2019 doesn’t sound like it was that long ago, but a lot has happened in seven years. In fact, looking at photos from 2019 is slightly disorientating as you realise how much has changed since then. It was before we had a global pandemic, before AI was making regular headlines, and when the whole world was waiting to watch the final series of Game of Thrones.

Seven years ago, starting in May 2019, I posted a five-part series on advanced keying in After Effects. It was something of a brain-dump, an in-depth dive into chromakey where I looked at everything from types of keyers, how keylight works, how I work, before demonstrating a range of techniques for dealing with chromakey shots and compositing.

At the time, it was the longest series I had made – this was before I posted over 18 articles (and 100,000 words) on After Effects and Performance, and before I made over 12 hours of videos on colour management. In fact the only bits in the chromakey series that aren’t correct are some comments I make about RED files and colourspaces. Having spent a long time unravelling the mysteries of colour management since then, it’s interesting to look back seven years ago and see a snapshot of myself dealing with topics that I was trying to understand.

Looking back at that series on chromakey- there’s actually 6 videos across 5 parts, because they got so long – I’m relieved that most of the information in them is still relevant, and they remain a valuable resource for anyone looking to learn more about chromakey.

But even old dogs can learn new tricks, and so when I learned about a technique called “additive keying”, I thought it would make sense to post an After Effects tutorial about it, to follow on from the original series.

There are lots of problems with chromakey, usually because greenscreens are rarely perfectly lit, and even if they are, fine details such as hair are still challenging.

Additive keying is a technique to recover these details at the compositing stage, and when you look at before/after comparisons you’ll be kicking yourself that you didn’t hear about it sooner.

I’m not 100% certain why it’s called “additive keying”, although it could have become shorthand from the more correct terms “additional keyers”, because the process uses two extra layers in compositing. However although one of them uses the add blending mode, the other uses multiply – so technically, only one small part of the process is really “additive” and neither of them use a chromakeyer.

It’s probably worth clarifying that the additive keying process is something that is done at the compositing stage, and to be pedantic it’s not actually keying. When I was planning this tutorial and was thinking about what footage to use as a demonstration, I realised that the last keying project I worked on didn’t involve any compositing. I was contracted to key a whole bunch of people and supply them as cards to be used in a realtime 3D engine. For that project, the additive process I demonstrate here wouldn’t work (as it doesn’t affect the alpha channel, and so I had to use regular methods to ensure I recovered as much hair and fine detail as possible using Keylight.

If you’ve found this video interesting then here’s a whole bunch of links to some other related articles.

Firstly, here’s a shortcut to the original series from 2019:

Part 1: Introduction and types of keyers

Part 2: pre-processing

Part 3: Keylight

Part 4: cleanup and post-processing

Part 5: compositing

In the video above, I demonstrate a simple grain-managed workflow. For more details on that process, check out this article.

The additive keying process I outline works best in a linear workspace. It uses the divide and multiply blending modes, and although the blending modes themselves will work in a regular rec709 project, you won’t get an ideal, colour accurate result unless your workspace is linear. The grain managed workflow that I use will generate negative numbers, and to deal with them we need to be using the 32bit float project mode. So the keying process I demonstrate above, like all the projects I work on now, is a 32bit, ACEScg project. This all comes under the topic of colour management, and I’ve posted 26 videos on that very topic. However I don’t expect anyone to sit through them all from the start, so there’s probably two that are worth linking to here.

The final video in the series is an express overview of Colour Management, and falls into the “everything you need to know” category.

But for those of you who are still working with regular rec709 projects, Part 20 in the series looks at the benefits of moving away from rec709, and using 32bit linear projects.

And if you’ve found these topics helpful, then I recently collated together a bunch of my older articles and tutorials in one spot, so you can check them out as well.