Onyx Illustration Guide


Most important thing to make note of: add layer prefixes to distinguish between different elements. This will make things much easier down the line. Photoshop’s later palette filtering will allow you to toggle only shadows, highlights, etc, for batch operations. It’s a nightmare to deal with otherwise.

  • [S] = Shape Layer
  • [T] = Texture Layer
  • [-] = Shadow Layer
  • [+] = Highlight Layer

For textures, we pull from The Walter H. and Leonore Annenberg Collection using The Met’s Open Access tool. These images are licensed as CC0, so we can do whatever we want with them.

Download or clone the repo from APPC’s GitHub to get all the images at once (requires login).

Organizing Files

We use the standard APPC file federation here… listing initiative code, project name, taxonomical descriptors, and subject using BEM-style separators. It sounds daunting but it’s reasonably straightforward in practive. Here are some examples.

  • source/ac-illo--topic__women.ai
  • source/ac-illo--utility__not-found.psd
  • source/ac-illo--constitution__amendment-09.ai
  • sketches/ac-illo-sketch--landmark-cases-banner.ai
  • renders/ac-illo--media__timeline--light--RENDER01.jpg
  • renders/ac-illo--utility__trophy--dark--RENDER01.jpg

In Illustrator

After creating the composition, the layers need to be broken apart and named. You can select a layer and choose Release to Layers (Sequence) from the Layers panel menu. This will put each shape object on its own layer.

Normally I try to keep multiple shape objects that make up the same element (like, light/dark components) as one layer, just to keep the number of layers from becoming overwhelming. Plus I find it’s handy to have a single layer that contains the entire shape that you’d be using for masking instead of having the different elements spread out between layers.

At this point, all the layers will get the [S] prefix because we only have vector shapes. The other layer prefixes aren’t used until we’re in Photoshop.

Go to File > Export > Export As… Choose PSD from the dropdown menu of file types. Make sure the Write Layers option is checked, and Anti-aliasing is set to “Art Optimized.” This will make the edges of shapes aliased, which is important for using them as selections in Photoshop.

In Photoshop


Super helpful tip: Hold command and click on a shape layer to make a perfect marquee selection of that layer. Then select the texture layer and click the Add Mask icon in the layer palette and you’ll have a wonderfully aliased mask applied.

You’ll probably want to unlink the mask so that you can reposition the texture freely without the mask moving.


I use a rich, dark purple (hex code #3b1b45) for any shadows.

This is super important: Don’t hand draw shadows. Since the shapes have such hard edges, dodging shadows by hand creates an odd dissonance. Instead, make a fill using an edge selection of the shape that’s casting the shadow.

Typically I’ll mask these, set the layer opacity to and add a Gaussian Blur of ~30px (depending on how “long” the shadow should be). If the shadows should be so dark that they’re almost black, you can change the blending mode to Linear Burn.

I typically use combinations of Linear Burn and Normal layers to get the depth I want.

Again, if you’re using a mask, unlink the mask before you apply any filters like Gaussian Blur. Otherwise the blur will apply to both the layer and the mask edges.

For harder, non-diffuse shadows, I get good results using the Path Blur and Field Blur in combination.


Since we need to export these as JPEGs in order to get a useable file size, we’re going to take the entire composition and flatten it into a new document. Select the entire canvas (cmd+a) and copy merged (cmd+shift+c), then paste into a new document. Set the background color to either Annenberg Classroom $black or $white (#3c3c6d and #fffdfa, respectively).

Export using the Save For Web (Legacy) option. For smaller elements like topic images I reduce the output size to 25% in order to get it about 1000 square pixels. Then adjust the quality slider to try and get a file size below 150kb (under 100kb if possible, but there are some quality compromises at that size depending on the tonal transitions in the image).