EQ, Compression & Reverb
Technique : Effects / Processing
This month, we tame a strange space, record a
sitar and dispense a little wisdom on EQ, compression and reverb, as we
sort out Dominik Johnson’s home studio.
Dominik Johnson plays his sitar, as Paul White searches for a good mic position.
Dominik Johnson
studies world music at Leeds University, and plays all manner of plucked
stringed instruments to a very high standard. His compact Manchester
apartment is crammed with classical guitars, banjos, bouzoukis,
mandolins, a renaissance lute, a tenor mandola, and even a sitar.
However, his attempts to record these instruments at home had been
compromised by the rather odd acoustics of his living space, and by the
layout of his budget recording setup, which comprises Logic Pro running
on a small Macbook laptop computer, an Edirol UA25EX interface and a
pair of tiny Alesis M1 active USB desktop speakers. He’d made a good
choice of microphones (a pair of Rode NT55s, to allow for stereo
recording), but at the time of our visit the lack of acoustic treatment
was working against him.
Starting Setup
Dominik’s flat is a nice, large space with high ceilings, but it was completely untreated when the team arrived.
We arrived to find that Dominik had enlisted the help
of his friend Joel Davis, who is at the University of Huddersfield
studying Music Technology. This turned out to be useful, as Joel is very
tall and Dominik didn’t have anything rigid to climb on when we were
fixing things up!
The studio apartment is a
rectangular room, around four metres wide and 10 metres long, with an
enclosed ‘island’ in the centre serving as kitchen, bathroom and
storage. This stops around 80cm short of the main room’s unusually high
ceiling (3.2 metres), leaving some storage space on top, which is
accessible by means of a ladder but also allows sound to pass over the
top. Essentially, this island divides the apartment into two roughly
square spaces, separated by two short corridors on either side (flanking
the island) — which unfortunately means that any sound created in one
space sets off resonances in the other.
Dominik
had his system set up on a small computer desk backed right up against
a side wall. Playing some known test material revealed that he had a
huge room-mode resonance set off by the bottom ‘A’ of his guitar.
Speakers Corner
An unusual open-plan layout with an ‘island’ in the middle caused problem resonances.
Because this was a rented apartment, we couldn’t do
anything non-reversible, and there was no budget for bulky bass traps
(although there was space on top of the island that could be used), so
our first step was to try moving the desk, to find a place where the
bass end would be more even to start with. Dominik’s little Alesis
monitors have rear ports, and his initial setup placed them very close
to the wall. That being far from ideal, we experimented by moving the
desk away from the wall, and then over to the corner of the room nearest
the glass windows and doors leading to the balcony. By arranging the
desk at an angle, we got the speakers further from the walls, and helped
reduce side-wall reflections reaching the listening position. We also
placed Auralex MoPad foam pads under the speakers (Dominik had already
bought some but wasn’t actually using them!). Because the speakers were
so small, only one half-width pad was used under each speaker, set up
‘reversed’ — so that the speaker was angled up slightly, towards the
listening position, and the foam that extended in front of it would help
a little to reduce reflections from the desk surface. Using both foam
pads would increase the area of foam and reduce reflections even more,
but the desk was quite small and the best working compromise, leaving
enough space for the laptop, was to use a single pad.
Packing
duvets into the storage area above the central ‘island’ room tackled
the unwanted resonances, as well as providing some useful bass trapping.
Further listening with test recordings and a
‘staircase’ of sine-wave semitones (from Logic’s EXS24) showed a
noticeable improvement, although the room mode was still there.
Examining the speakers showed that Dominik had engaged the bass-boost
switch: flipping this back to the normal setting cleaned up the low end
further, as did pulling the speakers forward on the desk so that they
were further from the walls, and closer to the listening position.
As
an experiment, we improvised some foam plugs to put in the bass ports,
which further tightened the low end, both with and without the
bass-boost switch engaged. My preference was to leave the bass-boost off
and insert the foam plugs halfway, which felt like a good compromise
between a clean bass and having enough low end: tiny monitors like
these are never going to be great at deep bass frequencies, but as
Dominik was recording non-bass instruments, the arrangement was
workable. Even so, we did recommend he buy some good-quality headphones
to use as a secondary reference.
All Of A Flutter
With plastered concrete walls, the apartment
exhibited very obvious and distinct reverb, and produced some of the
most dramatic flutter echoes we’d heard to date, so our priority was to
dry this up as much as possible. We’d brought half a dozen sheets of
two-inch Auralex foam with us, but this wasn’t going to stop sound
leaking around and over the centre island, and exciting the bedroom
space at the other end.
After considering the
problem over coffee and chocolate biscuits, we set out for the nearest
DIY store to buy self-adhesive plastic hooks, nylon cord, impact
adhesive and two king-size polyester duvets. The plan was to glue wooden
strips to the top rear edge of the foam panels using contact adhesive,
and then hang these, picture-style, on a self-adhesive plastic hook
stuck directly to the wall. We’d use the duvets to screen off the space
over the centre island by hanging them over a nylon line fixed to
plastic hooks and going all the way between the side-walls, via a
convenient centre support where we could put more hooks.
In
most respects this worked well, although the hot weather made all the
plastic hooks gradually peel off the walls under the very modest weight
of the foam panels, so we had to provide a helping hand using some of
our contact adhesive. We ended up with one foam panel each side of the
listening position (one on the wall and the other fixed to the metal
door/window frame), three on the wall behind the sofa, which was roughly
behind the mixing position, and one more set horizontally, as high as
we could place it to damp down some of the flutter echo that developed
high up in the ceiling space.
With the panels
and duvets in place, the liveness of the space was significantly reduced
but was still enough to flatter acoustic instruments, especially if
recorded with the floor rugs lifted to reveal the nice parquet floor.
Test recordings sounded cleaner, with better imaging, and although the
‘bottom A’ bass hump was still present to some degree, it was far more
benign now than it had been at the outset.
Recording System
When we arrived, Dominik had his small monitor speakers placed directly on the surface of the desk...
Having addressed the acoustics issues as best we
could, we turned our attention to the recording system. Dominik had been
plagued by high latency, and couldn’t get the direct monitoring on his
Edirol interface to work. We were initially concerned that as the
monitors and interface were both connected via USB, the setup might
require an aggregate device driver to allow both to work at once (this
approach can cause increased latency), so we wired the speakers into the
analogue outs of the interface, using a couple of jack cables, so that
the interface could use its own drivers.
A
quick look at Logic’s Audio Preferences showed that Dominik had it set
to a 1024-sample buffer size, and we reduced this to 128 to cut down
the latency. Most players can live with this tiny delay even if they
don’t have zero-latency hardware source monitoring, although we were
determined to find out what had been going awry with the Edirol. As it
happened, nothing was wrong, but Dominik hadn’t realised that you need
to turn down the track monitor level in Logic as you record. If you
don’t do this, you hear both the direct monitor signal and a slightly
delayed version coming back from the computer — and his initial
1024-sample buffer size would have created a very distinct slapback
echo.
We went back to the Audio Preferences,
ticked the Independent Monitor Levels box, and showed Dominik that he
could now turn down the fader level while in record mode, and it would
automatically spring back to its playback level when he left record.
Keeping the channel fader down while recording prevents the current
performance from being monitored in the DAW output while tracking, so
that only the direct source and existing DAW tracks are heard. We made
some test recordings in this way after selecting source monitoring on
the interface, and confirmed that all was now well, with no unwanted
echo.
Mic Technique
...so
we isolated the speakers from the desk surface, before putting foam in
the ports to get the most appropriate sound for Dominik’s music and
room.
For recording vocals, we took along a Reflexion
Filter kindly provided by Sonic Distribution. This would also be useful
for placing behind an omnidirectional mic when recording Dominik’s
instruments. As usual, we rejigged the mounting hardware to make it sit
closer to the centre of gravity of a typical mic stand. We also
stressed that for the best recorded vocal result, some absorbing
material should be placed behind the singer to intercept reflections
that would otherwise bounce back over their shoulders, into the
microphone, from hard surfaces behind them. Here, Dominik could use one
of the foam panels or an additional duvet. He asked what was the best
place in the room to stand for vocal recording, and we explained that as
long as he avoided the exact centre of the room, or positions too close
to the walls, the results should be OK more or less anywhere.
Dominik
was also keen to see how we’d mic the various instruments that he
plays, so we did some test recordings to see what worked best, using the
Rode NT55 with its omni-pattern capsule fitted. Acoustic instruments
always seem to sound more natural when recorded with an omni mic and, as
explained above, if the room is intruding too much you can screen off
the sides and rear of the mic using a Reflexion Filter or improvised
screens. Because omni mics have a much better-balanced off-axis
response than cardioids, positioning also seems less critical.
Above
right: To the rear of the mix position, we decided to hang three
Auralex acoustic foam panels to cut down on unwanted reflections. Right:
we also placed foam panels at the mirror points to create a tighter
stereo image from the monitoring setup.
We explained a simple ‘universal’ miking rule that
often yields good results when you’re recording an instrument for the
first time: you measure the longest part of the instrument that creates
the bulk of the sound (such as a banjo body diameter), and then place
the mic no closer than this distance, so that it has a chance to pick
up sound from all parts of the instrument rather than from just one
spot. Then you need to find the placement that gives the most natural
sound, which is best done by wearing headphones as you move the
microphone (or move the instrument relative to the mic). Most recording
textbooks give you typical mic placements, but every situation
(instrument, player and room) is different, and it always pays to search
for that sweet spot.
When recording the sitar, I
tried a mic position very close to the floor, to make use of the
boundary effect and boost the low end slightly. The Sitar is an unusual
instrument in which some sound is also produced by the (hollow) neck
and, where attached, by a secondary resonator gourd at the back of the
neck. We found that placing the mic close to the floor and aimed at the
bridge from around 18 inches away produced a well-balanced sound with
the required degree of warmth. The banjo was altogether simpler, as it
seemed to produce a consistent tonal balance over quite a wide angle,
so the distance between the mic and instrument was actually more
important than the mic position. We also experimented by having the mic
look over the player’s shoulder, working on the principle that if it
sounds OK to the performer, it will also sound good to the mic. This can
yield excellent results, but as we found here, the mic picks up more of
the player’s breathing (or humming, in Dominik’s case).
EQ, Compression & Reverb
As Dominik was a relative newcomer to Logic, Paul
created a ‘starting’ song template that included screensets for the
Arrange and Mixer pages, as well as a reverb set up on a post-fade
send. This lets you get up and running without having to set up a
project from scratch every time. He also gave a crash course on
compression, the most important lesson being that if you choose a
preset compressor plug-in setting, you still need to adjust the
threshold controls so that the gain-reduction meter is showing activity
on the loudest peaks: if the gain-reduction meter isn’t showing
anything, then you’re getting no compression. For gentle instrument
compression, a peak gain-reduction of 4-6dB is probably fine, although
more aggressive sound sources may need larger amounts of gain reduction.
Ultimately, of course, your ear has to be the judge.
This
led to the question of the best order in which to place insert
plug-ins. On one very transient-rich, percussive part that required very
assertive level control, Dominik had placed a reverb plug-in before
the compressor, creating a splashy, unpleasant sound that got in the
way of the instrument. Normally, you’d use reverbs via post-fade sends,
but on small projects there’s no problem using them as inserts. In this
case, we reversed the order, so that the compressor controlled the peaks
before the sound hit the reverb, and the result was less ‘splashy’.
EQ,
on the other hand, can go before or after compression, depending on the
effect you’re after: for natural-sounding tonal tweaking, I tend to
favour filters first (to remove unwanted out-of-band noises such as
rumbles), then the compressor, then creative, tonal-shaping EQ if it’s
required. Placed the other way around, the compressor will react more
strongly to any frequencies boosted by the EQ, and tend not to control
frequencies you’ve cut. We also demonstrated the way in which quite
severe narrow EQ cuts can deal with resonances, but narrow EQ boosts
rarely sound natural. For natural results, use as little EQ as you can,
and boost with fairly wide (low) ‘Q’ settings.
Our
final venture was to explore some Logic Space Designer reverbs to see
what was best for acoustic instruments. While the temptation is always
to go for a concert hall — or something equally familiar sounding — you
can get great results from less obvious spaces, such as tiled rooms,
parking garages, or even woodland and other outdoor spaces. Impulse
responses with complex early reflections and low-level tails often sound
interesting, but don’t over-use them: a 20 percent mix level is usually
a good starting point. About 60-90ms of pre-delay can add to the sense
of space, while rolling off some low end below 200Hz can clean up an
otherwise messy reverb.
As rush hour was
approaching, we decided to make our getaway before any more of those
B&Q hooks peeled off the walls — but we left Dominik with the
contact adhesive just in case!
No comments:
Post a Comment