The look ahead compressor setting can be amongst the most confusing of the audio compressor settings. You’ll also find conflicting information when trying to understand it, with some saying it squashes transients and others saying it preserves them. The truth lies more in the middle, as it depends on the kind of audio you’re feeding the compressor. Time to demystify the look ahead compressor setting – here’s what it is and does, and how it works.
Most importantly, I’ll answer SHOULD you use the look ahead compressor setting when setting up your compressor.
The Look Ahead Compressor Setting
First let’s identify what the look ahead compressor setting is and does.
Rather intuitively based on the name, the look ahead setting previews your audio a few milliseconds ahead of when that audio would play in real time.
This allows the rest of the compressor to anticipate what’s coming next in terms of the peaks and transients of your audio before it happens.
So how does this feature predict the future, at least in terms of the dynamics of your audio?
How Does Look Ahead Work
Look ahead on a compressor operates with a lot of behind the scenes work.
It first internally creates a (silent) duplicate of your track and moves it earlier by the amount of time you set, measured in milliseconds.
The number of milliseconds you can set will vary from compressor to compressor. For instance, the FabFilter Pro-C 2, which I recently called the best audio compressor, allows you to set the look ahead time to anywhere between 0 and 20ms.
For argument’s sake, let’s say we’re setting the look ahead time to 5 milliseconds.
The compressor essentially sidechains the original to the duplicate internally, using the peaks of that duplicate (which is “playing” 5ms earlier) as the threshold.
In other words, if you set the look ahead time on the compressor to 5ms, the rest of your compression settings will be applied to your audio 5 milliseconds in advance.
Look ahead allows the compressor to catch peaks earlier – something which can be useful when you’ve got tracks with fast transients and even the fastest attack setting isn’t catching them the way you want.
Incidentally and as an aside, before the look ahead setting was created on modern compressors, to replicate this effect mixing engineers would manually duplicate their track, mute it, move it earlier on the timeline by 5ms, then sidechain their main track to it.
It’s the exact same effect – the compression is acting on that earlier signal’s threshold, so it’s effectively compressing as if the peaks were 5ms earlier.
Because attack determines when compression begins after the threshold is exceeded, if that threshold is being exceeded 5 milliseconds in advance (in this example), you can also kind of think of look ahead as pre-attack (crazy, I know).
One last thing to mention – the look ahead compressor setting introduces a little latency to function so your DAW will use a bit more CPU to compensate for this. This isn’t an issue if you’ve got a decently modern processor which can handle the demands of mixing, but it’s worth mentioning.
Speaking of which, check out my overview of how to relieve your CPU mixing load to mix more efficiently, at least in terms of your CPU.
Getting back to the topic at hand, you might be wondering why would you/should you use look ahead in compression?
Should You Use Look Ahead in Compression
This is a fair question, because typically when we’re using a compressor, we want to ADD attack, thus delaying compression, not bring it on faster.
After all, the transients of our audio – that initial “punch” which helps a track cut through the mix, are at the front of each peak. They are key in contributing to a desirably punchy mix, so we generally don’t want to remove them.
This is why look ahead generally isn’t necessary in the realm of compression because compressing our audio BEFORE the peaks even hit will squash those transient peaks, killing the expressiveness of our track.
There are a couple practical exceptions to this idea, however, where look ahead can be useful.
Parallel Compression
Sometimes, not always, I like to add some look ahead when I’m using parallel compression.
Also known as “New York compression” because of where this technique was pioneered, this involves duplicating a track in your mix and absolutely flattening the dynamics of that duplicate via a compressor. This is accomplished using a high ratio and low threshold to capture the entire signal and flatten it to essentially a uniform output in terms of volume.
This flattened duplicate (devoid of peaks/dynamics) is then subtly blended in alongside the original version of the track to give it a desirable added fatness and energy in the mix.
When we’ve got tracks with extremely fast transients and peaks, occasionally even the fastest/instant compressor attack time can’t catch those initial transients.
Taking a snare drum for example, the crack of the stick hitting the top of a snare drum has some of the fastest, most immediate transients we can process.
In the case of parallel compression, if we want to absolutely flatten that track then we’ll want to eliminate all trace of that initial “crack” transient to bring it down along with the rest of the audio.
I say “if” because parallel compression isn’t ALWAYS about smothering the transients, but oftentimes when we want that classic “sausage” type waveform we get from extreme compression, we want to keep everything as uniform as possible.
Dialing in some look ahead on our compressor in the context of parallel compression accomplishes that because it anticipates that transient of the snare before it plays and is thus already pulling down the peak by the time it plays, thereby giving us that sausage-like waveform that we might want in this situation.
Limiting
Somewhat in the same vein, in the case of limiting we may also want to achieve that “brickwall” limiting effect on certain tracks. We don’t always need transients in our audio. In EDM some producers like to brickwall limit their kick to cut off the peaks.
Most limiters come with look ahead feature to accomplish the same thing as I just mentioned with parallel compression. In fact it’s a very similar process, the difference being that we’re applying this directly to our main track in this scenario rather than blending in a squashed version.
Regardless, the look ahead once again allows that limiter (which the difference between limiters vs compressors) to get ahead of those peaks and pull them down before they can trigger.
De-Essing
Another instance where we want to attenuate transients would be vocal sibilance – those abrasive and exaggerated high frequencies on certain consonants like “S” and “T” sounds.
Vocal sibilance is one of the rare cases where we don’t want those transients because the transient sounds themselves are harsh and undesirable.
De-essers are specialized multiband compressors which specifically target/compress sibilant rich frequencies to gently attenuate these sibilant transients.
De-essers generally don’t even feature an attack setting for this very reason; instead they typically have a look ahead setting in case the attenuation you need isn’t happening fast enough to get in front of those sibilant transients without it.
Look Ahead Compressor Setting Reviewed
- The look ahead compressor setting internally creates a duplicate of the track you’re compressing and moves it earlier by a few milliseconds, using that earlier duplicate via internal sidechaining as the threshold to compress your track by.
- The result is the compression settings will be applied to your audio by however many milliseconds you set before they normally would.
- This allows you to compress the fastest transients which even an instant attack setting can’t catch, like a kick drum, snare drum, or cymbals.
- Look ahead adds a bit of latency which your CPU compensates for, adding a slight bit more taxation on your CPU.
- Look ahead typically isn’t necessary in most compression as the goal in is oftentimes to maintain transients because they help those tracks cut through and add to a punchy, desirable mix.
- Look ahead is mostly useful when you want to attenuate those initial peaks. A few examples include (some cases of) parallel compression and or limiting, as well as de-essing the unwanted sibilant transients out of vocals.