looks, you need to reduce the opacity of the layer further. If it looks natural, toggle the layer on and off. If you don’t notice a difference, then the opacity is too light. It’s a matter of taste of course. But you don’t want your client’s first impression to be “eeeuw, I’ve been retouched!”

Now here’s a comparison, zoomed back into 100%. At a 50% opacity, the fully zoomed image might look a little too much. But that’s probably just right for making a print.

And comparing it with the degrunge layer turned off.

Degrunge is good for blemishes and wrinkles. It will also get rid of glossy, sweaty highlights. If you’ve got a giant shiny spot, you can even do a separate degrunge layer with larger values, and you can smooth out the highlights that way. It will also to some degree remove stray hairs. Although some cloning is usually necessary to finish the hair removal. If you have some deep shadows, sometimes large values will allow you to smooth the transition areas. It’s all dependent on your particular image situation.

If you work with the same types of images a lot, you can set up some Actions to use common values. The necessary values are dependent upon size of the skin area in pixels. So a headshot like the one above requires larger numbers than a family of four, with correspondingly smaller heads. I have a ‘headshot’ Action and a ‘small head’ action, that takes care of about 90% of my images. I just have to paint in the masks properly.

This is not a click-and-go technique, obviously! It requires manual labor, and good taste on your part. If you overdo it, you images will look horrible. But used properly, it makes skin look nice and smooth. And it retains texture at the same time, unlike techniques that simply blur the skin selectively.

For those of you who have been waiting for me to post this tutorial, you can see why it took me so long! I hope it’s been useful. Feel free to post questions if you have them.

Here’s a more in-depth tutorial on the “Highpass/Hard Light” look. By the way, this probably has a better name, but no one’s ever told me it if there is.

[click "continue reading" below for more…]

First the lighting: This is shot in my living room. There is a piano behind me, as well as a beige wall. Where’d it go?? I set my shutter so that the background fell almost to black. I tried to balance it so there was perhaps a little ambient fill in the shadows on my face, but that has probably been negated by processing. The main light is an SB-28, set to full power, through a homemade grid (aka grid spot or honeycomb grid). I made it using drinking straws and a plastic gutter accessory. But that’s a story for a different time. The rim light is a Sunpak 333 with a “window green” gel on it, set to (I believe) 1/4 power. I set the flash to wide angle, and aimed it at the back of my head. Both strobes are triggered using cheap ebay wireless triggers.

But hey, this post isn’t about the lighting! It’s about the Photoshop.

So here’s my starting point. I exported the RAW file from Lightroom, made some super secret retouching steps to make myself look pretty (yes, this is me). I’ll go over those in a future post. As you can see, a handsome model but the image is very normal.

The first step is to hit shift-option-command-E on your keyboard, if you’re using a Mac. If you’re using a PC…well do we really need to get into that? Go buy a Mac, I’ll wait here.

This key command merges all visible layers to a new layer. See below for the result, which is a new layer that looks remarkably like the original. You won’t see a change to your image just yet, since the merged layer is just a duplicate, sitting on top of your existing layers.

With this new, merged layer selected, select Image -> Adjustments -> Desaturate. Your new layer will now be in black and white.

Now go to Filter -> Other -> Highpass. Enter a value of 100. Why 100? Because it seems to work. Feel free to experiment with other numbers, for different effects. Click OK. You should see something like this:

You can see above how the highpass filter has brought out some detail in the background (over my right shoulder) , that otherwise was not very visible. I mask that out later, along with the green rimlight (see below).

Now go over to your Layers menu, and with the merged layer still selected, choose “Hard Light” in the pop up menu above the layers.

The end result is what you see below. You could stop now if you wanted, but realistically you’re going to want some adjustments. The effect has a tendency to emphasize edges in weird ways, so you’ll probably mask some of that out. You might want to back off on the effect when it comes to skin, as it can make people look a little ugly sometimes. Sometimes I’ll paint out the skin but leave the eyes and everything else effected.

You’ll notice my screenshot (the image before last) includes a Hue/Saturation layer as well. This is an optional step, but it seems to go well with the general look. In this case, I didn’t desaturate everything. Rather, I left the green at normal saturation so as to retain the edge light effect. (See more on that below.) I did this by selecting all the other colors besides green in the Hue/Sat window, and reducing them by about -40.

If I were doing a family portrait, I would use a mask to remove this highpass/hard light effect from the faces and skin of the subjects, as well as remove the desaturation. Leaving only their clothese and the background effected. With something more commercial in nature, such as the Keith Clark or Model Portfolio shoots I blogged about recently, I’ll keep the effect on the skin but perhaps mask some of it out. In the image we’re working on here, I left it set to 100% for educational purposes, but then I masked out the green rimlight area. The effect can sometimes blow out color on edges of things. Below is the mask I created for the highpass layer.

And the final result, cropped so that my head is partially chopped off (just the way I like it). Some brightness/contrast adjustments too.

I should have shaved before taking this self-portrait. Ah well. Feel free to ask questions in the comments section.