Mar 3, 2011

10 Things I've Learnt in AE That I Wish You to Know Before it's Too Late

Hey guys!
So it's been a year since I started in After Effects, a very fun and interesting year. We've done action scenes, light sabres, we've had failed videos, experimented, experienced tons of crashes, created presets and struggled with controls we didn't understand until now (and feel quite foolish about it).
I've learnt a lot from AE, it's changed my perspective on some things (and slightly affected my social life) but most of all, it's frustrated me tons of times. There's things I wished I knew sooner, some things I'd like to tell you now (so hopefully you won't go insane like I almost did).

"After Effects reminds us to save by crashing."
One of the funniest quotes I've heard because it's terribly true. After Effects (especially on a clean install) crashes a lot. You'll be in the middle of the project with tons of layers, turn the camera slightly and the next thing you know...


Luckily for you AE has an Autosave feature, however it's disabled by default so you'll have to switch it on yourself. Go to Edit > Preferences > Autosave...

This is mine, I doubt the default will look exactly like this.
Anyway, the "Save Every" is obviously how often it saves, and max project versions is how many autosave copies it will make. REMEMBER that Autosave will not overwrite the original file. What happens is that autosave will create a folder like this:
"Adobe After Effects Auto-Save" will be the name of the folder. If your project was called "Fight scene" you'd have "Fight scene autosave 1" etc.
Note, however, that you actually have to create the original file for autosave to work (otherwise it won't have anywhere to save to).
Btw, if you experience a crash while closing AE on an unsaved project, it's because of the autosave feature. You can download an update that fixes that, and many more problems, here.
Or just use Adobe Updater.
Or just go Help > Check for Updates...

That's right, they annoy the hell out of me, and pretty much everyone else who's worked long enough.
Cameras work with a Point of Interest, which is basically where the camera should be looking. When you orbit a camera (using only position), you'll notice it always faces a particular direction. That's because of POI, which kinda works like the anchorPoint.
Has this ever happened to you? You set a keyframe to position at one point, use the Unified Camera tool to turn the camera at about 2 seconds along the line. When you're satisfied, you RAM Preview, then this happens:

The camera comes REALLY CLOSE about midway through the animation!!
We expected this to happen:

There's also the thing of when you scrub the Z position and the camera flips and starts looking backwards.
Well, that's because, like I said before, the camera is annoying. Its relationship with the Point of Interest is kinda hard to understand. In the case of the Z position problem, it can be fixed by keyframing and moving the Point of Interest as well. But the problem above was not solved this way.

This method was solved by a common solution used by most motion graphics people: Null object parenting!
Create a new null object, make it a 3D object, and parent the Camera to the Null.
This process saves a lot, but also takes a little bit of getting used to. (The "camera win" uses a Null)
Working with the camera is, regardless, frustrating. Mostly because working with something 3D on a 2D screen can get very confusing.

I highly recommend you check out Andrew Kramer's tutorial on Video Copilot: Advanced Camera Tips! (That's where I learnt it :-D)

Especially solids. If you remember I talked about this in a previous post.
Solids usually span the entire comp, so if you just create a solid, animate opacity or something for only half the comp, you'll have a lot of excess. This isn't a problem technical-wise, but it is when it comes to future editing, or trying to find specific layers.

"Which layer was the fade-in again?"
Sure, you can find the layer by the name, but don't you think it would be easier to find a layer by just looking at the current time and finding which layer is active. In the image above, I can't immediately tell what I'm looking for... but in this one:
Look how much easier that must be to navigate. And working with 45 layers, it wouldn't be easy finding a layer by name. Sure there's the search box above, don't get all technical with me when you know exactly what I'm talking about.


Resolution let's you turn down the resolution of the rendered view. It reduces the number of pixels that need to be rendered. Setting down the resolution allows RAM Previews to render faster. It's a great tool especially when working with particles and High Definition. Think of it this way, if you had a 1280 x 720 comp and you set the resolution to half, it'd be rendering 640 x 360 :-D and so on and so forth....

"But my computer is powerful! Why the heck would I use resolution?"
Have you ever tried to tweak something like Turbulence in Particular and you'd like to see the effect? It's not fun waiting 10-15 seconds for a 2-second clip to render when it could've taken 5 seconds when set to Half resolution (not exact numbers, just for example's sake), especially when you're making constant, small changes. It's great for making adjustments and fast previewing. 
A consequence of setting down the resolution is that you lose quality. The lower the resolution, the lower the quality, but the faster the rendering. If you're previewing something big (like the motion of a layer), you can set the Resolution to "Quarter" to save some render time and quickly see how the objects are moving, since you don't need to see the exact details of the objects.
Note that if you set the resolution to AUTO, zooming out to 50% will set the resolution to Half, since from so far away you can't see that loss in quality anyway!

Text layers have screwed me over a few times, but mostly it's the Scale-Character size thing.
Have you ever typed text at like size 200 px but wanted it smaller, so you scaled it down... then you have a bunch of them... then you want to animate their scale...
LAYER 1 SCALE: 42.34232
LAYER 2 SCALE: 12.13232

Sure, they're all the same size visually, but this change in scale is screwing us over.
If it's never happened to you, lucky! If it has happened, I've felt your pain many times.
Now, I figured that creating a Null and parenting them all would help (heck, even throw in a valueAtTime expression to offset them a little), but wouldn't it have been easier to just set down the Character size to begin with.
So, head over to the Characters pallet, and set down the font size (which is a small T and a big T next to each other). This way, when it comes to animating scale, we have:

...and now we can just select them all, press S, and animate their scale :-D Then physically (rather than mathematically) offset them if we need to.
Simply put, get used to that.

If you follow Andrew Kramer loyally, you'd know that he's never used a shape layer in his tutorials. He'd always just create a new solid and apply a mask of the shape he wants. Do I lie?
Well, if majority of your techniques and processes are based off his tuts, you probably haven't used shape layers either (unless it was by mistake while creating a mask, which would be undone swiftly.)
Then came the brilliant and very professional Eran Stern (He's got an awesome accent, funny jokes, and very professional client-oriented techniques. I know him from Creative Cow) He's the guy who not only showed me how to use Mocha, but also made me realize that shape layers can be useful (can't remember which tut it was exactly).

Shape layers kinda work like text. You can set a fill color and a stroke. That may not sound so appealing now, but you might need this to create custom particles.

(people refusing to accept the reality would prob say "I can just duplicate the layer, change the color and scale it up", but then again that's a longer process, and that's an extra layer!)
Shape layers also feature very cool path adjustment thingies (I dunno what they're really called). I used them to create the free TV logo (on my Youtube):

These guys are awesome fun. You can make snowflakes from polygons (using Zig Zag + Pucker & Bloat), especially when keyframed! Let's see you do that easily with solids.
It's also great that you can have more than one shape inside a single shape layer, and each will be editable. You can add an effect and it will affect all the shapes within that shape layer. A great tool that's, sadly, neglected!

So take that into consideration the next time you're creating shapes!

Have you ever added something to the Render Queue and clicked Render, only to find the final 1minute video is... 3GB!!!!!!
Well, that's because the Render Queue starts out at Lossless AVI as the preset, which is HUGE, but keeps every pixels EXACTLY as it looked in AE, which is what I tell myself so I don't freak out like the guy on Caps Lock just did.
If you'd like to render out to good tiny sizes, talking about tens of MBs, I'd suggest rendering out to FLV. There's some consequences though:
1. There's lots of compression happening, but relatively speaking there's MINIMAL loss in quality. It's definitely noticable when compared to Lossless AVI, or even the AE version, but when it comes to uploading an originally 720p video to be viewed at 360p on youtube, only the super-technical, video editing veterans would notice the effects of compression. The layman would totally miss it! But when that layman sees the lossless AVI, they wouldn't flip, they'd probably say "ah it looks a bit clearer".
2. FLV tends to squeeze your widescreen footage into a square frame, which will add the black padding. If you check out our Deathmatch video, you'll see exactly what I mean!
You can actually fix this by going into the render settings and choosing "FLV (Same as Source)" when in Media Encoder, or just check "Resize" in the Output Module in the Render Queue and leave the settings as normal.

You can select FLV by clicking on the menu in Output Module and picking it.
Check out this blog post by Andrew Kramer where he discusses various output formats!

Please people, learn at least the basic expressions (like wiggle or loopOut or if). Expressions are little buggas that calculate the value of a property. Most tutorial sites use expressions frequently, but it's also good to go through them yourself (I'm still yet to see a tut where anyone uses loopOut or loopIn, as useful as they are!).
But do learn how to use them! I have a post on how to use expressions, this one, actually!
Chris and Trish Meyer are currently doing a series of posts on expressions that I HIGHLY recommend (they're the masters of AE, and they're married. How awesome is that). Check them out and catch up here!

Adobe's Help system is a great tool, so get over your 'pride' and get in there!
Relatively speaking, of course!
Compared to most help documents that come with software, Adobe's is perhaps the easiest to read and it actually HELPS! (Pun kinda intended). Unlike most Help systems, AE uses easy to understand language, gives links to tutorials explaining whatever it is you're looking at, and it's updated frequently.
If you don't have a constant connection (esp in CS5 where the Help is all online), you can download the help document and have it as a PDF in your system.
Also consider using the search to find tutorials in the Adobe Community website. According to some blog (can't remember which), it's updated frequently as well, and is set up to search popular and quality tutorial sites all over the web like Video Copilot (of course), Creative Cow, Motionworks, Maltaannon and lots more!

You can also follow @adobeAE on Twitter. They post interesting tutorials (some educational, some fun, some extremely informative). 

Mocha for After Effects is a great tracking tool... neigh, an AMAZING tracking tool. Since CS4, Mocha comes by default. It's a planer tracker. I'm sure that didn't explain it.
You know how the tracker in After Effects uses points? Well, Mocha uses planes and AREAS to track, which is awesome because where you'd want to place something on, say, a book, you'd have to do 4-point tracking with AE, where as drawing a rectangle around the book would track it in mocha just like 4-point tracking would.
It's also SUPER ACCURATE, since it's dedicated to tracking you know it'll be hella good at it!! Look for Eran Stern's tutorial on Mocha on creative cow, or this one  or Andrew Kramer's Blast Wave and Magic Tracking tutorials for info on how to use Mocha!

There's lots more stuff I'd love to throw in as well, which I'll be throwing out in various posts as time goes by.
Do you know something that you'd like beginners or other AE users to know before it brings them problems? Do you have an interesting experience with AE? Drop a comment below and get nostalgic (unless it happened a few seconds ago then it'll be HILARIOUS!)


  1. Great post!

    Regarding your crashes: Make sure that you've installed the recent updates. More information here:

    Actually, I think that telling people about the importance of the updates would go well in this post of yours, since they fall into the category of things that many people wish that they'd known about and applied sooner.

    Regarding point #9: That was probably my blog that menioned the frequent updating:

    If you like the @AdobeAE Twitter feed, here is a post with some more ways to follow us (the After Effects team):

    Regarding point #2: There's more information about working with cameras here:

    regarding point #7: This FAQ entry has some additional information and suggestions about compression formats and settings:

  2. Excellent information and a ton of great links, thank you very much.

  3. @Todd I spoke about updates at the end of the AUTOSAVE section, even posted a link :-)

  4. Nice info...still learning AE and all these insights help. Thanks

  5. That's great to hear. Looking forward to seeing your work