Compression

***Control The Dynamic Range

And or change its character
The distance between the loudest and quietest parts of a signal

We are fighting an opposing battle

1. Try to make things efficient
2. Trying to make things punchy
Reasons for using a compressor
● To make the dynamics more consistent
● We want to feel like there is a big dynamic range without there being a big one (maximum
illusion minimum voltage)
● FEEL!!!!!!! (release) Dynamic Pumping - The big intangible
○ the bounce of your track
● Punch / Presence (upfrontness) - Slow Attack. You can bring sounds forward or back in the
mix (depth). Release - can move the body of a sound forward or back in
● Bus compression to “Glue Elements and create inner bounce”
● Slotting - Side Chaining - Creating Space
○ having one sound make space for another sound.
○ Feel and bounce -
○ The sidechain source becomes dominant
● Tonality / Character - Amplitude Modulation Artifacts (compressor jittering), Analogue Modeled
(harmonic distortion modeling “clipping”) (like Glue Compressor)
○ Also can change the tone of transients
● As an effect - Over compression as a pumping effect (breathing) - MORE OF A PRODUCTION
MOVE THAN MIXING
○ Side chaining bass in dance music
 Input / Output In out

Threshold -
When the input level exceeds the threshold the compressor engages

Ratio - Sets the ratio between the input and output volume levels above
the threshold (gain reduction)
- Ratios above 10:1 are considered limiting
- Hard limiting is Infinity to 1
- 0 attack time
Attack - is engaged by the signal exceeding threshold
and is the time it takes to the arrive at the output ratio
(controls the transients or not) - affects punch, and presence (placement
Of the transient)
Release - after signal goes below threshold ..is the time it takes to come
back from the compression ratio (back to input level) (FEEL & placement
Of the body of the sound)
Attack (Presence and Depth)
● The closer a sound is the more transient there is in relation to the
body(reverb) of the sound
● Quick attack pushes the transient into the body and thus pushing it back
into the mix
● Say compressing a snare… quick attack feels like the drummer is using a
chopstick to hit the drum and a slow attack feels like the drummer is hitting
it with a thigh bone. On acoustics it quick attack seems like a small light
pick and long attack feels like a thick meaty pick.
 Release
Depth: quick release is close sounding, slow release is far sounding
Groove of the release : generally going for the release on the upbeat
Character of the release
Quick Release = Jittery, Aggressive, distorts, like too much coffee (uncontrolled) pushes upfront (hear the quiet details /
nuances)
Long Release = Controlled, pushed back, soft (tucks in nuances)
If you there is a zone in which the release is grooving,,, tend to push it later rather than earlier because its more
controlled
AUTO RELEASE - Automatically adjusts the release time depending on the content.. I don’t like it… I like to choose the
bounce! (except for multibands)
Make up Gain - Big newbie mistake
● Don’t compress for a volume boost…
Manually adjust output volume so that the
sound remains at a similar volume
○ Our brain automatically thinks louder is better, but in
the end everything can’t be louder and you screw up
your headroom not paying attention to that
issue…….
 Cracking the Compressor Code
Set up the compressor (maximum compression)
1. Fastest Attack Time
2. Fast Release Time
3. Highest Ratio
4. Sensitive Threshold (the gain reduction is constantly engaged)
5. Turn makeup gain off and manually move the output volume around to compensate for volume changes throughout the
process. Fabfilter Pro C2 has a good auto gain that does this automatically!
Order of Compressing (to hear what the subtleties of each parameter are doing)
1. Choose compressor algorithm if it has options for that
2. SET ATTACK Back off the attack to desired Punch, Presences, (how thick is the drumstick)
a. If there is a zone of good punch / attack.. Air on the side of quicker attack because it’s safer….
3. Set Release
a. groove / right attitude
b. & presence of the body of the sound
c. if there is a range of feel and body presence that works then tend to the longer release (safer)
4. BACK OFF THE AMOUNTS
a. Back off Ratio so the attack and release are perceived but more transparent (how processed do i want?)
b. Back off Threshold so that gain reduction returns to 0 between hits
c. & or apply soft knee curve
d. & or back off the Wet from the wet / dry (parallel=bottom up compression)
e. & or back off the RANGE (compress the medium dynamics)
5. Bypass on and off and adjust makeup gain so there isn’t a substantial level boost or drop. Use an LUFS meter if you need
*****Stages 4 changes depending how aggressive or transparent you want the compression processing.
Fixed Thresholds
● like LA2, 1176, Ableton’s limiter (ceiling),
Fabfilter Pro L limiter
Hard Knee / Soft Knee - Great for sounds with a big dynamic range.
It will smooth out the artifacts of some sounds being smashed because they are loud and others not being touched
because they are quiet

● Soft knee fades on the compression ratio

○ More transparent sounding because the threshold and ratio have a range to “morph the
ratio in” and “morph the ratio out”
○ Knee affects the threshold zone, makes a zone for the ratio, and the attack and release
are affected by that
● Keeps the sounds more even, smooth, safer, transparent (careful it’s not boring)
● Great on vocals to even out a dynamic performance
Range / Depth - effectively compressed the medium dynamics of the
sound (not the louds like regular compression)

● Limits the extent of the compression. 0 range is essentially bypassed

● Range backs off the ratio the closer the signal gets to the limit of where the
range is set to.
○ Sounds more transparent, retains more dynamic
○ Allows you to set a much lower threshold. Similar effect to soft knee
● Sounds great as gentle bus compression
● Range is rare, but you can find it on:
○ Glue compressor (cytomic), Fab Filter C2, ProTools Channel Strip &
Pro Compressor
Bottom Up & Parallel Compression
● Real Bottom up Compression is rare
○ waves MV2 and MAX Vol do it
○ Ableton Multiband Dynamics
○ OTT
● Parallel Compression / a.k.a. New York Style Compression- Add and Eq
smile on the parallel compressed signal
○ The smile adds “hype”
● Retains the peaks (punch) and brings up the body (subtleties) (rms) of the
sound
● Is the shit!!!
○ You can retain peaks with a quick attack
○ It sound more transparent than just top down
The result of parallel compression is
bottom up
DRY Compressed =bottom up

+ +

+ +

Try a smile eq (boosted lows

and hi’s and ducted mid
3 ways to squish the dynamic range
Top down (most common)

Squish the middle dynamics with “Range”

Bottom Up (parallel compression)

Peak Vs RMS- crappy thing is most compressors aren’t labeled peak or rms….????

RMS: Root Means Squared (Average) ~ 0.707…..

● Loudness perception is based on RMS levels
● Often used for bus compression (on the master)
● Nice for maintaining transients and evening our average volume
● Opto & Vari Mu Compressors are a form RMS compression
● RMS ironically is often more punchy than peak compression because it is too laggy to catch the peaks
Peak (necessary for limiting)
● what kills the dynamic range (make the mix inefficient)
● Good for controlling the peaks
Crest Factor: Peak to RMS Ratio: Difference in distance between peaks and average
- When mastering don’t let the peak to rms ratio exceed 6dB
Types of Compressors
VCA,Opto,Variable Mu, FET
● VCA stands for “Voltage Controlled Amplifier” and its compression behaviour is based on PEAK
● Opto (rms like) one will see an averaged over time version of it.This makes the opto compression much less
sensitive to transients, peaks and sudden spikes. For this reason, much higher ratios can be used.
● Variable Mu (rms like) compressors (tube) produces incredibly smooth compression. It’s transfer curve is far
from being linear. The actual ratio increases with gain reduction. That means that louder a transient is, the
harder it is going to be compressed. Another characteristic of this type of compression is the time constants. It
simply doesn’t respond as fast and impulsively as a VCA or FET. The tube compressor takes its time and never
over-react. It has this ability to glue a mix together like no other type of compressor because of that.
● FET :(peak) The slowest attack time available on the FET is usually faster than the fastest attack time on a
variable mu! Yet, it’s far from being as transparent as a VCA. It definitely has more character. Usually, when you
use a FET, you want to hear it working, because it’s a sound we all like. (Think of a typical rock snare).

http://quantum-music.ca/wordpress/index.php/2015/07/29/vca-opto-vari-mu-fet-compressors-when-to-use-which/
Peak Limiters
● Ratio of 10:1(soft limiting) or higher (infinity : 1 hard limiting)
○ 0 attack time
● useful for catching inefficient transients
● Look ahead: for pre-recorded tracks the compressor can react to a transient before it gets there

Maximizer - Specialized Mastering limiters

● peak limiter with auto gain compensation for mastering
● Fab Filter Pro L2, Limitless, Sonnox Limiter, Ozone 8 are the most competitive digital limiters
going
● Maximizer can be used as Subgroup Limiting but make sure they don’t take too much cpu and
delay compensation :Waves L2 is a good fit, Ableton’s limiter, Maxim is best for that and not for
mastering.
Artifacts of Compression
● Analogue modeled compressors listen for
the “tone” not just dynamic
○ caused by amplitude modulations
○ also can be from analogue modeling
○ can be very desirable or not depending what you
need
● Often eq is used to compensate for the tone
change that a compressor induces
Order of Compression EQ
● Additive EQ before a compressor has the compressor fight the boost and
squashes it back in
● Additive Eq Before Comp = Blending
● Additive Eq After Comp = Distinction (pops)
● Subtractive EQ Before a compressor (the compressor won’t fight to subdue
those freq
Sidechain Compression
1. Slotting: to fix masking... Ducking one sound from another sounds signal
● make one sound dominate the other
● Multiband side-chaining
2. To Help Dynamic Efficiency - when one thing is masking another you can take the other thing away without
noticing its gone,,, thus you have more dynamic room to play with
3. As a musical tool creating “feel & bounce” (dramatic side chain)
● typical dance 4 on the floor kick side chain making the mix “bounce off the kick”
4. Groove chemistry - real musicians play off of each other. Sidechain can give the effect to call answer groove / feel
5. To create inner dynamics
● Creative side chain ideas
○ **** Focus Mixing - can use SC compression in place of some automation for focus mixing
● vocal with a delay (or reverb) send - put a sidechain compressor after the delay and sidechain from the
dry vocal
● Side chain a lead guitar from a vocal,......
● Sidechain a group of sounds: Route or “send” tracks to an aux and sidechain the aux -

***Kick is the most common side chain source

Soft Clip - Black Magic!!!!
● Compression thru clipping
● don’t get compression Amp modulation artifacts… but you could get
distortion
● Cytomic Glue, T-Racks BrickWall limiter (set to clip), ozone maximizer
IRCIII (clipping mode)
● Link to free soft-clip plugins
○ https://www.youtube.com/watch?v=xcUL4I4Tn_0
● THERE IS NO COMPRESSION FFFFFFFUUP artifacts… but its distorts
when you hit it too hard
Multi-band Compression
● Crossovers
● Range
○ is sometimes reversible to make it an upwards expander (transient shaper)
● Main application
○ Being able to compress only certain frequency ranges
○ Control a sound,,, because it levels out the frequencies inherently
■ Flattens the overall dynamics and frequencies
■ Or target frequencies…. And instead of eq where you turn it and everthing below it
down.. You can just target dynamics of certain frequencies
○ Be careful that it doesn’t make your sound boring and static
● multiband sidechain **** really transparent side chain
○ Fabfilter Pro MB
○ Waves C6 Multiband sidechain
○ Pro-tools Pro Multiband
● Flip the range to become an expander!!!!! (make the louds, louder- careful of efficiency)
● Choose Cross over points based on similar dynamic content in the freq spectrum
Over - compression
● Ear Fatigue
● Hypercompression - Sausage
● Loudness war
○ Would you rather something be loud and annoying
or not as loud and makes you want to turn it up.
○ On a master you are in danger of over-compression
when you have less than 6dB “crest factor” distance
between your peak and rms (peak to RMS ratio)
ADVANCE DYNAMICS!
Internal Frequency Side Chain :
● We can EQ the signal going into the threshold and change what
frequencies the compressor dominantly reacts to.
● Most Common Internal SC is a HPF which let the lows pop through
○ Works great for master bus compression that doesn’t lose the low end
punch
● Can you turn a compressor into a De-esser?
● Thrust: analogue term for the signal with a HPF which kills the bass going
to the threshold. In turn the compressor doesn’t react to the lows as much
and the bass isn’t a subdued
L/R Link
● Unlink (independant) Separates the stereo compressor thresholds so that each side reacts
independently… but can enhance / smear your stereo image. Careful unlinking on the master
bus
○ FYI any stereo processor is actually 2 independent processors but the controls are linked
○ Protools pro tip: All Multi-mono plugins have a unlink button on the top right where you
can select and adjust the left plugin processor independent of the right processor
○ Ableton pro tip: Make an audio effects rack and put a utility device in a chain and select
left and pan it left in the utility. Duplicate the chain and in the duplicate select right and pan
right in the utility. That gives you an independant left and right chain for processing
■ ****Ableton’s Eq has a L/R Eq option built in
○ Problem with unlinking: It can smear mono sounds in your stereo image on a
compressor..ie when a sound panned to the right triggers the right compressor any mono
sound that is in the right will get ducked to meaning the mono sound will have more
volume in the left side at that moment.
○ Benefit of unlinking:
■ 1. More efficient: Why compress the side of the signal that isn’t loud?
Multiband Compression Vs Dynamic
EQ
● Multi-band compressors/expanders and dynamic EQs work in a similar way. They both split the input signal into multiple parallel paths.
Each path is filtered to restrict its frequency range, then sent into a compressor or expander.
● A dynamic EQ applies the gain reduction/expansion to the gain parameter of a parametric equaliser which processes the original input
signal.
● Multi-band compressor/expanders apply their gain reduction directly to each filtered signal, then combine them to reconstruct the
original wide-band input signal.
● This reconstruction approach has some downsides:
● Static phase shift
● When the band-limited signals are combined to reconstruct the input signal, a static phase shift is present at their cross-over point. The
band-limiting can be achieved using linear-phase filters, however these incur high latency and can degrade transient response at lower
frequencies.
● Spectral shape
● Multi-band processors typically afford less precise frequency adjustment than a parametric equaliser. If steeper filtes are used to
provide greater precision, phase shift at the crossover points increases.
● Bands cannot overlap
● Some multi-band processors get around the phase response issue, but are still left with a problem – the bands cannot be overlapped!
This can pose significant problems when a combination of gentle character modification and precise correction is required.
● But Multi-band processors do have an advantage over most dynamic EQs. As a boost or cut increases, the width reduces. This is
similar to the proportional Q response or gain / Q depenency of many well loved analogue EQ processors.
● When using an EQ which does not have this gain / Q dependent response, gain changes often require Q changes to maintain the
desired effect. Of course, this is highly undesirable if the gain is being modulated dynamically! See Gain-Q Dependency for details.
Dynamic EQ
● Basically Multiband compression without compression
artifacts.
○ Sounds more transparent than multiband
● Doesn’t use filters (cross-overs)- rather regular EQ shapes
which are less drastic than filters
● Plugins:
○ Izotope Ozone 8 Dynamic EQ

○ Izotope Neutron EQ (has a dynamic option ** with Sidechain)

○ Melda MDynamicEQ (has Sidechain)

Transient Designers
● Ableton - Drum Bus
○ Transients knob
■ Set to the left turns down the sustain
■ Set to the right turns up sustains
● Izotope - Neutron 2 has a multiband transient shaper
● Sonnox - Envolution
○ https://www.sonnox.com/plugin/envolution
 Spectral
Compression
● Melda spectral dynamics
○ Great for very advance side chain.. Take out the exact
frequencies that are in the side chain source