Limit Pro
Table of Contents
Related manuals
Introduction
"Limiting is about making the song loud."
Well, yes. Sort of. Limiting is used to prevent peaks from going over 0 dBFS (or 0 dB true peak) so that you can increase the gain while still keeping your audio below 0 dB. This will make the song appear louder. But the trick is doing this without affecting the song in a bad way, and the trick depends on your song!
In a very simplified manner, think of a limiter as a compressor with an infinitely high ratio and a fixed threshold at 0 dB. If the attack isn't fast enough, you will get overs and the limiter isn't doing a very good job. So early limiters basically had a single parameter: the release time. With that, you could go from a fast and distorting limiter that's almost like a clipper, to a slow and cleaner limiter with a lot of pumping.
As limiters evolved, more effort was spent on making the release curves better, mostly taking cues from older program-dependent compressors with adaptive release times. Some limiters became essentially two-step algorithms, with an advanced compressor before a simple clipper.
But neither of these addressed the core issue: if you want a loud mix, you need a fast limiter, and you will essentially end up with a clipper with a lot of distortion. There's no simple fix for this, since it depends on your audio material. A dense mix with electronic instruments will require a different type of limiting than a classic RnB song. And it's not only about release characteristics; to get a clear, loud, and transparent mix, you need to tailor both the type of peak suppression and the method of release to suit your music. In general, you want it as fast as possible without distortion. That's where the real challenge (and opportunity!) lies.
Peak Suppression
The peak suppression should be fast but transparent. But depending on your audio material, "fast" will mean different things. There's no single recipe that works for all kinds of music.
Most limiters only use a single peak suppression method, even if they have a lot of different "modes". These modes typically only change the amount of smoothing or the release characteristics. But the core of the limiter, the peak suppression, is the same. Some methods that sound great for dense and bright mixes will sound heavily distorted on a dynamic rock track. Methods that work for an acoustic recording can totally destroy the transients of a dense dance mix.
Limit Pro uses three different peak suppression methods: Standard, Adaptive, and Fast. They are all good at different things. Fast is almost like clipping, without actually clipping the signal. Adaptive tries to be as fast as possible, but can slow down if needed to prevent low-frequency distortion. Standard is predictive and reliable—a workhorse that gets the job done, but might not give you that extra dB of loudness.
On top of this, you have the Smoothing control, which determines how fast the peak suppression is. A dense mix with occasional peaks can have low smoothing. A bass-heavy ballad probably needs 2–4 ms of smoothing. Again, there's no single recipe that works for all kinds of music.
Release
In an ideal world, we wouldn't need to have a release control on a limiter. And in many cases, you don't need to bother with it. But when you do need to use it, it needs to come in different flavors.
The Transient Release is almost an extension of the peak suppression. It is used to reduce distortion from the peak suppression and is typically kept short (10–100 ms). This is similar to the release control that you see on most other limiters.
If you have a dynamic mix, the limiter needs to be a bit more understanding of what's going on. That's what Adaptive Release is for. It's program-dependent, meaning that it releases faster for transients than for sustained material, and you can set the Amount of the release with a slider. Adaptive Release allows you to use really long release times without having the limiter "punish" the mix.
The goal of the release is to be as fast as possible without distortion, so it needs to adapt to the track and suit different songs. But Limit Pro takes this adaptiveness one step further.
Frequency Dependent Release
The Adaptive Release can be made frequency-dependent, meaning that it has different release times for different frequencies. This is very different from a multiband compressor since it doesn't change the tonality or frequency spectrum of the audio.
To avoid distortion, a kick needs a longer release than a high-frequency snare. The frequency-dependent release behaves differently for different sounds, reducing the low-frequency distortion coming from the limiter.
In other words, Adaptive Release determines how the limiter responds over time, while Frequency Adaptive modifies this behavior according to the frequency content. This allows for shorter release times for high-frequency transients and longer release times for low-frequency material.
Low Frequency Protection
A fast and loud limiter is typically bad at handling low frequencies. Different limiters handle this in different ways, typically with long look-aheads (and lots of latency). We use different peak suppression methods, as well as program-dependent and frequency-dependent releases. But there's another trick we use as well.
It started with the question: Is there an optimal way of limiting a specific frequency with as low distortion as mathematically possible?
It turns out that such a method exists, and we chose to implement it in two different ways. The Frequency Adaptive release is one implementation, while the Low Frequency Protection is another. The purpose of Low Frequency Protection is to take over limiting for sub-bass frequencies to avoid distortion when, for example, the 808 kicks in.
Transparent Signal Path with Low Latency
When Limit Pro isn't reducing gain the signal path is completely clean without any box-tone or character. So you can safely keep Limit Pro on your master bus and have it catch peaks without worrying that it'll change the tone of your track. Together with the extremely low latency, that means that you can keep the limiter on your master bus while tracking and working in your project.
However, Oversampling and True Peak will add a bit of latency, around 3 ms for True Peak and an additional 1 ms for oversampling, and the Oversampling feature will also add a linear phase anti-aliasing filter to the signal path. So for the cleanest path with lowest latency, turn off Oversampling and True Peak.
What about Multi-Stage Limiters, or Stacking Several Limiters?
Stacking several limiters or clippers after each other can be a very useful creative tool. But if the objective is to get as much peak reduction with as little distortion as possible, it's generally better to stick to a single algorithm.
The different parts of Limit Pro work in tandem, and their interaction is delicate. This approach wouldn't achieve the same result if Limit Pro was built by stacking several processors (compressors, limiters, clippers, etc.). It wouldn't achieve the same loudness, and perhaps most importantly, each part would slow down the processing, making the total limiter slower. And speed is of the essence here. We want our limiter to react fast, so that it can be clean and transparent.
First Use
The two most important controls are Input Gain and Smoothing. They affect the sound more than anything else.
Step 1: Input Gain
Increase Input Gain to achieve the desired loudness.
How high can you push it until you either get distortion, or the dynamics of the song gets lost? The Loudness Loss and Distortion meters are good indicators of when pushing it too far. Keep Loudness Loss below 2 dB.
It's important to listen to the "groove" and "rhythm" of the song. If the track has a hi-hat pattern that makes you want to dance or a kick/snare relationship that has that special feel, make sure that those parts are intact. Loudness isn't worth anything if the song doesn't sing.
Step 2: Smoothing
Adjust Smoothing to control the distortion.
Increasing Smoothing will lower the distortion, but might make the track less dynamic. Dense tracks typically require less smoothing than sparse tracks.
Step 3: Peak Suppression Method
Go through to the different Peak Suppression Methods and choose your favorite, tweak Smoothing to suit the different methods.
- Standard: Predictive all-round method with low distortion.
- Adaptive: Has a snappy response and a solid low end, adapts to the incoming audio.
- Fast: Can be very fast, for shaving off the peaks of a track.
To be able to A/B test your settings, engage Gain Match and use the smooth Bypass button next to it.
Step 4: Release Curves
Fine tune your settings by working with the different release times. Try the following alternatives.
-
Dense mix:
Adjust Transient Release, keep Adaptive Release Amount low or at 0%. -
Pop/Rock mix:
Adjust the Adaptive Release, keep Adaptive Release Amount below 50% -
Bass heavy, sparse mix:
Set Adaptive Release Amount at 100%, turn on Frequency Adaptive, and tweak the Adaptive Release time.
There is no single set of controls that works for every song. Start with the default settings and adjust the limiter to suit the material. Flip through presets to find your favorite starting point. Do a lot of A/B-ing.
User Interface
The user interface consists of a top window with graphs or meters, and a bottom part with controls.
Display
The display has two different modes: one with scrolling waveforms and graphs, and another with just metering. Select which view you want to see by clicking either Waveform or Meters at the bottom of the display.
Scrolling Waveforms
In the top part of the scrolling view, four different curves are displayed simultaneously:
-
Light Blue: This is the total (combined) gain reduction from the limiter.
-
Darker Blue: Shows the gain reduction coming from the Adaptive Release.
-
Yellow (Line): Loudness loss graph.
-
Yellow (Filled): Shows the gain reduction coming from the Low Frequency Protection.
At the bottom of the scrolling view, a single curve is shown:
- Red: Amount of distortion.
Loudness Loss
The Loudness Loss is a heuristic measurement of when a gain increase doesn't correspond to a louder mix.
For example, if you push the gain 12 dB and the LUFS readout gets 12 dB louder, that's great! You might be able to push it more since a 1 dB increase corresponds to a 1 dB louder mix.
But if you push the gain by 12 dB and the mix only gets 9 dB louder, you have 3 dB of loudness loss (12 − 9 = 3 dB). This suggests that you might want to back off a bit because you've hit a ceiling.
You can also lower the loudness loss by decreasing the release times, at the risk of adding more distortion.
It is good practice to keep the Loudness Loss below 3 dB.
Distortion Meter
The distortion meter measures how the limiter changes the audio and analyzes how that will affect the output audio.
Limiting will always introduce distortion, but not all distortion sounds the same. The smooth saturation of a tube is very different from a bit crusher or wave folder. Some types of distortion blend easily with the original audio, whereas others stick out like a sore thumb. The distortion meter analyzes what type of distortion will be introduced and reacts more to distortion that is likely to be audible and "bad".
The distortion is also visualized to the left in the "Distogram".
Distogram
The distogram visualizes how the distortion sounds. A peak at the top of the graph means that the harmonics created by the distortion are located at a much higher frequency than the incoming audio, which is typically perceived as a "bright" (good) or "fizzy" (bad) type of distortion. A peak at the bottom of the graph means that the harmonics created by the distortion are close to the incoming audio, which can be perceived as "natural" (good) or "growly" (bad).
Since distortion is very subjective, and its impact depends on the type of music you are making, it's not necessarily a bad thing. But if you do get high distortion readings, you should be aware of it and make an informed decision.
Increasing Smoothing typically lowers the amount of distortion.
Meters
The meter view summarizes the waveforms using momentary, 4-second hold, and infinite-hold values. You have a total of five meters in this view.
The big meters have a max value readout to the left and the current peak value to the right. The left channel is displayed at the top, and the right channel at the bottom of each pair.
-
Low Frequency Protection: Left and right, measured in dB.
-
Adaptive Release: Left and right, measured in dB.
-
Total Gain Reduction: The combined gain reduction, measured in dB.
Distortion and Loudness Loss meters are displayed on the side.
-
Distortion: Amount of distortion. Unitless.
-
Loudness Loss: Measured in dB.
Clicking on any meter value will reset all meters.
LUFS and Peak Readouts
On the right side of the display are three readouts: LUFS, True Peak, and Peak values. The True Peak measurements are following the ITU-R BS.1770-3 standard.
If the numbers are yellow, Gain Match has been engaged, and the readings are taken before the output gain. It's worth noting the Monitor will affect the metering, since the meters reflect the audio as you hear it.
You can choose to display Integrated, Momentary, or Short LUFS from the "cog wheel" menu.
Controls
- Adaptive Release
- Sets the release time of the program-dependent release. If Frequency Adaptive is turned on, the release time will correspond to a range of release times, indicated by the bar under the control.
- Adaptive Release Amount
- Sets how much the Adaptive Release will affect the audio.
You will typically use a high Amount with a short Adaptive Release, or vice versa.
- Frequency Adaptive
- Enables frequency-dependent release times, making the release times longer for low frequencies. With Frequency Adaptive enabled, you can typically use shorter release times and still get a clean low end.
- Low Frequency Protection
- Enables an extra layer of protection for low frequencies that prevents loud low-frequency parts from becoming distorted by the peak suppression.
It is important to note that Low Frequency Protection doesn't prevent low frequencies from being limited, but acts like a gain rider that smoothly (but really fast) lowers the volume when the low frequencies hit. This prevents low frequencies from being distorted by the otherwise fast limiting algorithm.
The image below is an exaggerated example of the release times with and without the Frequency Adaptive release circuit. With Frequency Adaptive Release, the release time is shorter when the snare hits than when the kick hits. This allows you to use generally shorter release times without getting low-frequency distortion.
In the graph, you can also see the yellow graph, where Low Frequency Protection works to avoid distortion in the bass.
If you have trouble with low-frequency instruments getting distorted, such as an electric bass that sits up front in the mix, turn up Adaptive Release Amount to 100% and experiment with Adaptive Release and Frequency Adaptive Release.
- Transient Release
- Sets the release time of the peak suppression.
- Limiter Gain
- Sets the input gain to the limiter; this determines the loudness and is typically the first control you adjust.
- Smoothing
- Sets the speed of the peak suppression. A high smoothing reduces "clipping"-like distortion, typically in the high frequencies, but can lead to a "soft" and "pillowy" sound.
- Peak Suppression Method
- Sets the type of peak suppression.
-
Standard: A smooth, all-round type of peak suppression that fits a lot of material, except the most transient and dense mixes. Typically used with 0.5–2 ms Smoothing.
-
Adaptive: Similar to the peak suppression in Weiss MM1. Adapts to the audio material and can often be used without any Release at all.
-
Fast: The most aggressive type of peak suppression, designed for dense, electronic mixes. With Smoothing at its lowest setting, this method almost becomes a clipper.
- Bypass
- Bypasses the audio processing without any clicks or pops. Useful for auditioning in combination with the Gain Match function.
- Gain Match
- Automatically sets the Output Gain to match the Limiter Gain, making A/B auditioning easier.
- Output Gain
- Sets the final output gain of the plug-in.
Expert Controls
There are expert settings accessible from the "cog wheel" menu in the lower left corner of the display.
- Sidechain Link
- Sets the amount of gain reduction ganging between the left and right channels. As a rule of thumb, the more gain reduction you apply, the more you want the sides to be ganged to avoid the mix moving around from left to right. Sidechain Link does not affect the peak suppression, which is always unlinked. It only affects the slow release.
- Waveform Range
- Sets the zoom level of the gain reduction in the scrolling waveform.
- LUFS Mode
- Sets which LUFS mode to display in the window.
-
Momentary: Updates often (400 ms window).
-
Short: The typical "current level" estimate used. Uses a 3 sec window.
-
Integrated: Measures LUFS from the start of playback until stop (or reset).
- LF Protection Range
- Sets the (soft) upper limit for the Low Frequency Protection algorithm.
- LF Protection Mode
- Sets the type of Low Frequency Protection.
-
Fast: Reacts fast, but offers slightly less Low Frequency Protection.
-
Slow: Reacts slightly slower, but has better Low Frequency Protection.
Display Modes and Extra Controls
At the bottom of the display, you also have the following controls.
- Cogwheel symbol
- Opens the settings menu
- ? (Help)
- Turns help texts on/off.
- True Peak
- Turns true peak limiting on/off.
- Oversampling
- Turns oversampling on/off.
- Monitor
- Turns on delta monitoring, where the output is the difference: the audio that the limiter removed.
- Waveform/Meters
- Sets the display to either show scrolling waveforms or bar graph meters.
- Reset Meters
- Resets all meters. This can also be done by clicking on any meter readout.
Presets
The presets are generally categorized into two main categories: Mastering and Mixing, with the mixing presets made by Philippe Weiss, and the mastering presets made by Mat Leffler-Schulman.
By default, some controls are locked and will not be changed by the presets:
- Limiter Gain
- Output Gain
- Gain Match
- True Peak
- Oversampling
- Monitor
That means that you can easily browse through the presets without having to readjust Limiter Gain for every preset.
To unlock controls, right-click the control and select "Unlock" or "Unlock All".
Credits
Cameron Clark – Algorithm design
Fredrik Jansson – Algorithm design
Niklas Odelholm – Algorithm and product design
Patrik Holmström – UI programming
Arvid Rosén – DSP optimization
Johan Bremin – Quality assurance
Philippe Weiss – Presets
Mat Leffler-Schulman – Presets