
Could this be After Effects CS6?
I’m an After Effects user, and have been since version 6.5, when I taught myself visual effects to help my friend out with his film. I’ve since used it day-in-day-out in my VFX career, and for all my personal VFX and motion graphics projects. I’ve also dabbled in a number of other compositing programs such as Nuke and Fusion and it’s during those experiences that I sometimes feel After Effects is a few steps behind the competition when it comes to the features it offers compositors. I’m not alone. Just visit the After Effects user-to-user forums and check out the After Effects Wishlist thread and you’ll see a ton of AE users all with great ideas for new features they’d like to see implemented.
Don’t get me wrong – After Effects is an amazing program; there’s nothing else out there that is quite as geared up for both compositing and motion graphics, and it’s layer-based approach to working gives editorial-level control over timing. There are people who’d like to see AE become a node-based compositor. Not me – I think that would make it a completely different program, which would alienate large numbers of the userbase. The features I’m suggesting here I’ve tried to keep within the realms of reality, such that if they were implemented in CS6 or CS7 they would maintain backwards compatibility with older projects.
Adobe seem to be making a point of really listening to their users at the moment either through their various Twitter and Facebook accounts, but it would be remiss of me not to mention Adobe’s feature request form as the official way of putting your ideas forward. I’ve made an official feature request for all of the features I’m describing here, so fingers crossed that one or more of them will find their way into After Effects over the next couple of versions.
So here are numbers 1-to-5 of my top 10 features After Effects needs!
1) Per-Vertex Feathering of Masks
What?
After Effects CS5 saw the introduction of the Roto Brush, a very cool tool for giving you a head-start on the painful journey of rotoscoping. From my limited experience of it, it’s certainly very impressive tech, but this is a roto improvement at a more basic level.
A long-time feature of Nuke and Fusion, the ability to set the amount of feathering per vertex is a must for the next version of After Effects.
Why?
Sometimes you want to have a mask that’s sharp on one side and heavily feathered on the other. A recent example saw us adding some shadowing to an actor’s face – the profile of his face needed a sharp edge, but the soft shadow required a lot of falloff. Per-vertex feathering would give this control using a single mask. Yes, you can buy RE:Vision Effects’ PV Feather plugin but I strongly feel this should be built into AE’s native masks. Besides the plugin doesn’t work on Adjustment layers, which is where a feature like this would be very useful.
How?
CTRL+dragging on a vertex point could be the way to unlock the feather hull, with a per-mask interpolation type (linear, ease-in, ease-out, etc.). The existing Feather control would feather the mask as a whole, with any per-vertex feathering compounding the effect.
2) Flexible Processing Per-Layer
What?
Allow masks and effects to be mixed-up, and grouped in layers (This and the next two features are all companions of sorts. If they could all be implemented it would revolutionise the program, whilst keeping it very much After Effects!).
Why?
Currently After Effects forces an order of operations on its layer processes: Masks, effects, transform, Layer Styles. This is quite restrictive and in many cases forces precomposing to change the order of operations. Say, for example you’ve got a Fractal Noise effect on a layer that you want to blur heavily but cut out with a sharp mask. Currently, as masks are processed first, any blur effect will blur the masked edge so you’d have to precomp your Fractal Noise & blur and mask the precomp layer. Annoying! Solution? Drag the mask below the effects!
At the moment, of course, After Effects is very neat & tidy – you know exactly where everything will be within a layer. If this restriction were to be relaxed then you’d lose that order, but just think of the possibilities!
How?
Working the same as it does now, but allowing masks to be dragged from their default position to inbetween or below any effects in the layer. Things could be kept neat by introducing groups, which serve no other purpose than for organisational purposes, along with a master on/off switch for that group. And to top it all, provide a mini-UI for visualising & editing the processing order of effects & masks on the layer (kind of like the flowchart view but for layer processes). I’m quite happy for Transform to remain at the bottom of the processing stack, after all, there’s the Transform effect if I want to change the order (though while we’re talking about it, how about making the Transform effect 32-bit, and offer a few alternate filtering modes..?).
3) Mask Grouping and Transforms
What?
Expanding on the previous feature, allow multiple masks to be grouped together logically, including nested groups, and give each mask (and group) its own optional transform control, similarly to how Shape Layers work.
Why?
By adding transform parameters that would at least give us the option of moving a mask as a whole via expressions. Grouping masks will allow simple and identical transforms of masks in fewer steps, and keep things organised during heavy roto sessions.
How?
Masks are already grouped together by default, however this group should be able to be renamed, and other groups created separate to, or nested within it. An “Add” menu (like Shape Layers) would be used to create new groups, and add transform controls. Within the layer you’d drag and drop masks into their respective groups. Also, masks shouldn’t necessarily need to be part of a group.
4) Layer Groups
What?
Group one or more layers together for organisational purposes, and for pre-processing without having to precompose.
Why?
Photoshop has provided layer groups for master manipulation of multiple layers for a few versions now, and being able to group a series of contiguous layers within an AE comp would be, in my humble opinion, one of the most powerful new features Adobe could offer After Effects users. A group of layers would become a super-layer, essentially a nested precomp within the main comp. Let’s look at the potential benefits:
- No need to leave the comp to edit effects and timings of the grouped layers
- They would be processed like precomps (i.e. before non-grouped layers) so their output could drive compound effects in other layers in the same comp. In fact, by grouping a single layer with, for example, a Fractal Noise effect, thus forcing its pre-processing, this could mean the end of precomposing altogether for driving compound effects!
- A group transform would mean simple move, scale & rotate of multiple layers without parenting or precomposing
- Collapse the group to keep things neat and tidy in your master comp, and turn all layers in the group on and off with a single swich
Of course I’m not proposing this as a replacement for precomps, as they’re useful for many things including reusing elements.
How?
Create a new group using Layer->New->Layer Group, then drag the layers you want into the group in the timeline. Grouped layers would be indented, and their label colour would also feature the label colour of the group. As they would behave like nested precomps, a layer group would share all of the switches and modes of regular layers. How far this could be taken (adding effects to groups, nesting groups) is up to Adobe!
5) Native Multi-channel/OpenEXR Integration
What?
Use from EXR & Photoshop files, and create, new channels in comps for use further down the line.
Why?
Adobe currently provide fnord’s EXtractoR plugin for accessing OpenEXR channels. Whilst I applaud Adobe for bringing this functionality to AE users as part of the standard install, I wish they would develop something completely native to the program that goes a bit deeper. The problem isn’t the speed, but more the execution. For starters I’d like to see support for more than the standard 4 channels (RGBA) throughout the application rather than solely at layer-only. Creating new channels from layers is also a must when dealing with multi-channel projects. For example, you might have a comp with multiple layers making up single matte, but each layer deals with something separate in the shot. Later on you might want to reuse those layers in another comp, which is where such a feature would work wonders.
How?
By default comps should be RGBA only but if a multi-channel layer is inserted into a comp there should be the option to add one or more of those channels to the comp’s output via a Channel Manager dialog. Similarly for any of the standard RGBA channels from regular layers, which could be turned into new, custom output channels. If that comp then becomes a layer in another comp, its output channels will be available to effects applied to it. Extra channels should be accessed via a modified version of the Shift Channels effect to keep things properly integrated.
Here’s my mocked up all-new Shift Channels effect!
Notice the new Channel Layer dropdown, which lets you choose a specific render pass (left) or channel group, like RGBA or HSL (right).
Coming soon, part two, featuring suggested improvements to the 3D environment, colour correction and more.