Too many compression tutorials jump right into the specs. Of course they tell you which instruments you should apply them to, settings to use, etc., but rarely do they offer a perspective into the true Art of Compression. I hope to clear some of it up in this quick read.
If you look at a drum track you will see triangular shaped waveforms. A drum waveform is usually loudest at the beginning, the Transient, and then rapidly trails-off with very little Sustain. On a violin, a Staccato note is nearly all Transient like a drum. It is quickly bowed against the string and ends as soon as it begins. It also has impact, and very little Sustain. A Legato note on the other hand is nearly all Sustain. Yes, it will have a Transient as well, but the Sustain from the Staccato note will dominate the Transient comparatively, while the Legato will be mostly Transient.
When it comes to compression, many plugins offer a lot of power over Transients and Sustain. But most of them will specialize in one or two of the following classes: Leveling, Transient, or Sustain compression. Some compressors are really good at Leveling or Transient compression, while many others offer the ability to compress Sustain. And as we will see, some are a clear choice.
In the following sections, I will describe the nature of common compression techniques, why you apply them, and demonstrate which plugins appear to work best under some recommended settings.
LEVELING COMPRESSION
Leveling Compressors allow you to present audio with clarity, free from offensive spikes or inaudible dips in volume that naturally occur as a result of a recorded performance.
When the level of a recording is not enough to carry the track in the mix, you will likely increase its gain or push the fader, thereby boosting both the Transients and Sustain. This can not only be a problem for Peak Management, the Transient levels increase along with Sustain, but also to the performance itself. A singer, bass, or guitar spiking-out of the mix is undesirable.
Leveling Compression is considered a must for most tracks, especially when they are performed by a human and not via MIDI where velocity can be programmed. If used conservatively, it will greatly enhance the quality of playback. Leveling Compressors essentially allow you to ride the fader with the dynamics of each word or note played, ensuring the Transients do not overpower the track, meanwhile allowing the Sustain present itself at an acceptable, consistent level.
The first compressors were made to limit transients while simultaneously allowing you to increase the volume of non-transient sounds. This means spikes would be tamed and lower volume non-transients could be increased with the goal of providing a more even presentation of the audio. The most famous of these compressors is usually found in any professional studio, the LA-2A by Teletronix. Since its creation, many clones of the LA-2A have been created; one of the most popular of which is the CLA-2A by Waves.
The original LA-2A was a tube compressor that relied upon optical sensors to trigger compression. The LA-2A also does not have Attack or Release settings. Attack in the traditional hardware was fixed at ~10ms. By comparison, 20ms is the average attack of a snare. So depending on the amount of Attack you want to apply to your source, an LA-2A may not work. Likewise, the release of the LA-2A is around 60ms, although this is somewhat variable as a result of its design.
Some may say, however, the natural sound of the LA-2A (and many clones like it) rivals most other compressors. And I won’t disagree. Tube compression results in a very unique, colored sound that cannot be replicated by solid state. That said, we are not talking about saturation or distortion here, only Transient control, which is why I recommend a solid state FET (Field Effect Transmitter) compressor to perform Leveling. It is colorless and provides the fastest Attack of all compressor designs. With the ability to adjust Attack and Release to my needs.
Recommended Compressor: 1176 (FET) - Waves CLA-76, Arturia FET-76, or Logic Pro Vintage FET shown above
Threshold: ~-10dB
Attack: Fastest
Release: Slowest will squash everything, Fastest may produce some artifacts
Ratio: 2-5:1
Avg. dB Reduction: ~5dB
Application: Any instrument or voice
Caution: Where extreme Transients exist in your recording, a Leveling Compressor will likely need to be followed by, or can be preceded by - depending on your ears - a Transient Compressor in your FX chain.
TRANSIENT COMPRESSION
There are few compressors that are capable of microsecond-quick Attack; some are simply too slow to grab the signal immediately as it is being presented. Audio with extreme transients like a Snare or Kick may require a compressor that allows very little of the transient to get through unmanaged. So here you want to employ a compressors with the fastest Attack possible.
Typically this compression technique is applied in parallel to the source track to taste. Transient Compression may cause your source to sound unnatural, however, when applied in a Parallel Bus, Transient Compression brings-out the Sustain of a source and can extend the tonality and loudness of that instrument. When added back to the original audio, it can also be quite musical. So unless you are looking for an edgy sound design, it should be applied with some degree of caution.
VCA (Voltage Controlled Amplifier) Compressors rely upon the detection of a signal by measuring the voltage being presented. They are able to quickly identify transients. But I have found most VCA and even FET compressors fall short when it comes to Transient Compression. All compressors can control some of the Transient, not all can give you the Transient-crushing control we are looking to achieve here without a lot of clicks or distortion.
After testing the CLA-76 and the FET-76 against the Logic Pro Vintage FET, I was surprised to conclude, Logic Pro has the best representation in this class. The others simply do not compress the Transient enough to win-out in this category. I have also tested a number of other emulations from SSL and API and found they didn’t produce the results I desired either. Although I am impressed with both the Split EQ from Eventide and the Transient Shaper from Softube which are actual transient shapers, not just compressors.
What I look for in Transient Compression is the transient to be caught by the compressor very quickly; none of it gets through. And with a fast Release, the Sustain is left untouched, as close to the original source as possible. Even much louder than the source if I choose.
Recommended Compressor: 1176 (FET) - Logic Pro Vintage FET (shown above)
Threshold: ~-30dB
Attack: Fastest
Release: Fast, ~20ms. Slower Release times may begin to compress the Sustain
Ratio: The highest ratio recommended is ~8:1. Higher ratios will compress the transient more, but distortion starts to take over. Too low, and the Transient may not be reduced at all
Avg dB Reduction: 10-20dB
Make-Up Gain: You may find you need little additional gain here if you are blending it back into the original source
Application: Any instrument with extreme Transients
Caution: Over-compression or pushing these settings beyond what is recommended can introduce a lot of distortion if not applied musically.
SUSTAIN COMPRESSION
What I am looking for when applying Sustain Compression is a crisp Transient and very little Sustain. The Attack must be somewhat delayed in order for all of the Transient to get through untouched, but then once compression kicks-in, the Sustain level is dramatically reduced. This technique is also commonly employed in Parallel processing. By highlighting only the Transients of a track, the source track can be brought-forward into the mix, given extra punch. A lot of extra punch!
Most compressors can perform Sustain Compression fairly well, even those without Attack and Release dials. But before we go further, we should revisit the discussion around the LA-2A.
Remember the Attack and Release specs for the LA-2A are 10ms and 60ms respectively. Well it just so happens this is an excellent starting place for Sustain Compression.
On your drum bus, add a send to a parallel bus at 0dB. Put an LA-2A in the chain, and turn the Gain and Peak Reduction all the way up. Then set the Compress/Limit switch to Limit (100:1 Ratio). This will generate a lot of gain so you can either add a gain plugin and set it to about -20dB or pull back the fader about 20dBs. You should be close to unity gain. It’s that simple.
Notice there is still some sustain coming through the compressor, which is due to the increase in gain. It is especially annoying for instruments with long sustain like cymbals, tends to result in a pumping sound. So this technique should only be applied to tracks without a lot of sustain.
I actually prefer the CLA-3A since it is a solid state version of the LA-2A. Less tubecoloration. But either will do a fabulous job.
Recommended Compressor: (Opto) CLA-2A or CLA-3A
Gain: Highest - This pushes the compressor hard
Peak Reduction: Highest - This crushes the Sustain
Compress/Limit Switch: Limiter (100:1 Ratio)
HiFreq: Turn all the way clockwise, flat, this will ensure the full frequency spectrum till triggger compression
Avg dB Reduction: ~20dB increase
Make-up Gain: You will need to reduce gain by ~20dB to level match with the original source
Application: Any instrument or voice where you want to compress Sustain such that the Transient is the primary focus of the audio. You will typically see Sustain Compression employed on a parallel Drum Bus that excludes cymbals or room mics, anything with a long Sustain, and is mixed back in with the original drum bus to taste.
Here is an example of the same technique in Logic Pro although it does not result in the same Sustain-crushing effect as the LA-2A.
