#parallax effect

Part 1: In-game Screenshot with Depth Buffer

1) If you are using ReShade 3+ you can follow this tutorial regarding its installation, otherwise copy the ReShade folder, ReShade.fx and ReShade64.dll (or ReShade32.dll if the game is 32bit) into the main game path (where the exe is, so to speak), then rename ReShade64/32.dll as dxgi.dll.

2) With ReShade 2 go to ReShade directory > Presets > Default and drag Shaders_by_CeeJay.cfg into Notepad, then scroll down to the very last line where you see a shader called DisplayDepth. Next to #define USE_DisplayDepth change the value from 0 to 1. With ReShade 3 press Shift+F2 and look for the same name, if you properly downloaded it along with the other shaders of choice it will most likely be already activated. It is highly recommended to set a hotkey for this effect so that you can easily turn it on and off in game (for example, #define Depth_ToggleKey VK_F12). 

3) Open Pipeline.cfg and look for the line #include EFFECT(CeeJay, DisplayDepth). Select it, cut it away and then paste it to the very top before any other shader listed. This will allow you to obtain smooth shapes instead of jagged edges (you should also play around with AA options, specifically SMAA and/or FXAA, both included in CeeJay.cfg). Sadly I’m not sure how the setup works in ReShade 3+.

4) Now that your options are properly set you can start the game. Find an area you like and want to shoot, then press F12 (or any other key you chose for it). If the game has a photo mode or custom made free camera tools, you should totally use those for the best results. You can check this post for a few listed titles and all their respective links.

ReShade allows you to automatically store your screenshots directly into the game’s folder by simply pressing the print screen key. This is how the scene looks normally, without UI and with the free camera activated.

This is what happens when you turn DepthBuffer on.

Note: If all you get with Depth is a fully white or black screen, chances are you need to “reset” by simply entering the fullscreen mode (and switch back to borderless/windowed if that is what you usually like playing with). ReShade usually reloads whenever you apply a modification to the settings, but if even after this the effect doesn’t show up it probably means the game has no DepthBuffer support (or there is some incompatibility somewhere, which may be related to other shaders).

Part 2: Editing

1) Grab your depth screenshot and open it with your editing program of choice, I use Photoshop. In order to animate this, you have to create different layers (renaming them for an easier use in AE, but that is up to you) being sure each hole is covered, or else the empty part will definitely show up in your animation.

2) Make a selection around the areas you want to separate, I’m starting with what is close to the “camera”. Try to be as precise as possible to keep most of the details, especially around the character. Once you’re satisfied with your selection press CTRL+X to cut it out and then paste it into the same position, a new layer for it will be automatically created.

3) If you hide the layer you’ve just made, you will obviously notice an empty shape on the background right where you cut it. The easiest way to fill that properly is by selecting the layer (CTRL+LMB on it), inverting selection and going to Modify > Fill > Content-Aware.

As you can see the empty area has been filled completely, but a few fixes are required. You can use the clone stamp tool or the healing brush, in this case I choose the latter because it makes the process quicker.

It can be a bit tedious in some places... especially if there are a lot of buildings or trees, depending on your scene. Photoshop covered the outline using the existing content to generate the missing parts as I dragged the brush over it, but you may want to use the clone tool as well and even draw here and there where the brush failed.

Repeat the process for all the other layers you want to separate and then rename everything to know what you’re gonna edit in AE (again, this is optional). You can also make the back layer bigger to hide imperfections behind the other layers, which won’t be too visible while animating. Further edits will be done in AE later anyway.

Part 3: Compositing

1) Save the PSD file and load Afer Effects. I’m currently using the CC 2015 one, but even if you have an older version this method will work without problems.

Go to File > Import > File and select your PSD. 

Be sure the importer is set on Composition and Editable Layer Styles is checked, this is essential if you want to work on each layer.

2) Double-click on the left to preview the PSD and convert every single layer into 3D. To do so simply check the third square on the right, under the cube icon:

Now we need to add a Camera. Go to Layer > New > Camera and leave the options untouched (the default configuration should be okay for this).

3) Right under your preview you should see a button called Active Camera. Switch to Custom View 1 to move the layers and create depth between them.

All you have to do is select each layer in the lower panel and move it using the blue arrow (z direction) to push it backward or forward. Use the order you’ve got as reference for the best result, then change the view to Active Camera again when you’re done.

4) Now that you have changed the original positions you need to fix their size to fit the composition. Select all layers at once, press S and activate the little watch icon that appears followed by the effect name (Scale). You will see three values then; place your mouse over the 100,0% one and drag it to the right or left in order to resize the layer(s). Use the directional arrows to also move them up or down eventually.

The stopwatch is absolutely fundamental, because it tells where to place each frame in the timeline. Whenever you apply an effect or any modification you need to be sure it’s recorded at the very start, this way it will affect the entire sequence and you won’t risk to only see it taking action in a few frames later instead.

5) You can move the upper bars in your timeline to change the final time. A few seconds for this kind of animations will suffice, I always set it around 4s.

Part 4: Animating

1) Now we can finally animate, woo!!

It is highly recommended to make all the planes bigger than the composition, so that you can move them around without risking to get empty spaces. In this case we want to fake a camera movement with a sort of 3D perspective, like the gifs I’ve posted in that photoset.

Let’s start from the first layer. Click on it and press P (position) to store the current location on the first frame, then move the time indicator at the end of your sequence. I want the main layer to slide horizontally, thus I will move the red arrow on X axes.

As soon as your first animation is stored, another frame will be generated at the end of the timeline. After Effects automatically creates all the frames between the first and the last one, if you are happy with it select the second layer and do the same, then the third and repeat. You can also use the R (rotation) key to add more realism. Remember to ALWAYS check the stopwatch on frame 1 before doing anything else.

You can preview the animation with space bar, the rendering times change according to the amount of virtual memory installed on your PC.

For a nice camera shake you can follow this tutorial.

2) Once you’re happy with your animation you’re ready to export it. You can pick a simple video option, but that usually requires a second rendering in Sony Vegas or any other program for video editing to fix the bad-looking fps drop. I suggest you to export in frames instead, especially if you plan to make gifs. Go to Composition > Add to Render Queue, a new panel will appear on the bottom. Click on Lossless and pick PNG Sequence next to Format, then on Not yet specified (output to) and select a destination on your computer. Press Render, wait and it’s done! Again, the rendering times will highly depend on your RAM and composition size. The bigger, the slower; if you play in 4k you should reduce the screenshot size to a common 1920x1080 at least, even smaller after that for the final gif (tumblr has a 3MB limit).

Result: