Sonible smart:gate

Just why would you make the humble noise gate ‘smart’? Sonible asked that question, and this plug‑in answers it!

Sonible were pioneers in the use of AI and machine learning in DAW plug‑ins, and in recent years they’ve delivered an impressive range of processors that make use of that technology. Their latest offering is smart:gate and, as might now be expected from the ‘smart’ series, this looks to apply the benefits of AI technology to the noise gate.

A traditional noise gate essentially opens (to let audio through) and closes (to block it) in response to the incoming signal level — the user can usually set the threshold at which this open/close state changes. Most modern gates, whether hardware or software, also allow the user to control details such as how quickly the gate opens, the minimum time during which the gate remains open and the opacity or floor of the gate when closed (if it’s only part‑closed, it brings the signal level down to that floor). Of course, smart:gate can be used in this fashion — in fact, as a conventional gate, it’s very well specified — but of course this gate is also smart. So let’s see what else it can do...

Context Is Everything

One thing smart:gate offers that’s out of the ordinary is a ‘content aware’ mode. As shown in the main screenshot, Sonible provide a number of common target profiles, and once you’ve selected one of these the plug‑in analyses a short section of the incoming audio. It attempts to identify, within that input signal, a match for the selected target and then configure initial settings — once done, the gate should open only when the level of the selected target exceeds the threshold. Other sound elements within the signal will not trigger the gate. And for many sources, this might be all that’s required.

smart:gate analyses the incoming signal for the information that most closely resembles the chosen target profile — it will only open when it detects those sounds in a signal.

As with all Sonible’s smart plug‑ins, the UI provides plenty of graphical feedback that enables you to visualise exactly what the gate is doing. When working in its content aware mode, this includes identifying exactly which parts of the input signal’s waveform are identified as a match for the selected target, as well as the state of the gate as it changes over time. If you feel the need to fine‑tune the conventional gate controls (attack, threshold, release and hold, for example), this visual feedback is particularly helpful; those controls operate exactly as you’d expect, but only on the selected sound. It’s a very neat trick and...