Choosing The Right Compressor For The Job
Technique : Effects / Processing
Now that we have digital models and dynamic
convolution technology, we’re spoilt for choice when it comes to classic
compressor flavours. But whether hardware or software, which is right
for which job?
Whether
they truly recreate the real thing or just give you a hint of their
flavour, there’s now a bewildering number of classic compressor
emulations in software.
At its most basic
level, compression can be seen merely as a labour-saving solution to
mix-balance problems: if you control the levels of an overly dynamic
instrument, it becomes easier to find a consistent fader setting for
it. While the practicalities of setting up compressors can take a
little time to get your head around to start with, there is still a
certain logic behind the process, as I’ve tried to demonstrate in
preceding pages of this issue of SOS.
As you venture beyond general-purpose balance
discussions, though, the subject of compression becomes shrouded in
mystique, particularly with respect to the qualities and quirks of the
most famous hardware compressors, many of which are prized as much for
their tonal qualities and audio side-effects as for their ability to
control dynamic range. For years, such debate remained the preserve of
those few lucky souls in professional studios who had the opportunity to
compare a range of such rare and expensive hardware, but recent
advances in the DSP emulation of those same units has brought the topic
slap-bang into the mainstream.
That’s great news
in some ways, but it also presents many recording musicians with a
significant problem. Seasoned studio jockeys can bring a wealth of past
experience to bear on the task of choosing the right model of
compressor for any task, but those without the benefit of such an
apprenticeship can now find themselves adrift in a sea of complex
possibilities, with little but guesswork and experimentation to assist
them.
In this article, then, I’ll try to shed
light on which classic compressors tend to be used in which processing
situations, based on my ongoing research into the techniques of ‘A-list’
engineers and producers (see the box at the end of the article for
their backgrounds). I’ll also look at a few of the design features and
side-effects that adapt the different processors to these tasks, as well
as talking about some of the ways they’re actually combined in
real-world sessions.
Classic Vocal Compressors
Let’s
start where the money’s at in most productions: those all-important
vocal parts. There’s no shortage of conflicting opinions from those in
the know, but certain compressors do seem to crop up a great deal more
than others.
Where transparent level control is
required, the Dbx marque is a popular choice, with engineers such as
Steve Churchyard, Humberto Gatica, Eddie Kramer, Ed Cherny, and Mike
Clink all specifically name-checking the Dbx 160-series processors. The
clean, predictable response of the low-distortion, Blackmer VCA
gain-reduction element is clearly part of the appeal: “A Dbx 160 works
really well if you just want to protect the tape,” comments Ed Cherny,
for example. The majority of SOS readers working with high-headroom,
24-bit A-D conversion will be less concerned than analogue engineers
about fitting their recordings into the dynamic range of the recording
medium, but cleaner-sounding compressors can be just as useful while
mixing, especially on instruments with complex harmonics (12-string
guitars, pianos and harpsichords, for instance), which can respond badly
to distortion.
However, the compressor’s speed,
achieved through a feed-forward circuit design, also seems to be a
key feature for many engineers, with Mike Clink referring to his ‘fast
and quiet’ Dbx 160s and Steve Churchyard commenting that “typically the
Dbx 160 seems a little faster than 1176s or Fairchilds, so if I need
something with a faster attack time, I’ll use that instead.” In fact,
so fast can this compressor react that it can easily generate distortion
as it tracks individual waveform cycles, especially on later models
with variable attack and release controls, so you do need to listen
carefully for this. Whereas the original Dbx 160 was a hard-knee
device, the later 165 retained the same basic design while adding in an
‘Over Easy’ soft-knee compression curve, and this markedly improves its
capabilities with regards to unobtrusive control of extreme level
fluctuations. In the words of Mark Nevers: “You can squash the shit out
of it, but you don’t hear it.”
Despite their
reputation for cleannness, later Dbx models do still offer subtle colour
and flattery to the vocal tone, by virtue of their transformer-coupled
inputs and outputs. High-spec transformers can contribute a kind of
euphonic, low-level, frequency-dependent distortion that regularly
elicits descriptions such as ‘warm’ and ‘fat’, and which is often called
upon to make vocals seem somehow ‘larger than life’. The combination of
VCA gain-control and audio transformers can also be found in
Focusrite’s Red 3 and Red 7 units, which also exhibit this kind of
subtle warmth and are used for vocals by producers such as Steve Power
and Stephen Hague.
The ability to heap
gain-reduction onto a vocal without audible side-effects is associated
with another well-known unit, the Teletronix LA2A (now re-issued by
Universal Audio), which achieves its low-distortion soft-knee
characteristics, as well as a programme-dependent release time, using a
unique electro-optical gain-reduction element. This unit has been in
regular use since the ’50s, and could lay a strong claim to being the
most popular vocal compressor of all. Where the Dbx’s feed-forward VCA
and solid-state circuitry are fast, controlled, and clean, the LA2A’s
electro-optical feedback design works more slowly and in a non-linear
fashion, which nonetheless produces very musical results, preserving the
impression of performance dynamics despite quite extreme level
management. While it, too features audio transformers, it also rounds
out the tone via its internal valve amplification stages. “I added the
LA2A for warmth,” comments Serge Tsai, for example, echoing similar
‘gentler and warmer’ epithets from producer Glen Ballard.
The
hugely popular Teletronix LA2A (now built by Universal Audio, and one
of the first units modelled for their UAD platform) is great at
transparent gain reduction, but can also yield the kind of vocal
overdrive favoured by artists like Jack White of the White Stripes.
However, a few high-profile engineers have also learned to take
advantage of the extra attitude on offer from this king of smoothness as
you push it outside its comfort zone. “Sometimes the LA2A works really
well [on vocals],” says Joe Barresi. “If you’re looking for more dirt
you can turn a tube compressor up a little more.” Joe Chiccarelli
recalls a similar tactic when producing the White Stripes: “Jack
[White] always wanted more distortion on the vocals... When it came to
the mix, Jack wanted still more edge, so I overloaded an LA2A
compressor... This meant that I was getting the distortion from the
last tube stage of the compressor, which creates a really beautiful
distortion.”
The LA2A’s winning combination of
electro-optical compression, transformers and valve amplification can
also be found in a couple of other units well-respected for vocal use,
Tube Tech’s CL1B and Avalon’s VT737SP, so if you like what an LA2A can
do for your vocals you may find that these models provide useful
alternative colours. For faster, cleaner, more VCA-style control, and a
slightly crisper transformer-free sound, Summit’s TLA100A and DCL200
are also popular with audiophile engineers such as Al Schmitt, Jon Gass
and Tony Maserati. “I found I could use it on vocals that I needed to
squish a fair amount but didn’t want to lose all the nice top I got
off the microphone,” explains Joe Chicarelli in an interview on www.summitaudio.com.
If
you like the character of the LA2A’s non-linear optical gain element,
but the valve distortion artifacts aren’t appropriate, then the slightly
faster, valveless version of the same basic design, the Teletronix
LA3A, provides another option favoured by two of the most respected mix
engineers on the planet: Bob Clearmountain and Tom Lord-Alge. “To make a
vocal command attention,” says Lord-Alge, “I’ll put it through an LA3A
and maybe pummel it with 20dB of compression, so the meter is pinned
down. If the beginnings of the words then have too much attack, I’ll put
the vocals through an SSL compressor with a really fast attack, to
take off or smooth out the extra attack that the LA3A adds.”
Further Vocal Coloration & Enhancement
Beloved of Beatles engineer Geoff Emerick is the rare and mind-bogglingly expensive
Many
engineers use the Fairchild 670 for the subtle sonic enhancement of its
circuitry as much as they do for gain-reduction. Although the original
is now prohibitively expensive for all but the most maniacal of
collectors, there are some excellent software emulations, such as this
one from Universal Audio.
Fairchild 670 stereo valve limiter (and its pretty much identical 660
mono sibling), and his penchant is shared by other high-profile names
such as Roy Thomas Baker, Steve Power and Tom Elmhirst. Where the LA2A
will give you fairly slow attack times, the Fairchild has a much
faster, sub-millisecond, onset, as well as a different type of
extremely soft-knee compression curve, courtesy of its valve-based
‘vari-mu’ gain-control circuit — the compression curve is unusual in
that the ratio just continues increasing as the input signal level gets
higher.
While these characteristics mean that
the 670 can control levels very effectively, probably the main reason
that top engineers call on it is actually for the sonic enhancement of
its circuitry, which incorporates a total of 20 valves and 14
transformers! Geoff Emerick always put the Beatles’ vocals through it
for precisely this reason: “Just the sound of the amplifier, even if you
didn’t do any limiting, just added a certain presence.” Steve
Churchyard also talks about using the 670 “more for the sound than for
the compression,” a sentiment echoed by Tom Elmhirst in SOS August
2007.
The Fairchild 670’s extreme rarity,
combined with its ability to make notes, in the words of Jack Joseph
Puig, “long and lush and beautiful” has spawned not only fastidious
modern hardware replicas (such as Anthony DeMaria’s ADL670 and the
Analogue Tube AT101) and heavily-inspired variations (EAR’s 660,
Fairman’s TMC, Manley’s Vari-mu, and Pendulum Audio’s ES8, for example),
but also a slew of software emulations — so you don’t need to look too
far these days to find something akin to the Fairchild sound. If you
like what you hear, you might also want to hunt down emulations of some
other vintage valve-based vari-mu designs, such as the Altec 436,
Collins 26U1 or 356E1, Federal AM864U, Gates Sta-Level or Level Devil,
RCA BA6A, and Universal Audio 175 or 176. Fairchildren they aren’t, but
each is nonetheless packed full of its own unique vintage valve
character, and is frequently referred to reverentially by mix engineers.
Michael Brauer, for example, used a Gates Sta-Level and a Federal
AM864U alongside the EAR 660 for his mix of Coldplay’s recent hit
‘Violet Hill’.
The Urei 1176 is famous for adding a sense of edge and excitement to vocals.
The final classic vocal-processing contender is Urei’s 1176 (and the
same marque’s near-identical stereo 1178). Whereas the 670’s valves are
smooth and warm, the 1176’s equally fast FET-based design brightens the
tone and adds extra edge and excitement in a way that has made it a
fixture on vocals for many top engineers for a half-century. People
differ in terms of how exactly they use it, particularly with regards to
the choice of ratio setting, so it pays to experiment with this aspect
of any emulation — and don’t forget to try the unusual ‘all buttons’
mode if it’s on offer, as it is, for example, in Universal Audio’s
software 1176LN for the UAD platform. The 1176’s enormous popularity has
given rise to legions of impersonators, including hardware units such
as Universal Audio’s re-issued 1176 and Purple Audio’s MC76, as well as
myriad software emulations — so this should be one of the easiest
compressor sounds to get hold of, even if you’re working on a budget.
That
pretty much wraps up the vintage units, but we can’t leave the topic of
vocal compression without touching on one very popular modern processor
too: the Empirical Labs EL8 Distressor. This phenomenally versatile,
digitally-controlled, analogue compressor, introduced in 1995, was
designed partly to emulate the characteristics of different classic
compressor actions, including those of all four principal units we’ve
discussed so far: Dbx 160, Teletronix LA2A, Fairchild 670 and Urei 1176.
As such, it’s not easy to say exactly what kind of compression you’ll
get if you dial up a Distressor emulation, although a little knowledge
of the way the aforementioned classic compressors work may provide you
with some useful pointers. What is worth a special mention, though, is
that the Distressor has a number of built-in, side-chain EQ options,
one of which emphasises the upper-mid frequencies in the detector
circuit and thereby reduces sibilance problems that can often accompany
heavy compression.
Smash & Grab: Compressors For Drums
It’s
pretty clear, looking at the practice of real engineers, that
fast-acting compressors and limiters are the most popular ones for drum
processing. No prizes, then, for guessing that, of the models we’ve
already encountered, the LA2A doesn’t really get a look-in. Units from
Dbx receive mentions in this role from the likes of Tom Lord-Alge, Jack
Joseph Puig, and Butch Vig, while Glen Ballard singles out the 160X
variant for special praise: “I think the 160X is the best compressor
for drums that I’ve found. You just give it a little bit on the kick
and snare while you’re recording and it brings them into focus. I don’t
think you can find anything better for kick and snare — they’re
amazing! You can really slam them too! I love the fact that they have a
really quick recovery, and you can exaggerate it if you need to.” The
extremely fast attack and release times available on many Dbx units are
particularly worthy of investigation if you find yourself in a situation
where you really need to sculpt the front end of your drums with
precision.
The Fairchild units have been
responsible for much drum-mangling ever since the 670’s prominent use by
Emerick on Ringo Starr’s kit, and since then engineers including Steve
Churchyard, Tom Elmhirst, Manny Marroquin and Elliott Scheiner have all
used the Fairchild brand for drum processing, on account of its
colourful sonics and speed. “The 660 has a nice smooth sound,” comments
Fabian Marasciullo, “and it really controls the peaks.” However, the
670 also has an unusual trick up its sleeve in that it has M/S matrixing
built in, allowing you independently to compress the Middle and Sides
components of a stereo signal. By plugging the 670 into a drum bus and
then compressing the Sides signal more heavily, you can emphasise
and/or pump the room sound without compromising the transient snap of
your main kick and snare sounds.
It’s no
surprise that Urei’s 1176 and 1178 processors also crop up frequently,
with mixing heavyweights such as Chris Lord-Alge, Michael Brauer and
Rich Costey all partial. Despite the reputation these units have for
fast attack, though, Lord-Alge was actually using an 1178 more for its
sonic flavour with slower attack settings when we interviewed him back
in SOS May 2007. Rich Costey’s preference for a parallel processing
setup with this compressor also suggests that the tonality of the Urei,
especially when driven hard, was of more importance than its ability to
catch transients. (See the ‘Parallel Compression Explained’ box for
more information on parallel techniques.)
It’s
to be expected that Empirical Labs’ Distressor also turns up frequently
when drum processing is discussed, but that’s not just because of its
emulative power: it has also earned a reputation of its own,
particularly for its ‘Nuke’ brickwall-limiting mode, which has a
logarithmic release time originally optimised for pulverising room-mic
recordings.
Empirical
Labs’ Distressor is one of a few ‘modern classic’ hardware compressor
designs, and is incredibly versatile. As well as mimicking some
longer-established designs, it brings some of its own special tricks to
the table.
The side-chain high-pass filtering is handy, too, where you want to
compress a drums bus heavily without the kick drum triggering excessive
gain pumping, and the ability to switch different harmonic distortion
flavours in and out of circuit provides a great deal of tonal
flexibility. Perhaps because of the compressor’s enormous power to mould
the envelope and timbre, Michael Brauer, Joe Chiccarelli, Fabian
Marasciullo and Robert Orton all use parallel processing setups to give
extra control over the extent of its contribution to the mix. Take
Marasciullo, for instance: “On drums [the Distressor] gives me an attack
that no other compressor can give me. I will have one drum just raw,
exactly as it was given to me, with just a bit of EQ on it. I will
then mult it into a Distressor and that will give it this extreme
attack, and then I blend it in the way I like it.”
Moving
on from the vocal workhorses, another compressor renowned for drums is
the TG12413, a limiter originally built into each channel of the
Neve-designed custom consoles that EMI studios used for famous records
such as Abbey Road and Dark Side Of The Moon. Although original units
are as rare as hens’ teeth, the widespread availability of Chandler’s
TG1 re-issue (and its more flexible TG12413 Zener sibling) and EMI’s
more recent plug-in emulation has ensured this sound a regular role in
mainstream commercial mixing.
“I like the EMI
compressor for its fast attack,” remarks Joe Barresi, and this is a
recurrent theme in other producers’ descriptions, but the tone of the
TG12413’s unique diode-based gain-reduction circuit is also an important
factor. Although Joe Chiccarelli characterises it as “more of a round,
warm, vintage sound,” pushing harder transforms it into something more
aggressive. Nick Launay: “I don’t like a lot of compressors because I
find they’re not fast enough or they’re a bit too subtle or they make
things sound too nice. I like the stuff that’s got a bit of grit, like
the old EMI compressors, if you can find them.” Michael Brauer concurs,
speaking on the Chandler web site: “When used across the drum room [the
TG12413] brings out an aggressive quality that I’ve not yet heard from
any other compressor. When used as a send, it brings out the punch and
clarity of instruments without affecting the overall level of the stereo
output.”
The
RCA BA6A. Like the Fairchild 670, this compressor is often used as much
for it’s sound as it’s ability to achieve the desired gain reduction.
If you read Will Haas’s article about mix-bus
compression back in SOS May 2008, you’ll already be aware of many of the
industry-standard mix-bus compression choices, but it tuns out that two
of these regularly appear in the role of drum submix processor too: the
API 2500 and the SSL G-series console master compressor. Both of these
feature fairly clean, VCA-based gain reduction and are good at
increasing average levels without trampling on the punch of transients
or robbing too much low end. However, the SSL also has a reputation for
adding a dose of desirable mid-range crunch when more heavily
provoked, and Smart Research’s C2 emulation emphasises this aspect of
the unit further with its additional ‘Crush’ mode.
“[The
C2’s Crush mode] adds field-effect transistor distortion,” explains
Jason Goldstein. “I’m a big fan of using distortion, in small amounts.
Tubes distort more harmonically, whereas transistors and Class-A stuff
are more aggressive, which is why I think many guys still like to mix
on the SSL 4000 — those consoles are always just shy of distorting. It
adds to the overall aggression of the mix.” But that’s not all that
Crush does. According to Tom Elmhirst, it “completely annihilates all
the transients and brings up the air.” Small wonder, then, that
Goldstein, Elmhirst and Rich Costey all take a parallel processing
approach with the C2, for greater control.
Max Bass
Bob Clearmountain.
That the Urei 1176 convincingly topped the polls for
bass processing in my survey of producers might raise an eyebrow, given
what I’ve said about it thus far. After all, fast compression can
introduce serious distortion on slow-moving bass waveforms. However, as
long as the time constants are kept away from the fastest settings (in
other words, somewhat counter-intuitively, away from the
higher-numbered, clockwise end of the control settings on the unit), the
circuitry’s tonal brightening and subtle FET distortion really help the
bass to cut through the mix in a way that other designs can’t match.
The lowest ratio (4:1) is the most common choice, as evidenced by
independent recommendations from Steve Churchyard, Chris Lord-Alge and
Ralph Sutton, although Tom Elmhirst likes the result you get when you
push in the 4:1 and 8:1 buttons together.
Not
far behind the 1176 in popularity is the Teletronix LA2A, numbering
people like Mike Clink, Mark Endert, Tony Maserati and Jim Scott amongst
its friends: the slower-moving, electro-optical design is clearly
better suited to clean level management of low-frequency sounds. The
valve sound, though, is clearly a crucial part of the equation too,
lending thickness and warmth to the tone, as evidenced by the wide range
of other valve units from Altec, Fairchild, Gates and Summit that
turned up in the interviews from time to time. That said, mix engineer
Chuck Ainlay raves about the original Joemeek SC2 solid-state optical
compressor for bass, so that might be worth a look if FET or valve
additives don’t suit your taste.
What About Other Instruments?
Joe Barresi.
I’ve spent a good deal of time looking at compressor
choices for vocals, drums, and bass, and in each of these cases I’ve
been able to distil some degree of consensus from the available
evidence. What I’ve discovered in the process of my research, though, is
that compressor choices for other common instruments, such as guitars
and keyboards, are much less commonly discussed in interview and,
furthermore, that there also seems to be very little agreement between
different producers as to what processors to use on them.
For
example, with electric guitars, aforementioned units from Dbx,
Fairchild, Gates and Urei all turn up sporadically, as well as Neve’s
2254 and 33609
Joe Chicarelli
models (both of which use similar bridged-diode gain-reduction elements
and lots of transformers, to deliver a silky-smooth tone that often
seems to be called on to remedy thinness or abrasiveness in a
recording). However, little justification is given for these processing
decisions, so any attempt at generalisation remains firmly in the realms
of surmise.
Perhaps the lack of information on
processing other instruments simply reflects that vocals, drums and bass
are more important than other elements in modern productions — so by
the time an analogue engineer gets to mixing the brass parts, for
instance, the most desirable compressors will probably already be in
use, and he or she will, in practice, be more concerned with the
compressor settings than the specific choice of unit. Or maybe it’s that
vocals, drums and bass just tend to need the most dynamic-range control
in a typical mix, so the choice of compressor (and indeed the use of
compression at all) becomes less crucial with instruments holding a
more consistent level.
Well
known for bus compression, amongst other things, the Neve 33609, which
offers silky-smooth gain reduction, is equally at home on electric
guitar.
Another argument for the lack of consensus with
regards to compressor choice for things like electric guitars and synths
could simply be that parts like these vary so much in character from
mix to mix. A clean and choppy funk guitar is clearly a very different
beast to a heads-down Emo wall of noise in terms of attributes such as
transient definition, dynamic range and frequency balance, so it makes
sense that the same compressor would not necessarily suit both.
Whatever
the reasons, though, hopefully this article has provided enough of an
insight into the basic characteristics of different classic compressor
models that you can begin to extrapolate towards more educated guesses
when it comes to processing other instruments. The aforementioned funk
part might benefit from the speed of a drum compressor to catch its
transients, for instance, while the Emo landslide might need some of the
1176’s tonal definition to make it bite. Once you get used to the way
in which compression can react on vocals, drums, and bass, you’ll find
that there’s a lot less guesswork involved in processing pianos,
guitars, strings, brass, zither, nose flute, one-handed clapping...
No comments:
Post a Comment