So this tutorial (among 2-3 others) is long overdue. For that, I apologize. I had a tutorial block for awhile, and though I had 80% of a big guide + another tutorial written, I wasn't comfortable posting them as I didn't think they were...I dunno. ANYWAY.
Today I'm going to talk about composing icons with multiple images, blending with silhouettes, and, in particular, this Snow White icon. This fills the requests of the following people: deternot, fouroux, rowofstars, and applepips16.
I honestly don't know if this will help anyone, to be honest; it's kind of a long ramble about things that I think. Hopefully it answers some questions? Anyway, I'm going to link other guides at the end that I think are more helpful. I just didn't want to let the requesters down. So here ya go!
I would absolutely love to tell you all that I have this very scientific process brought about by lovely, mathematical formulas for picking screencaps and that I am unequivocally a genius. Alas! This is not the case.
No screencap selection, for me, tends to be me sitting in front of a gallery for whatever show I am iconning and browsing through for caps that catch my eye for whatever reason. Yes, this process takes ages for me just like it does for you. It is also boring. I tend to pepper my screencap searches with actually working in Photoshop. This breaks up the monotony.
I, personally, find that it helps to go ahead and pick out a whole bunch of pretty screencaps that you might find fun to icon. Then you have a short list selection once you start the actual iconning process. I tend to have at the very least ten caps open at once so I can see what works and what doesn't. Sometimes I have to go find something different because my eyes overestimated my ability. However, the process is less daunting this way.
Things I might look for in my screencaps: pretty colors, pretty people, emotional expressions, and contrasts; occasionally, I'll pick out screencaps that look challenging just for funsies.
The important thing to keep in mind while making icons that include multiple caps is a) how they might fit together spatially and b) how they might fit together thematically.
For example, the Snow White icon, thematically, shows both sides of her character, the one in the real world and the one in fairy tale land; that was the point of the icon. However, I also had to take into consideration how I would fit images together spatially. I found the silhouette image first (I flipping love silhouettes; you might have noticed), and I knew I needed something that fit into the silhouette in an aesthetically pleasing way. I wanted to make sure I wouldn't awkwardly crop any other image that went inside the silhouette. I also wanted to make sure the image would be visible inside the silhouette. This knocks out extreme close up caps and also extremely wide caps as contenders for the blend. Just like that you've narrowed down your search field and made your screencap hunting mission easier!
So the tl;dr of the cap hunting section: go to a gallery (or to your cap folders), find a bunch of pretty images, and get to work. if you're going to use lots of caps into one icon, do a little mock up of what you want the icon to say and then figure out how certain caps will fit together in a pretty way so you don't get any awkward surprises.
You've got your caps; now it is time to blend them together!
Layer masks + screen/lighten/hard light mode (these are the modes I typically use; try others though!) + soft brushes = fun times for blending!
The usual process goes like this:
1. Find caps (see section one).
2. Do base prep (blurring, levels, lighting maybe, a little color tweaking, etc).
3. Put caps on the same canvas and move around trail-and-error like until a good figuration is found & 4. Do the same thing with blending modes. Usually screen works pretty well (and you can tweak using levels and curves to get the right depth if the cap you are blending in doesn't look quite right), but lighten is a good fall back. Recently, I've been using hard light as well. It can also have the added benefit of doing fun things to the colors of the icon.
4. Use a layer mask and a brush (I usually use a soft brush, but any type of brush will work as long as you know what you're doing with it) to get rid of anything in the blend you don't need.
Quick visual aid:
For the Snow White icon, I blended on the original screencap canvases. Currently, I do my base prep on the original screencap as well. I find this gives me a little more wiggle room later on when I'm cropping. If I've got my base prepped, I already know where the highlights and shadows are going to be. I know which parts of the cap probably won't have any color life (and should be cropped out), and if I change my mind halfway through, I can go back to my base cap and play around with crops some more. This method works for blending as well. Sometimes (most of the time, lbr), you can't just layer one cap over the other and make magic. You need room to move them around and resize the caps and whatever else. Starting big (and then working smaller) helps me do this.
So I had my two Snow White images (which I would link except they were from the Rawr-Caps gallery, and I dunno if the Grande-Caps galleries have the exact same caps or not. Both images were from Snow Falls, I think so if you absolutely need to go out and make your own copy of this icon for practice, you can find the screencaps yourself). Keeping them at full size, I did my usual base prep routine (brightening them up & adding contrast back in), and then I put Snow White on top of Mary. The Snow White layer was then put on screen. Then I used a layer mask to get rid of anything I didn't want. However, judging by the screencap, I probably didn't have to mask overly much (white or light colored space are good like that). Then I chose a boring center crop because it showcased the blend the best and resized down to 200x200 'cause that's how I roll.
Before we move on, a few quick tips for making your blending work:
[x] Unless you are using a silhouette or one subject is significantly smaller (and cropped differently than the other), don't do face in face blending (* this is just a general rule of thumb; once you get things down pat, you can kind of make things like this work; just go with your gut). That's how you get ugly double face monsters of doom. With a silhouette, one face should be more or less completely blacked out with just the profile showing which means whatever you are blending into it won't merge in weird ways with the facial features of the other. If you aren't using a silhouette, but are using a regular close cropped face, try making sure your second cap is a mid crop or a wide crop and blending that way (example).
[x] This will sound weird, but just work with me: Cap A is your base cap, and you layer Cap B on top and try and blend it in; it doesn't work. So try the opposite! Put Cap A over Cap B and blend that way. Strangely enough...this sometimes works or gives you options you didn't have before. Weird, I know.
[x] TRY INVERTING THINGS. I know, you are giving me a weird look, but try it out. You might get some fun results! I got this, for instance (which, okay, was originally inverted the other way, but whatever).
[x] Leaving in bits and pieces of the cap you are blending in is okay. See Snow White up there? The fun textured bits in the background are trees from the forest in Cap B.
[x] To repeat: When I work, I try blending at full cap size and if I get nothing like that, I also try blending at 200x200 (which seems to be the canvas size I like to work with right now). It may also be that, in order to get a good blend, that you have to make one image smaller than the other or only use pieces of it. Just...don't be afraid to play around with sizes.
[x] YOU CAN MAKE YOUR OWN SILHOUETTES. Try using levels, curves, soft/hard light, and other tools to darken the profile shot to silhouette acceptable levels. Also, a tip I gleaned from alfiri, try upping the saturation up all the way. This, if done correctly, can smooth out the features of your subject and make it easier to turn into a silhouette. Brushing also works, of course, but this is my least favorite method because it involves me coloring inside the lines. Boo.
[x] Quick tip (which is less important if you are using silhouettes, to be honest, but...) for blending in general: it helps to blend on a "line" so to speak. You'll make less deformed face tumors if you work this way. Find some sort of edge to blend into like a shadowy side of the face (like this) or the edge of someone's hair (like this).
Stop! It's coloring time!
Firstly, I wanted to take care of what, at the time I saw as a flaw, and that would be that some of the darker parts of the icon look a little LQ because of the brightening I did back in base prep. So I went into variations and made it a little bit darker (Image > Adjustments > Variations > Darker [1 click]). It's a very small adjustment, and I'm just taking a shot in the dark as to my though process behind it because looking at the .psd file now, I'm kind of laughing at myself. Ahem, moving on.
The biggest coloring adjustment in this icon is a Hue/Saturation level. Yes, the old school PS 7.0 user's answer to vibrance. Whatever. I like it. So I added a new Hue/Saturation layer and upped the saturation to a whopping +45. Hot damn. This over-saturated Snow White's face so I masked a little of it away (soft brush on black with an opacity of about 50% I'd say; alternatively, you could use a soft brush on grey to achieve the same effect). Makes a huge difference.
And that's basically it for deliberate coloring decisions made by me. I know. It looks nothing like the finished product, right? WELL, THAT IS BECAUSE WE AREN'T YET FINISHED. It's just that the rest is...
LIGHTS. CAMERA. MORE LIGHTS!
1. A really, really bright white/yellow blurred six ways from Sunday layer on screen. [x]
3. Copy merged onto a new layer + auto levels to add some serious contrast back in! [x]
2. Two curves layers. Layer one: adds darkness back in. Layer two: Okay, I lied; technically this is a color layer. But only because I needed to add contrast back into certain places. This layer messes with the green and blue settings. [x] | [x]
3. Aaaaaaand then, because I am contrary, a variations layer to counteract the cold colors that the previous curves layer added in. So the icon becomes a lot warmer. More reds, more yellows. [x]
Text is your friend, not foe!
Okay so...the text part of this icon was definitely a difficult thing to get right. I still don't 100% think I achieved what I set out to do with it.
Anyway, let's talk font choice here. There's two different fonts in this icon: FairydustB (for 'SNOW') and Sweetly Broken (for 'White'). I wanted my font choice to reflect the theme of the icon (so Snow's tow different personas) so I chose an ornate font for 'SNOW" to reflect the mannered princess personality of Mary and a free, wilder font for 'White' to reflect Snow's rebel streak.
Font color: Basically just used the eye dropper tool and chose a yellow color from the icon for 'SNOW'. Layer one of 'white; is, obviously, white. Seemed fitting. There's a second layer of the same text behind it in the same yellow color as 'snow' to make 'white' pop a little more.
Font positioning: Oh god. This was so difficult. There's hardly any room for text! I wanted a longer quote first ("He thinks you're Snow White", I think), but nixed that idea really quickly for obvious reasons. So I shortened it to just her "name" which is super uncool, I know. Fun fonts make up for it, y/n? Anyway, putting the text at the bottom was really my only option. To save room, I put 'white' on top of 'snow' which is actually a technique that I had been dying to pull off well for ages. I had tried before, but it never seemed to look right. All text layers are on their own separate layer for easy moving.
All the text is blurred (gaussian at a 0.3 radius). And that's it!
I sharpened Snow White only (not the text) and then resized down to 100x100.
Other blending guides of note: by alfiri | by tinebrella | by raiindust | by sarisafari