From their ubiquitous budget Scarlett interfaces to their pro-level Red series, Focusrite is a name synonymous with audio hardware. However, they have recently expanded into the world of AI software with a suite of three plug-ins going by the acronym FAST. But what does it mean? All will become clear.
The three plug-ins currently available are an EQ, a compressor and a third called Reveal that does something clever with levels. This trio is a sensible first move into the world of AI plug-ins for Focusrite because these types of processors tend to be used on most tracks.
However, many production tools can be daunting to someone without years of experience. Compression parameters like threshold and ratio probably don’t much to the casual user. Equally, EQ requires knowledge of frequency bands to get your sounds sitting correctly and blending with other parts.
This is where the FAST plug-ins come in – these are not more emulations of vintage compressors and channel strips. These plug-ins use AI learning to analyse the incoming audio and then apply appropriate settings to the plug-in for you. You can tweak them manually if you want, but that slightly misses the point of them. If you let them do their thing, they really are FAST! Even installation is painless. You authenticate online within the demo version of the plug-in, then input a one-time verification code to get unlimited access to the full version. However, you will need to be online to continue using the plug-ins.
The first plug-in we tried was the compressor plug-in because this tends to go first in our signal chain. We try it on some dynamic vocals because they have a clear transient response for checking the compression’s speed and accuracy.
Because these plug-ins are designed for users that may not be that familiar with a compressor’s functions, it helpfully explains each process that is taking place. When opened for the first time, the plug-in runs through a swift tutorial to explain how to get the AI to analyse the incoming audio. You start by selecting the source that is the closest match to your sound (Vocal High in our case), start your audio and press the learn function. The plug-in then analyses the audio’s dynamics and applies what it thinks is appropriate compression to the sound.
We were pleasantly surprised at how well it worked – and it sounded good too. It’s not trying to be another 1176 clone, but it added a depth to the sound that wasn’t as obvious in the raw audio. The level of compression was pretty good, but we would have gone for a more extreme setting, but that is just preference.
Once the AI has done its thing, you can either leave it alone or explore the Detailed settings. These reveal the usual complement of compressor controls, allowing you to change any setting that you aren’t happy with. It’s also handy if you’re learning how compressors work. Play with the functions to see what difference they make, knowing you can safely return to the AI settings.
The Equaliser works similarly. On opening, it asks you to pick the type of sound source, and it listens to the incoming audio to apply an appropriate curve. This is relatively quick and only takes a few bars to analyse the sound.
Once done, it opens up the typical EQ curve. However, rather than showing it in Hertz and Decibels, it describes frequency as sound characteristics such as sibilance, nasal, and body for vocals or snaps. For drums, descriptors include knock, thump and so on. As with the compressor, this is a fantastic way for aspiring audio engineers to understand the different elements that constitute sounds.
From this view, you can boost or cut the relevant characteristics to enhance the sound based on your preferences. The Detailed view converts the sound descriptors into a 6-band parametric EQ, too. The curve doesn’t change, but it allows you to see which frequencies have changed and by how much.
The final plug-in in the suite is slightly different in that it is essentially a very clever ducking processor. If you’re struggling to balance two sounds, Reveal uses its AI to identify the relative volumes and frequencies to make space in the mix for both.
The initial tutorial this time is a little more involved because it requires sidechaining and, although still pretty straightforward, there are more steps involved. Once set up, Reveal shows a spectral frequency graph of the two signals in real-time to show how the two sound sources are interacting.
The only control here is to adjust how intense the ducking process is on the background track, and even in the Detailed view, there are very few adjustable parameters. In our opinion, this is the weakest of the three plug-ins. The effect it had on the masking track was too subtle and still required us to apply other processing to achieve the desired effect. That isn’t to say it’s bad, because it isn’t. It’s just not as effective as the compressor and EQ.
Fast and furious?
As a foray into AI processing, the Focusrite FAST plug-ins are an impressive first step. The dynamics and frequency analysis in the compressor and equaliser is excellent. Moreover, the ability to fine-tune manually makes them both practical for production and excellent learning tools. Reveal, the ducking plug-in, also works well but is a little too subtle in its effect for it to be equal to its siblings.
These plug-ins are a great introduction to processing for less-experienced engineers. They’re not aimed at people looking for a specific vintage sound from their compressors and EQ, but there are plenty of plug-ins out there that do that brilliantly. However, these are perfect for people that want excellent processors that are easy to use.
- Windows 7+ / MacOS 10.8+
- Intel DualCore i5 CPU
- 4 GB RAM,
- GPU with OpenGL Support
- 64-bit VST, VST3, AU, AAX
- AI-controlled processing
- Compression, equalisation and spectral ducking
- Manual control available
- Clear in-app tutorials
- Rent-to-own payment model