AE Tutorial

dry version: Adobe After Effects is a compositing program, which, simply, means it is for assembling various picture components, including bitmap and vector images, and video sequences. It offers tremendous control over complex layering and matting, in a simple, powerful interface, as well as a plug-in architecture for access to tons of third-party filters. It outputs Quicktime movies, and still files and sequences in a variety of formats, including layered Photoshop compositions.

soaking wet version: Adobe After Effects is truly the bomb. Many people feel they owe their lives and/or sanity to it; it's just the ticket many of us stray creatives needed to thrive. I am SOOOO in love with this program; I have made complete videos using only self-generated imagery, no footage or cameras used. You tinker, preview, make a laundry list, make your fixes, and repeat as necessary until you have sidled right up to exactly what you want: perfect.

It is the perfect combination of intuitive and powerful. The basic metaphor can be clearly explained over the telephone. You can download the free demo version, comes with lots and lots of tutorials/help, etc. With a little diligence, you find yourself up to speed surprisingly soon, and find yourself spontaneously getting ideas already formatted in AE, and next thing you know, it's done.

I recommend very quickly browsing this document all the way to the end, just skim it all and then toss it down; start again when you're ready. It's not meant to be exhaustive, but only sufficient to get you safely through the initial bumps until such time as you are well bitten. Throughout, a leading C will be used to indicate the Command/Apple key, as in C/O to indicate the key combination Command/O. Likewise, an O will be used for the Option key, an S for the Shift key, and C+ for the Control key. This is for Macintosh; translation to Windows is the same as fr any other program, subsituting Control for Command, and Alt for Option, more or less, not my world. I place strong emphasis on using key commands rather than menus; it's much faster. If you go to a menu and see a key command, come off the menu and use the command instead, and you will find you use pick them up quickly. Feedback is always appreciated; please let me know if anything needs clarification, fixing, etc.

I’ve tried to keep this simple and organized, but complete enough, too, so I'm sorry if the rhythm’s clunky. This was last updated for version 5, which should hold you through version 6.5, which for beginners is the version I would recommend; versions 7 and CS3 utilize a powerful new Graph Editor in place of the more immediate curves in these versions*. It shouldn’t take much practice to completely absorb everything in this tutorial. Just start and continue, and wait for the addiction to kick in, and you won’t believe how quickly you’ll be up to speed; this program figgin rocks. It might be daunting at first, but it will soon reveal itself to be pure elegance, racing skis for your brain. If you are working from a cracked copy, please fall in love with it then come legit and buy it; the AE team is small and really cool, and they worked hard to make that insanely fun new toy you now adore. Eventually you’ll need to buy the Production Bundle version; it has a few extra features that are really wicked. There are also loads of third-party filters, pretty out of control actually.

*you can still use it for these versions; please access the Help menu to find the button required to access the Graph Editor, and also the button which selects menu type. Note that you have to have a parameter selected for its curves to show up.

First: A basic RGB document has three 8-bit (2 to the 8th, or 256 shades of grey; the range of values is given as 0-255) channels, which correspond to Red, Green, and Blue. 24-bit color (256x256x256) gives a color depth somewhere in the millions of colors. It takes a while to learn to work intuitively in RGB color space (like, to make it more yellow you take out blue), but it’ll come. You can also add a fourth channel, called an alpha channel (more details follow near the end); pixels in this channel control the transparency values of the corresponding pixels in the RGB channels (RGB+Alpha is also called Millions+ or 32-bit). By definition, a bitmap file is rectangular, so for anything with any other shape, you'll need to accommodate the transparent area around it. So, every layer/image is just a matrix of solid squares (like tiles) with 6 numbers attached: 3 for color and one for transparency, and one each for the x and y address. Get far enough away, it’s a picture, but when it comes time to try sort out problems, you’ve got one set of numbers that you’re trying to modify into better ones, and thinking of it in these terms will come in handy. Means nothing to you now, but you’ll see.

If no projects are open, C/N will create a new project, which is the master file. Only one project can be open at a time. Normally, C/N creates a new Composition, the basic workspace-unit of AE. You can also create folders in the Project Window with C/S/N. To retitle anything, select it and press the Return key to highlight it, and hit Return again (or click away) to accept the changes. C/I will import one image or file into the project window. C/O/I will allow you to continue to import multiple files until you click "Done". If you select a layered Photoshop composition, you will be asked to choose a layer, or import the entire file merged (flattened) as a single layer. There’s also an option to import an entire layered Pshop or Illustrator file as its own composition, with all the layers retained; one advantage is that alignment and transfer modes are preserved. You can import other projects into projects, but there is no connection between the imported and original, i.e.: if you edit the original, after it was imported, the imported version doesn’t update, and vice versa.

To modify these settings, use C/K with the composition open and active, or when selected in the project window. Each composition has two windows, the Time Layout Window (TLW, aka Timeline), where you work on the structure of the layers, and Composition (Comp) Window, where you can see the visual result. Along the top of the TLW you’ll see two timelines, each with a blue Current Time Marker (CTM). The bottom represents the portion of the comp currently displayed; the top represents the entire composition, with a white center section corresponding to the bottom portion, and grey areas corresponding to areas not currently displayed. Clicking anywhere on either moves the CTM to that point.

If the Tool Palette isn’t visible, Window>Palette>Tools. The topmost layer in the TLW is the frontmost. Each gets a corresponding line in the TLW; the default length for stills can be set in File>Preferences>Import, while footage files show their exact length, i.e. a 2-sec Quicktime will display a 2-sec bar. The head and tail can be trimmed by dragging; O/[ and O/] trim the layer to the CTM.

This will create a new keyframe at this point in time, corresponding to that value. As long as the CTM is on that keyframe (S-move it to snap to the nearest point), any changes to that parameter will change that keyframe; otherwise, if you make a change to a parameter, and at least one keyframe has been set, a new one will be created at the CTM, whether you want one or not. If no keyframes have been set for a parameter, changes do not create keyframes.

In the Comp Window, a selected layer will display handles; the arrow tool can resize the layer by moving these (use S on a corner to preserve the aspect ratio), or move it if you grab it inside the boundaries; there’s also a Rotate tool.

Note that it has the same keys as the one you duped. Any modifications to the original Composition, Small1, will be updated in A, B, & C.

See what’s going on? Duplicates in the TLW are aliases to the original; duplicates in the Project window create independent copies.

Note that the default black in the backgrounds of the comps is actually transparent (C/S/B lets you set the background color of the active comp). If you really need a solid, opaque background, go to time 0:00 and create a solid (C/Y; C/S/Y to modify) and drag it to the bottom of the stack.

Now you understand the next big building block, compositions. It will become very apparent how useful this is. An example would be a word made of five letters: in one comp you have them take 2 sec. to spin into position, where they hold for 3 sec; you could then place that comp in another, and do nothing to it until 2 sec, at which point it could, as one unit, uniformly decrease in scale, while a copy blurs and grows.

Note that the image now only shows within the boundary of the solid, and that the solid’s eye icon is off.

An "Effects Control Window" (ECW) will open; this is available anytime by selecting a layer and hitting C/S/T.

You can also delete effects in the ECW, or change their order. All Effects are heavily keyframeable for most of their parameters. That’s building blocks 3 and 4, and we’re nearly finished with the essentials.

Every layer can have quite a few masks, each a simple path (like Illustrator), which in combination determine the transparency/matting for the layer.

The Pen Tool lets you draw a single freeform "vector" path, a la Illustrator curves; it takes a little time to get used to, but it's actually fairly simple. There are points, and handles which control the curves between the points. The points can have no handles (corner point), two linked handles (smooth point), or two handles moving independently of each other (it used to be called a cusp, but Adobe changed that, but I still call it that). You can add or delete points, change the points from one kind to another. The Pen tool only makes new points, the arrow tool selects and moves them. If you have one selected, holding the Command key will toggle the other. If you are new to vectors, then just practice a lot, and you'll see it's that simple, but it takes a little time until it flows.

You can key changes in this mask shape as well, and change its Feather Values. You can also combine multiple masks on the same layer, with various modes such as Add, Subtract, Intersect, in descending order from the top mask down. If you’ll be animating the paths over time, it’s best to just draw one, and modify it at various points in time, rather than redraw it or add new points, otherwise the interpolation can be unpredictable, as you’ll see as you work with it. You can also do basic Transform moves on any mask (or selected portion/keys), either double-click it or hit C/T, then move the handles or rotate/scale etc. Vector maps are the basis of rotoscoping, i.e. groups of paths adjusted to fit moving footage to isolate objects for compositing.

In the Layer Window, selecting Anchor Point Path (via the top right options arrow) allows you to move the Anchor Point; this is the effective center of the layer, from which all scaling, rotation, and position operations proceed. If you already have Position keys set when you move the anchor point, you may have to adjust or reset them; you'll see. You can also move the Anchor Point in the Layer Window; pull the top-left menu to Anchor Point path; you can also adjust the actual path here from the default linear interpolation.

Limit your overhead and you'll work faster: while AE does its best to keep up with you, every change requires AE to rebuild the whole screen, and impatience gets very stressful; try to avoid creating it. Usually you try to have only one window open, zoomed out so it’s as small as you can stand, and on Draft resolution except as necessary.

In the TLW: The Lock icon makes the layer unavailable for selection or modification. There's a column to choose Draft or Best quality.

The M turns Motion Blur on or off; this adds convincing motion blur to any moving layer, but is processor intensive. It only works if the M button at the top of the comp is on; usually best to turn it on only as necessary or until it's needed for rendering

With any layer selected, (S-select to add layers to the selection that are adjacent to each other in the stack, or C-select to add non-adjacent layers...try it, you'll see the difference), the following shortcuts bring up the corresponding parameter:
P Position
S Scale
R Rotation
T Transparency/Opacity
A Anchor Point
M Mask
F Feather
E Effects
U the "Uber" key- all animated parameters (yes!)
UU will show you all parameters for the selected layer that have been edited (yes yes!)
,/. will zoom you out/in without increasing the size of the pane onscreen
There’s others, but that’s enough to start. If you have several layers selected, hit S, then click on and modify any of their scale numbers; all will update to reflect that value. Same with P, T, etc. Hitting S then S/P gives you scale AND position, etc.

You can drag the layer handles to get an irregular aspect; this will split the scale into x/y values. Or you can hit C/S/S, pull the constrain menu to None, then modify the values.

Have a ball with the filters, but it can get boring to wait for them to update every single time. The Effect checkboxes (a little f) in the TLW turns on or off all effects for the layer, plus there's individual checkboxes there for each effect.

There is a Path Text filter, which is fairly involved, but you can do some unique text animation with it. It's sometimes easier to edit the text elsewhere, including returns, then paste it in. (There is a new type tool, which isn't covered here, yet: it is fantastic. For now just realize that your type has to be selected for Character pallette (Window > Character) changes to affect it)

Adjustment layers apply their assigned Effects to all layers under them, based upon the Adjustment layer's alpha/transparency info, i.e. where the adjustment layer is 50% transparent, the effect will be applied at 50% strength. Usually you will just use a full-size solid. To make a layer into an ADJ layer, click the half-black circle in the switches panel; it disappears, but remains active. Clicking the visibility (eyeball) off disables the effect. You can animate the transparency of this layer; this doesn't so much ramp up the effect as dissolve from no effect to full on, not quite the same (i.e. Blurs won't go from 0 to 20, the 20 blur will just fade in over the uneffected layer): experiment. Masks and feathering can be used to control/localize the effects.

Clicking the stopwatch once sets the first keyframe for that parameter, at the CTM; clicking it again removes all keyframes. Clicking the parameter name selects all of its keys. If you select any keyframe(s), and hit C/S/K, you’ll get Keyframe Interpolation options; Linear means straight lines, while the various Bezier options can give you smoother transitions. For smooth ease-in and ease-out, you need "S" curves; experiment. Hold means the value freezes until the next key frame. Once keyframes are established, an arrow appears next to the parameter, giving you access to the curves; experiment to get a feel. If you uncheck the tiny little box below a key (the ones at the ends don’t have them), it will "rove" in time so as to smooth the animation. You can also manually edit the shape of the Position path in the Comp window. If you select several P keys, you’ll see them become available (in Prefs you can set how many show); the selected keys can be dragged as one group. With Bezier interpolation, sometimes AE draws really weird curves as it tries to guess how to connect points; if your layer doesn't settle but loops past wildly, this is your problem; make the intended settled key a Hold key, and/or adjust the handle on the path. Notice the little dots on the path; the closer they are, the slower the layer is moving in that section. Hint: if it's stopped, the velocity is flat at zero.

Generally, for creating files with an embedded alpha channel, you’ll need to output in "Straight Color" (which you set when you render), which is best explained by example: an image of wispy smoke would look solid white in the RGB channels, just a big, blocky blob, because all of the color values are pure white, no shading. However, if you looked at the alpha channel, you’d see the transparency information, because the alpha channel would look like smoke: when properly composited you’ll get a nice white veil. If you used that same alpha channel for a pure red layer, you’d get red smoke. If it was output "Premultiplied with black", all of the transparent areas and edges would have a black ghosting, and in the main channel, instead of solid red you'd see it fade to black at the empty areas. It’s possible to compensate for this in Interpret Footage, but ultimately Straight Color works more consistently. It boggles my mind how many professionals don't understand it.

It takes a fair amount of tweaking sometimes to get convincing movement. Try to get things to bounce and jiggle, and slip and snap. Look at old cartoons, those guys were geniuses, with nothing more than pencil tests. One trick: move your rotation keys a frame or two after the position keys, so that you get a lag. The word animate means to give it life; like many things, only the mistakes show, but if you get the nuances right (jiggle, anticipation, reaction, etc.), then viewers will stay sucked in.

B sets the beginning of the Work Area at the CTM, N sets the end. Keypad-0 gives a preview of the work area; S/keypad-0 gives a half-frame-rate preview (or whatever playback rate you specify in the playback controls menu).

C/M is Make Movie, which will lead you to the Render Queue, where you choose where and how your movie will be saved. When you output, there’s an option to only render the work area. File>Preferences>Render/Output Templates allows you to set up custom modules for exporting movies, and specify defaults, so you don’t have to correctly check 15 buttons everytime you want a quick preview, etc. Animation Compression at Best Settings, leaving the Keyframes field blank, is lossless, and good for most instances of final output; for smooth playback of previews you can use Cinepak compression at lower quality settings. Note the "Scale" and "Crop" option in Output Settings.

When it comes time to output to actual tape, requirements vary, so you’ll have to ask. Find a friend with access, because otherwise it’s very expensive. MiniDV systems look pretty decent for some things and are cheaper to get into, but it is NOT broadcast quality for most uses. Anymore, you can probably just put it on a data DVD and hand it off that way.

If you make a transparent 320x240 Photoshop document and create a 100x75 type layer, AE will import the layer at the smaller size and ignore the space around it; sometimes blurs and effects will not extend past this edge, though in most cases that seems to have been fixed. A common workaround is to set a brush to 1% transparency and draw small marks at the top left and bottom right corners; AE reads these as valid data and sets the layer boundaries wide enough to accommodate them. Usually they’ll present no problem, but if necessary the mask can be trimmed in AE to exclude them. Illustrator files have the same problem, but you can cure them by making a larger rectangle in AI and turning it into Crop Marks; no masking is required. Easier is just to pre-compose it in a larger box; in actual workflow, things that come from clients, like logos, get changed, and having it precomposed means you only swap the source, without having to go through and re-do all of your position/scale keys, etc. Not all characteristics of Illustrator documents import; experiment. It’s also usually best (though not necessary) to turn all Illustrator type into Outlines before importing, though you should also save the original type in case you need to re-edit later (either as a copy or a turned off layer). C/E lets you update any file in its original application; if you just modify it without using this command, AE doesn’t update to reflect the changes until the project is closed and re-opened.

Do not retitle or move files in Finder once they’ve been imported into AE, or it might lose the links. If this happens, double-clicking on the name in the Project window will allow you to navigate to restore (or replace) them. You can also simply replace a file with C/H; anywhere it was used will now use the new file.

C/S/D splits a layer at the CTM, keeping all keys intact on each portion. One possible use: if a lot of filters or unusual matting takes place only on one small part of a longer layer, rendering may go faster if you split the layer and remove the unused complexities from the long part. You can also use this to change a section of a layer’s position in the stack there.

Hitting the Asterisk (*) key on the numberpad leaves a marker on any selected layer at the CTM. C/clicking will delete it. Double-clicking a marker lets you assign a name. If there is a layer with Audio data (has to be saved as a Quicktime movie), and the Audio icon is on, hitting the period key on the numberpad will give you an audio preview. You can set the length of this preview in Preferences. If you click the arrow next to Waveform (or hit the L key twice quickly) in the TLW layer, you can see it, which is useful for really locking onto peaks. These work great together for synching with sound; just select a layer (or create a solid), move the CTM where you want it, hit period, then tap the asterisk at every beat, word, crash, whatever. Adjust the markers, then use them as guides to place keys on other layers. Sometimes I make a separate layer for the kick drums, another for the snares, etc.

You can Time Stretch or Time Remap layers (under the Layer Menu); on footage (i.e. Quicktimes) this just doubles or drops frames and can look lousy, but on comps made of animated elements it works fine.

Get in the habit of organizing your projects, and giving all layers/comps clear and consistent names. I usually put a bullet (O/8) in front of any layer that, for whichever reason, stays turned off.

In normal video (as opposed to computer monitors), the image is painted to the TV screen in alternating (or "interlaced") scanlines, with every other line put in on the first pass, then the remaining lines on the second pass about 1/60th second later. Each pass is called a field; two fields make up a frame. Most video footage will need to be set to Lower Field First when it’s imported; AE does an okay job of faking up still frames. It's always a judgement call; you have to see it on a real NTSC monitor to be sure, and that's not usually at all convenient. If you are using lots of motion, it can look better rendered as fields. Sometimes you can get away without field-rendering, but sometimes that makes things look even more jagged.

Standard video output in the US is called NTSC; it is 29.97 fps, and the frames are 720x486 pixels (or 720x480 for DV/DVD). However, unlike the Macintosh environment, the pixels are slightly taller than they are wide, by like 10:9. If you create a document in Photoshop at 720x486 (never do this) and import it, with its pixels set to D-1 pixel aspect, it will be squished and appear narrow when it is output, though it looked fine in AE. Either set it to square pixels (File>Interpret Footage), or create the originals at 720x540 (if you have a lot of tight thin horizontal lines, you'll have to go with 640x486; you'll be seeing all kinds of new problems when you see your work on an actual TV monitor), and scale it in AE to 720x486. Interpret Footage also gives several other options; experiment. Yes, this is a giant pain in the ass, but it'll be a while before you have to be an expert; enjoy your ignorance alibi while it lasts.

In the Production Bundle, there are Audio filters, pretty well neglected as far as I can tell. They are MAD fun, and often they introduce all kinds of odd artifacts. It's a unique means of producing/modifying audio, hardly general purpose, but you can get into very unusual results.

bonus: my rotoscoping tutorial
a great way to get your entry-level foot in the door.

Hmmph. That’s it. There’s plenty (like parenting, motion blur, the extensive and powerful Type tool, layer-controlled effects like displacement mapping, collapse transformations switch, 3D/cameras/lighting. tons and tons of filters waiting to be crossbred into weird mutant results...here kitty kitty) left over for you to figure out on your own, but this should be a sufficient head start. Rove around, click on everything, go snooping. The tools are simple, but the combinations can get really out of hand, and take you far. Make a great mess!!

Several excellent books are available for deeper digging, notably "After Effects Hands on Training (H.O.T.)", which is the book I'm using for the AE class I teach at Villa Julie College. Anything by Trish Meyers (three volumes available) is going to be a treasure trove of juicy tech tips, but I wouldn't recommend them for first timers; they're just a dense jumble. Look on Amazon for cheap used books, chew 'em right up.

The way I handle books is as follows: sit someplace comfortable, away from the computer, and first flip through the entire book, just glance at each page once, lingering here and there if you want, but otherwise keep moving. This will dissolve the initial overwhelm somewhat, and maybe even displace it with excitement. Then, start at the beginning; anything not immediately intuitively obvious (most of it will be), anything you want to be sure to remember (such as key commands), just hit with a highlighter and keep moving; don't stress details just now. Whenever you want, take the book over to the computer and work with the highlighted areas. After you finish the book, assess your development and see where you think you need more work. Flip back through the book and look at all the highlighted areas, see which ones are still fuzzy. Eventually, you will put the book on the shelf confident that you have absorbed it. Seems like a lot of work, but me, I really don't like being a beginner, so this is an efficient way to get way past that. And I'm lazy, so I well understand how a little work now saves a lot of work later.

Brian Maffitt’s Total AE DVD series is legendary for creating mastery. I always recommend "Drawing on the Right Side of the Brain", by Betty Edwards (commonly available); it can teach you to actually draw, which is one of the best tools any digital artist could acquire. How can you hope to make visuals if you don't even know how to look?

Loads of online resources are avaialble, such as the extensive tutorials on creativecow, and the subscription-based resources on lynda.com.

Take notes. Type and organize your notes (and leave a section at the end for questions you don't get, then figure them out). Trust me, this is a huge help, plus you can take it with you as a cheat sheet on jobs.

Welcome to the world of motion. An excellent exercise is to select a short (15 seconds or so) piece of music, perhaps even edit it so you have an intro, middle, and end, and try to make some abstract visuals to fit it. You can use simple solids, and focus on the basic transforms: rotation, position, scale, opacity. The buzzwords I always emphasize are attention (distraction) management, dynamic disequilibrium, and counterpoint; try to learn about leading the eye around, creating a lack/desire and then answering it, and then moving again before it gets stale, which usually takes all of a couple seconds. Start watching how things move, how the quality of their motion gives them character. Animate: to make alive.

If you are in the DC/Baltimore/Philly area (NYC also occasionally possible), I give lessons; two or three hours is usually a great start, and if you've already been working with this tutorial a bit you'll be that much further down the path.

DJ - Computer Graphics - Paintings - Miscellaneous