|
Controlling Bass Response in Control Rooms
An article by Brian Gaylor
(appeared in Audio Media March 2001 UK)
WAVE WALLS - Part 1
Two years ago I finished the design and construction of the control
room at Stanley House Studios in west London. There was a problem
with the bass end acoustics which I know to be fairly common in
other rooms. The bass varied as you moved about and most importantly
there was a lack of bass at certain frequencies at the listening
position seated at the mixing console. After some experimentation,
I believe that I have come up with an easy solution which may benefit
others with similarly shaped rooms. I am not aware of the method
I employed being used now or in the past, but I could be wrong here.
I would be surprised if the concept is novel as it seems too basic
to be overlooked. However it seems to offer a solution that could
be tried in other rooms without too much expense or effort.
The Problem
Inside any building shell the internal dimensions of the room will
allow sound waves to resonate at predictable frequencies. A standing
wave is created which causes areas of maximum and minimum pressure
variations at set points in the room, making bass notes appear loud
in some areas and weak in others. Whereas the higher frequencies
are easily deflected and absorbed the lower ones, especially near
the fundamental carry a relatively large amount of energy at longer
wavelengths, and are difficult to control. The principal dimension
of concern is the front (speaker) wall and back wall.
The Room
Fig.1 shows the microphone output at the listening position at the
mixing desk when the speakers generate pink noise, before the room
was modified. It should be noted that on my spectrum analyser pink
noise slopes off at 3dB per octave. Peaks at 50Hz, 78Hz, 112Hz and
125Hz with troughs at 69Hz, 83Hz and 100Hz show the problem room
resonances.
There is a further region between 150Hz and 180Hz which is obviously
another poor area. This was improved by building screens to block
off the back of the desk. The back of the J series SSL offers a
large surface area just a short distance from the main monitors.
It is inevitable that if the speaker wall is non absorbing, reflections
between desk and wall will cause problems. Absorbent screens help
dampen these resonances and prevent reflected waves passing beneath
the console.
At Stanley House the back wall is totally lined with membrane absorbers
tuned to the fundamental of about 50Hz. This type of absorber is
constructed of a limp layer of rubberised matting pinned over an
air tight cavity, like a soggy bass drum. The mass of the material
and the depth of air behind define the frequency of operation, like
a weight bouncing on a spring. If the material absorbs energy as
it moves, and the air vibration behind it is damped then air movement
at the face of the membrane will lose energy most at the oscillating
frequency of the system. Lining the cavity with fibrous material
both aids the damping of the air movement and detunes the absorber
such that it should work over a slightly broader frequency range.
The spectrum analysis of a microphone placed inside the cavity in
response to pink noise from the room showed a peak at 50Hz tailing
off at about 10dB per octave. If the absorber is working properly
then this graph provides an indirect suggestion that all is working
as well as could be expected.
At a short distance from the front of the membranes hanging absorbers
complement the deadening of reflections. These consist of a layer
of roofing felt with loft insulation stuck to either side freely
hanging from a top fixing. Air movement due to sound waves is impeded
by the inertia of the felt matt and the damping of the fibrous material.
However the graph of the mic output at the desk still shows a peak
at the fundamental frequency. A general hole from there to over
100Hz causes the listener at the desk to miss the 'kick' from bass
instruments, implying still not enough absorption to control standing
waves at these frequencies.
Fig.1 Original
room configuration: Amplitude plot per frequency.
Part 2 >
Part 3 >>
Part 4 >>> |
