Musician's Friend Tech Tips: Compression
Producer's Edge Magazine
Home       Members    Calendar    Who's On
Welcome Guest ( Login | Register )

Musician's Friend Tech Tips: Compression Expand / Collapse
Posted 8/9/2007 4:18:44 PM

Forum Moderator

Forum Moderator

Group: Moderators
Last Login: 7/30/2015 12:52:10 PM
Posts: 1,901, Visits: 2,928
Musician's Friend Tech-tip of the Week

TECH TIP: How To Set A Compressor
For Voice and Instruments

By Darius Van Rhuehl

This tech tip will give you real world, practical information on how to set a compressor for various applications. We’ll start with a very brief review of the compressor controls and then get right into it. You’ve probably also noticed that there are different operating designs of compressors e.g. optocompressor, VCA, variable Mu etc., and software variations as well. In the second part of this tip, we’ll discuss what they’re best used for. We assume that you’ve read the countless descriptions of what a compressor does.

There are basically two types of compressor, the leveling compressor and the limiting compressor. The only difference between the two is that limiting compressors have a separate threshold and ratio control, whereas the leveling compressor (also called leveling amplifier) has a single control for gain (or peak) reduction. With leveling compressors, threshold and ratio (slope) are interrelated and change depending on the input signal. While these compressors were initially designed for broadcast in order to control overall levels of program material, leveling compressors such as the Teletronix LA-2A and Summit Audio TLA-50 are in constant use in the studios for vocal and guitar tracking. For our purposes, we’re going to limit ourselves to limiting compressors. (Well I couldn’t say gain-reduce ourselves now, could I?)

Controls of the Limiting Compressor

There are four basic controls that affect compression. In no specific order, they are Attack, Release, Threshold, and Ratio. Put simply, Threshold tells the compressor when to work; Ratio tells the compressor how hard to work; Attack tells the compressor how fast to work, and Release tells the compressor when to stop working (or how long or short a time to return to its natural state).

Hard/soft knee: This is selectable in many compressors. It determines how gradually or abruptly output gain reduction occurs depending on the ratio. A soft knee is a gradual transition, which makes it more "musical" sounding, and particularly useful for vocals or any continuous material. It also allows higher ratio settings should they be necessary. A hard knee is good for explosive or percussive sounds, such as drums and heavy brass.

Makeup gain: This is essentially an output volume control. Its job is to counteract the gain-reduction effect of compression and give you the output level you desire.

How To Set a Compressor For Tracking

These are your basic ballpark, jumping-off-point settings that will work for just about any application. Naturally, the settings you wind up with in your work will vary based on the music and the response of your equipment.

Setting the Ratio: Rule one: Don’t squash (unless there’s a crying need to, or a creative reason). Start with a ratio between 2:1 and 4:1. A ratio of 3:1 is good if your compressor allows or has a variable ratio control, such as the Tube-Tech CL1B.

Setting Attack: This setting will vary depending on what you are compressing. Start with a relatively fast attack time (around 25ms).

Setting Release: This control will vary the most. Start with a medium release (50-100ms) and be prepared to adjust. Too short a release will give you an unnatural pumping effect, and too long of a release will prevent signals from returning to normal levels soon enough, causing subsequent notes not to sound as loud as they should and a dull overall mix.

Setting Threshold: As you run signal (your track or instrument) through the compressor, start to adjust the threshold control (usually counterclockwise) until you see the gain reduction meter read a consistent –2 to -3dB of reduction. That should be sufficient compression to handle most tasks while leaving a decent dynamic range. Another common setting for vocals is to set the ratio at 2:1 and threshold such that no gain reduction occurs during soft passages, with 3dB to 6dB of gain reduction during normal singing.


Putting it all together, the general, applies-to-all-areas setting for tracking with compression is a relatively fast attack, medium-fast to medium release, ratio between 2:1 to 4:1, and threshold set for a fairly constant 2dB to 3dB of gain reduction. This is a basic starting point that will work effectively with almost any compressor for almost any application. In the next part of this tip, we’ll discuss various types of compressors, give you problem-solving and creative applications, and settings that are good starting points for different instruments and vocals.

Post #208
Posted 8/17/2007 7:50:48 AM

Heavy Cat Status

Heavy Cat Status

Group: Moderators
Last Login: 6/20/2017 1:07:35 AM
Posts: 483, Visits: 1,192
Musician's Friend Tech-tip of the Week

TECH TIP: How To Set A Compressor
For Voice and Instruments Cont'd

By Darius Van Rhuehl

How To Use A Compressor – Part II

What type of compressor should you use?

In the first part of this tip we gave you a basic setting to start with that can handle just about any tracking application you might have. In this part, we’ll discuss compressor types, applications, and give more start-off settings for specific instruments and voice.

If you’ve shopped for a compressor, doubtless you’ve seen compressors of various types including optocompressor, VCA (Voltage Controlled Amplifier), FET (Field-Effect Transistor) OverEasy (the ratio increases as the signal gets hotter), and variable Mu, which uses the natural compression characteristics of tubes. Each has its own method of reducing gain by virtue of what type of components act on the incoming signal, and each has its own sound. It’s hard to say which is the most effective for a particular job, but they seem to have divided themselves into accepted uses. Designers such as Joemeek’s Ted Fletcher believes that optocompressors are more "musical" than VCA compressors, which to his way of thinking stonewall the signal as opposed to the non-linear behavior of the optocompressors. Conversely, the linear behavior of VCA compressors actually makes them more transparent in operation, which Transparency combined with the VCA’s ability to stonewall a signal makes them a great tool for overload and speaker protection, as well as audiophile recording, where the natural sound of the instruments is preferred over affectation. Optocompressors tend to color a signal more and are therefore used for effect. (Sort of throws a crimp into the whole musicality debate, doesn’t it?)

Optocompressors are popular for treating vocals, drums, and bass guitars in pop and rock music, where a larger than life sound is desired. Stereo optocompressors such as the Joemeek MC2 are also popular for stereo buss bus compression. By the same token, the Universal Audio 1176 LN, which as an FET compressor, is one of the most popular of all time for drums, electric guitar, bass, and bus compression (two linked in stereo). The dbx 160A OverEasy is another favorite for drums. Compressors that use tubes for gain reduction (variable Mu) are most popular for overall mix compression and are mainly used in mastering. They provide the glue that holds a mix together while allowing each instrument to retain its sonic integrity. To find out what works best for you, you’ll need to hear it in context. With Musician's Friend’s 45/45 Double Satisfaction Guarantee, you can audition any compressor to see if you like what it does for your sound.

Software compressors digitally mimic the effects of hardware compressors during mixdown but can’t really control peaks and levels while you record. They don’t reduce actual voltages in response to input signals, they just do math. As such, a software compressor should be thought of as an after-the-fact creative effect and used accordingly. You’ll still need a hardware compressor for DAW tracking. Speaking of DAWs, while it is tempting to drop a software compressor on every track because you can, be aware that they also mimic the effect of making noise louder as well. Plus, they can also affect the overall sound of your mix in ways you don’t expect and won’t want. Since they are really just crunching numbers, the more math your computer does, the longer digital words get. Your CPU can only handle words of a given length, which means a word that’s too long gets truncated (a nice way of saying chopped off at the knees or guillotined, if you will). In that truncation, you can be losing sound quality and low-level detail (such as reverb tails). The same rule especially applies for overall mix compression. Like any seasoning, it’s best used sparingly, or you may be sending a mastering house a mix they can’t fix.

Application Tips:

  1. When in doubt, don’t compress. It can always be added later, but never undone. For example, if you have an electric or acoustic guitar part that sits fine in a mix as is, leave it be.
  2. If you plan on having your music professionally mastered, avoid putting a compressor on the stereo L/R main bus to compress the entire mix. If you do put compression on your mix (some people like it regardless), send two mixes to your mastering engineer, one with compression, one without. If they can work with your compressed mix and feel that it doesn’t compromise their process (and ultimately your music), then fine. If not, they have what to work with. Don’t use compression solely for the reason of making a mix louder. While many engineers do this, the danger is that an over-compressed mix will lose contrast and will sound dull and boring, losing the listener’s ear. Some mastering engineers actually wind up using expanders (the electronic opposite of compressors) to try to bring back the dynamic range lost by over-compressed mixes.
  3. A good way to use stereo bus compression is when you route drum tracks to a stereo bus (not the main L+R) to create a submix. Insert an compressor and an EQ on your stereo bus (this works in hardware and software), squash the drums, boost the highs at 10kHz, lows at 80Hz (leave everything else flat) and send the bus output into two open channels of your mixer. Pan one channel hard left, the other hard right. Now, sneak the drum submix into the mix along with the normal drum tracks so that it widens and reinforces the drums without overtaking them. Congratulations, you have just learned how to create what’s called the New York Drums sound. This is great for any percussion-driven music such as rock or hip-hop.
  4. To get more snap out of a snare, set an attack time that allows the entire transient to pass. Start with a ratio between 2:1 and 4:1. You’ll need a slower attack time and a fairly quick release. This one requires experimentation. Run your snare track through the compressor starting with the fastest attack possible. You’ll hear a very clipped attack. Now start dialing the attack back (slower) until you hear more of the transient. Don’t be afraid to go beyond the transient and dial back to a faster release until you know that you’re getting just the transient. (It’s like setting the water temperature of a shower, where there seems to be only a micro nano-meter between fantastically hot and freakin’ freezing.) Once you have it, set a fast release time. What you’ll hear is the entire attack passing while the sustain portion of the sound dips down for a moment and then rises quickly. This increases the level between transient and sustain giving the illusion of a louder attack combined with a subtle pumping of the snare, which contributes to musical drive.
  5. If you’re working with sampled cymbals and want to create a more realistic decay, as though a drummer hit the cymbal and moved on (so to speak) set a longer release time on the compressor. You can adjust the release time to coincide with the next musical event so that there’s not too much overhang, just enough to give the impression that the cymbal is fading of its own accord and not by gating or by fixed sample length.
  6. Ducking rhythm tracks is another way to use compression. Let’s say you have guitars and synth pads that by themselves sound good and balanced. The bad news is that when the vocals come in, they are fighting to be heard. This technique requires the ability to send a signal to the sidechain of a compressor that will control the volume of the guitars and keyboards. With plug-ins, this is easy if they have sidechain access. Just bus an output from the vocal track to the sidechain key in of the compressors. The rhythm tracks will duck when the vocals come in and go back to normal when the vocals exit. For rhythm tracks, you can use a higher ratio, ranging from 5:1 to 7:1. This also helps to keep them even throughout the song*. Set the threshold so that when the vocals come in, the rhythm tracks simultaneously decrease (or duck) by 2dB to 3dB. Set a fast attack and a long release (1 sec-2 sec). Naturally, you’ll have to adjust the release to suit the music. A fast release time will cause the tracks to pump, which you can also use to your advantage in terms of creating musical drive by adding a subtle pumping to heavy guitars. Again, experiment.
    *When you have tracks that need to keep fairly constant in level throughout a song, such as rhythm guitars (clean or distorted), synth pads, and backup vocals, it’s okay to use a higher compression ratio. Some engineers will use ratios as high as 5:1, 6:1, and 7:1. As ever, it depends on your tracks. You may choose a lower or even higher ratio depending.
  7. To remove sibilance from vocals and harshness from cymbals, a technique called de-essing is called for. This uses the sidechain input of the compressor and a parametric equalizer. Split the output of the vocal track so that along with your normal main out, you are sending to the equalizer as well. The output of the equalizer will go into the sidechain key input of the compressor. To start, you will be monitoring the EQ only. Usually, vocal sibilance occurs in the range of 3kHz to 8kHz. For cymbals, harsh sounds can occur between 5kHz to 10kHz. We’ll talk about treating vocals, but the technique applies to both. On the equalizer, turn the boost all the way up and sweep through the frequency band (while the vocal track is playing) listening for sibilance (the hissing sound of "s", also "t", "ch", and "sh"). When you find the most obnoxious frequency, set a very narrow Q (bandwidth) so that the compressor will just react to the evil frequency and not squash the good and pure. Leave the signal boosted anywhere from 10dB to 15dB. Now drop in the compressor, and with a high ratio (8:1-10:1), listen to the track. Start with the Threshold set high and then lower it until you hear and see the compressor acting on the sibilance. Gain reduction should show between 3dB and 6dB on the meter. If the compressor is acting on some but not all of the sibilance, the frequency may not be set right or the bandwidth (Q) may be too narrow. Try widening the bandwidth a tad and adjusting the frequency until you get the behavior you’re after. By the way, don’t worry about the oddly EQed vocals, you’re not going to hear them.
Basic Settings By Application

While the basic setting we gave early on will work as a jumping off point for most compressors and applications (particularly well on the Tube-Tech CL 1B), keep in mind that each compressor reacts differently. Even two of the same model will have slightly different characteristics. (Electronics is not an exact science.) Here are some basic settings by instrument that work well for VCA compressors. Again, there are no hard and fast rules or formulas that work in all situations. First and foremost, decide if you even need compression and go from there.

Attack/Release-Time Key:
Fast = 25-50ms; Medium = 100-500ms; Slow = 1sec–2sec (all are plus or minus).

Basic Instrument Settings

Distorted guitar: Fast attack, slow release, ratio of 4:1, or higher for more sustain.

Acoustic guitar: Medium attack and release, ratio between 2:1 to 4:1.

Drums: Fast attack, fast release, ratio between 2:1 to 4:1.

Bass: This one’s tricky. To tame aggressive attacks use a fast ratio, to accentuate attacks use a medium attack with a medium-fast release (good for slap bass), to increase sustain use higher ratios with a slow release (depending on music).

Horns: This depends on what you require: for smooth background pad-like performance, a fast attack, slow release, and higher ratio (4:1 and up). A fast attack and fast release will add punch. The requirements of the music dictate here.

Piano: For lively rock piano style, use a medium attack to let the transient through (like the snare) and a medium-fast release. Ratio of 2:1 to 4:1. For ballads, use a fast attack and a slow release.

Vocals: Medium-fast attack, medium release or slow depending on music, ratio of 2:1 and no higher than 4:1. To make the vocal sit in the track more effectively, try a higher ratio with a higher threshold, or a lower ratio with a lower threshold. In either case, compression will be heavy, relatively speaking. Gain reduction can be either 2dB to 3dB at the higher ratio, or 3dB to 6dB at 2:1, provided there is little or no reduction during passages sung at normal levels.

Our Final Reflections

While it may seem that this tip is quite lengthy, we’ve haven’t covered all of the uses and possibilities of compressors, but this should be enough to keep you going for a while. There are also other dynamics processors to talk about, which include levelers, limiters, expanders and gates. Look for them in upcoming tech tips.

Post #216
« Prev Topic | Next Topic »

Permissions Expand / Collapse

All times are GMT -5:00, Time now is 6:54am

Powered By InstantForum.NET v4.1.4 © 2018
Execution: 0.031. 9 queries. Compression Disabled.