Cat's Continuous Banners Tutorial
Today, we'll be making +these continuous banners from the September '17 UC
Before we get started:
- This is semi-translateable: as this is mostly composition focused, program shouldn't matter too much. There is selective coloring and topaz filters. I used Photoshop CS6 for these.
- This is an intermediate tutorial; it assumes some fluency with your program.
- I keep everything as smart layers (I almost always add images by dragging the image file from the group straight onto the photoshop canvas.) You will end up with a giant file (this one ended up being almost 400 MB), but it helps with image quality and indecisiveness, since nothing is permanent. I always use layer masks to blend for the same reason, either with a soft brush (0% hardness, opacity and flow at 100%, size varies), or the polygonal lasso tool for hardcutting.
- Actually, I usually hardcut on the full size image on a separate canvas, so I can save the png to reuse or share.
- There are a lot of layers- well over a hundred. Groups will help. The background stock is put into groups by type of stock (lightning, trees, etc.) and the people and text are put into groups based on which banner they're from. Coloring and filters also get their own groups. Seriously, please use groups.
- I made these over the course of a few days and not in any kind of reasonable order, but I have attempted to have a bit of a process here.
- If I don't say what mode a layer is on, assume it's on normal. For opacity, it's 100% unless stated otherwise.
- Any images provided will be linked within the tutorial.
Step one: spend a few weeks picking out songs for the fanmix and crying about NaNoWriMo and your OTP.
No, but really. When you're trying to make eight graphics in one go, it helps to have a plan- like a list of songs/texts already picked out that can guide you in image choices. Before I opened photoshop, I knew that I didn't want to repeat any fonts and that I didn't want too much of the text and people overlapping multiple banners. Since The 100 is set in the mountains of
Step two: The Base!
First, throw a bunch of my +lightning textures onto a 700*880 canvas.
The one lowest on the stack is on normal, and the rest are on lighten. I used layer masks and a giant soft brush (~300 px) to tone down the brightest parts, and make sure there were no harsh edges. All of those layers went into one group.
I added some very basic tree "textures" next. These are literally just desaturated stock images from unsplash (similar to +this) with the contrast adjusted (so is the lightning tbh). One of them is inverted. There are five layers of these total. I used all different ones to add some variety.
The tree layers covering the brightest parts of the lightning are set to soft light, the layers on the darkest parts of the lightning are set to lighten, and the inverted stock is set to hard light, which, in this case, is sort of the middle ground between lighten (too bright) and soft light (basically disappeared), all on 100% opacity. I masked out quite a bit of these layers, to uncover other things and remove harsh edges. The lightning should be the brightest part of the base. This is a pretty vague and sloppy part of the process. The layer modes are pretty forgiving, and everything will vary based on your exact images. Put them all in another group.
Time to add some color! I took a few leaf/ivy stock images and added them on, once again in a new group, also once again on lighten or soft light and softly blended in. The top half ended up being pretty distinct from the bottom half, due to the progression of the fanmix/story: the first half is more wintry, the second half is more homey. I also picked the window because of the convenient dividers (which make placing people later on a little easier) and the vertical shapes (which helps with the 'these are very much continuous graphics' thing).
Step Three: Lighting textures!
Open a new canvas, 700*110. Make some lighting textures, one for each banner. Copy them into a new group in the original file and set to soft light.
I used old graphics of mine (desaturated + lens blur + play with the contrast using your contrast method of choice) to make the lighting textures, but you could also just paint blobs of black and white onto the canvas and blur it a bit. You could also just use the gradient tool to make some very basic lighting textures, but the more interesting they are, the better the coloring will turn out.
*reminder: you can only recycle your own graphics like this otherwise it's ripping which is a. uncool b. rude and c. will get you demoted and possibly banned*
Everything seemed a bit too dark, so I duplicated the lighting group, and set it to screen ~40%. Later on in the process, one banner needed more depth so I added a gradient in here (just white to black), and another was a little dark, so I brushed some white onto a new layer over that lighting layer. You can always play with the lighting as you place your other images. If you don't put things into groups, you now have nearly twenty lighting layers floating around to deal with.
At this point, I added a text layer with all of the song titles, in order, to make it easier to pick images. That made me realize that banner #4 was a bit gray and bland, and also the song references a mountain top. Clearly, it's time to add some mountain stock.
I dragged +this image from the folder into photoshop, and it landed under the leaves group. It looked pretty cool, as is- a giant white/purply blue blur- so I set it to soft light. I then duplicated it, set it to normal, resized it, and soft blended it in. These two layers got put into a group and labeled. The group order is presently, from bottom to top: lightning, trees, mountain, leaves, lighting. These groups got put into another group, called Base.
Step Four: Coloring!
For anyone wondering why you'd want to put the coloring on now: Consistancy. It makes it easier to get all the people to look like they belong together, if the coloring affects them all the same way. I did spend a lot of time toggling the coloring group on and off, but I did add it at this point. I also had already done the coloring, since I wanted the banners to match my NaNoWriMo cover and the fanmix, so really, might as well just toss it in here.
This has also been, essentially, my coloring process for every graphic over the last several months. Basically, I darken the image and reduce contrast, reduce the colors present, then add back color and contrast using selective coloring and curves, then finish off with a couple vibrance layers. Sometimes I throw in some color fills or photo filter or extra gradient maps, but this is pretty much it.
1. Gradient map, #290a59 to #ff7c00 (the default purple-orange one) - Difference, 15%
2. Gradient map, #431605 to #a5b0ad (reddish-brown to greenish-gray) - Multiply, 60% (got this from Violet's monochrome tutorial)
3. Curves - looks like this:
4. Hue/Saturation - reduce saturation to -100 for everything EXCEPT FOR Master, Red, and Yellow
5. Gradient Map - #b9a08b to #ffffff (light brown to white) - Color Burn, 40% (this is the one I add whenever I want a bit of yellow. sometimes it's on soft light, opacity varies)
6. Selective Coloring -
Reds: Cyan -6, Magenta, Yellow, and Black all approximately +25
Yellows: Cyan +8, Magenta and Black +25, Yellow +100
Whites: Black -50
Blacks: Black -5
If you don't have selective coloring, the goal here is to generally brighten, and then intensify the reds and yellows. Another Hue/Saturation layer and any kind of brightening layer would work.
7. Photo Filter - default Sepia (or a brown color fill, set to hue or color on a low opacity)
8. Vibrance - Vibrance -63, Saturation +80
9. Vibrance - Vibrance +58, Saturation +31
(Or just increase Saturation)
At this point, when comparing the coloring to the cover, I decided it needed to be more green. So, underneath all of the other coloring layers:
0. Color fill layer - #a3a46d, Overlay, 25%
This was still insufficiently green, but raising the opacity was too much, so, under that layer:
-1. New layer, Overlay, 100% - just brush on some blobs of that same green color. Wherever. Just don't overdo it- we still want it to be mostly brown.
Stick all of these into a new group.
Step Five: People and Text!
Most of the images I used are either in my The 100 PNGs pack or will be eventually, once I finish with the later seasons. Pretty much any people will do, though.
I had three groups of people: One goes over the lightning, under the trees. Two goes underneath the leaves, above the mountain. Three goes above the Base group. Additionally, I made a group for each banner in each of the main groups. Unless otherwise specified, the banners go from the top down- Radioactive is at the top of the stack, North is at the bottom:
Put some background people into group One. Don't worry about how washed out or difficult to see they are, we'll fix that later.
More (mostly background) people into group Two. Most of the images were already cut out before I placed them onto this canvas, so most of the blending on here is either done with a large soft brush or I'm just masking out the top/bottom to avoid overlapping onto adjacent banners.
Some of the images are quite dark (thanks, screencaps and night scenes), so we're going to fix that. For each image, create a new layer and clip it on, set it to soft light, and use a proportionate soft brush and add some white. For the screencap of Bellamy's hand on banner #5, that wasn't enough, so instead, I added a black-white gradient map set to screen, clipped on, and then duplicated the layer until it was bright enough (I was trying to match the screencap next to it). For this one it took four.
Text for two of the banners went into this group:
'Radioactive' is in +Time Machino, #ebc693 (pale orange), 48 pt. It's in this folder so that we'll get the cool effect from the leaves.
'One of us' is in +Broadcast Matter, 98.17 pt, #bdc1cd (pale blue). 'One' is on its own layer, and 'of us' is on another, placed so that nothing overlaps. Each layer has a gradient map on overlay (the layer effect) - on 'One' it's white to black, on 'of us' it's black to white.
Everything else goes into group Three. Started with the text for some of these.
'revolution' is in +Ruthless Dripping Two, #c5b29f (medium beige), 60 pt, with a white-black gradient overlay on soft light, and a drop shadow in black (1, 0, 1) on normal (my default drop shadow). I wanted to make sure there's some variety in text placement. Most of the other banners have text either centered or toward the right. The drips are also a bit much, so I knew I'd be putting something above the text here.
'us' is in +Vif Argent, 120 pt, #8e8e87 (medium gray), with a black-white gradient overlay on overlay. It's set to Color Burn, ~85% because I didn't want it to be too dark or cover up too much of the background. I went bigger with the text here mostly because I could. The people were placed around the text.
'I need my people' is in +Signerica Medium, white, 31.1 pt with a drop shadow in black, 1, 0, 1. I added a couple of spaces between 'need' and 'my' to leave room for Clarke, and made sure the smaller Clarke was in the gap between 'my' and 'people.'
Time to fix some of the insufficiently visible people, which I mostly did by duplicating the layers and moving them up into group Three, then blending as needed, except for on 'Us' which I dealt with later.
The hardcutting on Bellamy on the Radioactive banner looks atrocious. I hid that by darkening the background- I used the magic wand tool to select the background around the png, then, on a new layer, just stamped some black with a large (~300px) soft brush just behind Bellamy's shoulder (looks like +this, with white added so you can see).
Next up: foreground people. There are two things I tried to keep in mind here. First, variety in placement and sizing, which is actually easier when you can see all the banners at once. It's often a good idea to try to place people differently than in your other recent graphics, or if you're hoping to put the graphic in your gallery, you can reference your gallery and avoid image placements you've used recently. On the 'One of us' banner, for instance, the main picture of Bellamy is from the eyebrows to mid-neck, which I didn't really do elsewhere in the set. While the other three people are similarly cut (mid-chest to above their heads), two of them are upside down, and the third is in profile. Second, make sure you leave some room for the text.
Now the rest of the main text:
'Anna' is in +806 Typography, #29292c (dark color eyedropped from Bellamy), 36pt - same font used on the cover/fanmix
'Sun' is in Sudestada (now a paid font), 97.92 pt, #d5ae6c (another pale orange), with a 1px stroke in white and a white to black gradient on soft light.
'trouble maker' is in +Antro Vectra, 92 pt, #cb9861 (orange), line spacing at 34.33 pt. I put a few spaces before 'maker' to make it line up nicely.
- Stroke: #555762 (dark cool gray), outside, multiply, 50%
- Gradient overlay: white to black, soft light, reflected, scale 104%, make sure 'align with layer' is checked
- Drop shadow: black, 1, 3, 4
Last group of people. I basically just filled in any empty central space that I wasn't saving for subtext. Any model images that aren't meant to be the focal point should either be smaller than the focal people (usual method here), obscured in some way (most useful for bigger images, like on the Anna Sun banner,) or a less eye catching color (darker or less saturated, usually). In a set like this, you definitely want to vary the ways you place people. I also tried to use existing elements to guide image placement: the text on Radioactive and Troublemaker, the leaf on Anna Sun, the windows on the bottom four.
You'll notice a lot of dark or red faces. Usually, that would be a problem for the coloring or lighting to fix, but we already did that part, and it's impractical to try to alter the coloring to work with every single face here. So, we're going to fix the coloring on each image instead, either with black to white gradient maps, set to screen, or soft light layers with white, duplicated and adjusted as needed and clipped onto their respective layers. On the Us banner, to darken the background screencaps, I used a soft light layer with black.
If, in the process of brightening the images, they start to look washed out or gray, I like to use a vibrance layer, clipped on directly over the image, underneath the brightening layers, and increase vibrance and saturation as needed (usually vibrance +100, saturation +20 to 60).
On the Revolution banner, I also added another screencap with a couple of torches on it. The torch from the screencap on the right was the brighest part of the banner, but erasing it didn't really work. If you have one quite bright but uninteresting element, it will call attention to itself, but if you have three bright but uninteresting elements, they'll basically just fade into the background, as long as there's something else more interesting to look at.
Step Six: Filtering!
I always filter before the coloring, so go ahead and turn off that group. I also copy-merge for every step and apply the filters to a new layer. I do use Topaz, so sorry if you don't have/use it. All of this is in a coloring action, so I don't even really think about it anymore.
- Topaz Clean - Edge Boost, 60% opacity
- Topaz Detail - my settings +here
- Paint Daubs - 1, 1
Step Seven: Lightning! (again)
The lightning wasn't showing up as much as I wanted, so I duplicated the group, then merged it, and dragged it up between the filtering and coloring, then set it to Lighten. Then I stuck a layer mask on it, and erased over faces and wherever it got too bright.
Step Eight: Subtext!
I almost always do any subtext last, so that it isn't affected by the filtering and coloring. Alternately, I'll put it on early, but keep my subtext group at the top of the stack, and toggle it off when filtering so it isn't included. I'm almost always thinking about subtext placement, though, even if I don't put it on until I'm almost through. Just like with the main text, it's much easier to build a spot for subtext to go, where it'll be legible, rather than trying to change things later on to make sure it works.
For continuity's sake, I used the same font (Orator Std, but any sans serif text would work pretty well), size (9 pt), and tracking (200) on every banner, and just two colors: one light (#c7c4c0, greige) and one dark (black). The line spacing did vary a bit (from 7 to 10 pt) depending on my text to available space ratio. Two of the subtext layers had drop shadows:
- 'they'll name a city after' - #d2cecc, 0, 30%, 19
- 'everybody needs a place to call their home' - black, 0, 0, 5
And that's it! Feel free to post here or pm me if you have any questions, and if you try any of this out, I'd love to see
cat edit 2/20/2018: +here is another set of banners made with the same process