Professional Analog Mastering.
Professional Analog Mastering.
222K+ subscribers
Modern limiters, or brick wall limiters utilize lookahead.
When enabled, latency is introduced - this allows the processor to measure the peaks of the signal before processing.
With this info, it knows when to attenuate the signal without the need for an attack time - it doesn’t clip the peak like a clipper, it reduces the gain of the peak right when that peak occurs. It’s not reacting like a typical compressor, it's already mapped out how it will alter the signal.
Most limiters allow for amplifying the signal into the limiter, which is the only real benefit of the processor. The quieter parts are amplified, causing a louder sound which is generally preferred by people, and reducing masking to quieter details.
Lookahead-guided attenuation causes less harmonic distortion than clipping since the wave is shaped to a lesser extent, but it affects transients in a distinctly unpleasant way.
You may have noticed that limiters reduce the impact of transients, causing a blurred sound. Some limiters are better sounding than others, but the algorithm they use is really what changes the sound.
The release time and shape, the attack shape, if transients are expanded post attenuation, if oversampling increases the accuracy of amplitude detection if the channels are attenuated in tandem or processed independently, and so on, all affect a limiter’s sound.
But, here’s the point.
If all we want is to amplify quieter details, there are much better ways to do this than by using lookahead attenuation on peaks.
If your goal is to create a dense, detail rich, and nuanced sound, but without messing up transients, then stop limiting. Limiting will always affect the transients by altering the amplitude of peaks - that’s what they’re designed to do.
That makes sense, but aside from clipping, what’s the alternative?
Mastering with Various Dynamics ProcessorsAll the benefits of limiting can be accomplished with parallel compression, maximization, and careful monitoring of the peaks.
Let’s look at this graph - the x-axis is the input, and the y-axis is the output.
If a limiter pushes the input until the loudest parts hit 0dB or a set ceiling, then I want to try pushing the input, while leaving the loudest parts alone.
That is, to take the input of every aspect of the signal that’s below say -10dB, and amplify it, while leaving everything above -10dB alone.
If I can continue to push up these quieter details while leaving the peaks alone, I can achieve the same loudness as limiting, but without the need for attenuating peaks.
The most common method for this is parallel compression.
If I use this Pro-C2 plugin and compress heavily, you’ll notice the peaks are attenuated; however, if I blend in the dry signal so that I achieve equal parts wet and dry notice how the peaks are unaffected, while quieter details are amplified.
I can vary what gets amplified and to what extent by adjusting the knee, the threshold, the ratio, and other parameters like the algorithm or compressor type.
Although the peaks will be affected, it's to only half of the signal since half of the output is the original, unprocessed signal.
It’s not perfect, but it allows for much less noticeable attenuation to peaks while introducing all of the detail and loudness we’d achieve from limiting.
Furthermore, since we don’t need to worry about clipping and subsequently don’t need an instant attack, we can vary the attack time and release to optimize transient retention.
Also, if you’re worried about harmonic distortion, bus or mastering algorithms work well at reducing that - or you could try different models out to find some distortion that suits the track.
With other plugins you may need to combine the wet/dry and the output to achieve a similar effect, but I recommend this plugin if you plan to use parallel compression for mastering.
I’ll cover some different methods in a moment, but for now, let’s compare limiting and parallel compression. I’ll set both to the same loudness so that we can more easily tell the difference.
Watch the video to learn more >
A maximizer like the Oxford inflator pushes quieter signals upward before slowly reducing the amplitude to reach 0dB.
Some people call this a soft clipper, but looking at the waveforms, it seems like it imparts gradual waveshaping—not squaring the waves like a clipper, just amplifying quieter details.
Another great option is Vintage Warmer 2 by PSP.
Its knee function is basically a control for amplifying quieter details.
Like the inflator, compression only begins at a higher amplitude, and even then, it’s very subtle. Unlike other limiters, the brick wall function can be turned off, allowing for a careful balance between maximization and monitoring peaks.
If we lower the input drive by -4dB, and increase the knee, we can achieve a lot of maximization without any attenuation until -12dB, and no significant attenuation until about -3dB.
Furthermore, we can vary the mix to incorporate more of the dry signal, resulting in additional transient detail - again, while achieving the benefits that would come from brick-wall limiting.
Lastly for this video, although I’m sure there a more plugins you could use for this purpose, let’s cover Saturn 2.
Saturn 2 offers subtle settings, like subtle tube or subtle tape, that when combined with the dynamics dial can achieve a similar effect.
It’s almost the same as the parallel compression effect when using the dry dial of the Pro-C2 but without the need for attenuating the signal first.
The drive function isn’t even needed unless you want to impart more harmonics.
Also, since this plugin allows for multiple bands processed separately from one another and routed in parallel, we can introduce varying levels of maximization to differing frequency ranges.
And in case you’re worried about varying the mix dial, or maybe aliasing distortion, both oversampling and linear phase processing are available.
All this to show, there are multiple alternatives to limiting. All of which amplify the lower amplitude range without the need for peak attenuation, or in the case of parallel compression, peak attenuation that can be controlled and made to sound much better than limiting.
Let’s compare limiting with these 3 processors. Notice that we can achieve a competively loud and upfront sound without the need for brickwall limiting.