Part I

Before recording
Chapter 1
Acoustics and the recording
venue

Capturing the musicians’ performance in a suitable acoustic space is

central to the art of recording classical music. A good sounding space
will enhance both the recording and the performance itself; playing
music in a rewarding acoustic is a much more enjoyable experience
than playing in a deadened room or a space with distinctive echoes or
other problems. Reverberation from the room enables the players to
hear their own playing better and thus helps with tuning and other
performance attributes. Commercial record companies will spend a
lot of time searching for halls and churches with a suitable acoustic
characteristic, and as a result, many venues are used by multiple
companies and freelancers for making recordings. However, it is not
always possible to pick and choose the venue in a non-commercial
situation, and so we need to consider what can be done in a less than
ideal environment. The use of additional artificial reverb is an
essential skill in classical recording, and this will be considered in
Chapter 17.
1.1 Brief introduction to room acoustics
This book is intended to have a practical approach, and so for detailed
discussion of theoretical acoustics and mathematical modelling of
spaces, you should seek out an acoustics textbook. However, there are
a few basic concepts that are worth including here.
Reverberation is built up from the repeated reflections of sound
waves within a building, and after the very early reflections, it
quickly becomes a blend of thousands of reflections and interactions
that are no longer individually discernible. At every reflection of a
soundwave, some energy will be absorbed and some will be reflected.
This behaviour varies with the surface involved, and most
importantly, it is frequency dependent. In a room where most of the
sound energy is reflected at each interaction with a boundary, the
sound will continue bouncing around the room for a longer time and
produce a longer reverberation time. In a room where most of the
sound energy is absorbed at each boundary interaction, the reverb
will die down very quickly. After the first early reflections in the
room, the reflections quickly multiply to form a characteristic reverb
signature which is determined by a number of factors (see sections
1.1.1–1.1.4).

1.1.1 THE MATERIALS MAKING UP THE

SURFACES IN THE SPACE
Any very smooth, hard surface will reflect more high frequencies
(HF) than a rougher surface, which tends to scatter reflections and
dissipate energy. Stone and glass will generate the most HF
reflections, followed by painted, modern plastered walls and wood
with a hard modern varnish. The more HF reflections there are, the
more HF content will be in the reverb, and the space will have a
bright reverb characteristic. Rougher surfaces such as oiled and
waxed wood finishes will reflect less HF energy at each reflection,
and carpets, curtains and upholstered seating will absorb higher
frequencies very quickly.
Low frequency (LF) energy is absorbed very little by soft
furnishings, but is taken up by structures with low resonances such as
sprung floors and partition walls that are not solid. A lot of the energy
from the LF waves will be taken up by these structures, making them
move slightly, and will then not be reflected to form part of the
reverb. A room with a sprung floor or partitioned walls can tend to
sound bass-light when compared with one that has a solid floor and
walls. An interesting example of a sprung floor was Brangwyn Hall
in Swansea that had a floor with adjustable springiness so that it
could be used as a dance hall. When Decca recorded there, the floor
was adjusted to its firmest point to minimise LF loss.
Kingsway Hall (which was situated at 75 Kingsway, London WC2)
was another very well-used recording venue. It had a sprung floor
with a large void underneath that affected the sound in a different
way. The void was used to store large rolls of theatrical fabric which
were very absorbent, and they reduced the sound of the underground
trains that ran beneath. This became apparent when the void was
cleared out in the 1980s and the train sounds became more intrusive
and amplified by the resonance of the now empty space. The floor
and void in combination had a resonance in the 600 Hz region that
actually enhanced the viola section.
Figure 1.1a and 1.1b show images from Kingsway Hall.

1.1.2 THE SIZE OF THE SPACE

The larger the distance from the performers to the first boundary (a
wall, ceiling, or floor), the longer the time taken for the first early
reflections to arrive back to the players and any microphone placed
near them. This timing gives a characteristic signature to the sound
that enables us to judge the size of the space. A large space will also
tend to have a longer reverb time because of the distances involved
between reflections. A bathroom or a stairwell is likely to have a
longer reverb time than a large softly furnished sitting room because
of all the hard surfaces, but it will not sound like a large concert hall
because the first reflections happen after a very short space of time. A
useful rough guide is that sound takes about 3 milliseconds (ms) to
travel 1 m (3′4″), or roughly 1 ms per foot. As noted in Chapter 17,
artificial reverb is generally better at modelling the reverberation tail
than the early reflections.

1.1.3 THE GEOMETRY OF THE SPACE

The geometry of the space includes both the shape and size of the
room. Surfaces might be flat or curved, parallel or divergent, and
these affect how sound waves behave in the room. Simple-shaped,
small rooms with dimensions of around 3–10 m with parallel walls
are susceptible to LF standing waves because the dimensions
involved tie in with the wavelengths of frequencies in the 30–100 Hz
regions. Parallel walls that are very reflective can produce a distinct
series of echoes as the sound is bounced back and forth between
them. These ‘flutter echoes’ are easy to perceive using a transient
sound such as a hand clap to set them off, but they will be an
unpleasant colouration on most sound sources placed between the
walls. Rounded wall sections, as are sometimes found behind a
church altar, can focus sound or can reflect it around the edges as in
the whispering gallery at St Paul’s Cathedral in London. Temple
Church in London has a rounded west end which produces some
interesting focussing effects on the organ pedal notes, which are
particularly noticeable just before the point where the rounded walls
start. (See Chapter 14.)
Figure 1.1a
Figure 1.1a A view of Kingsway Hall taken from the stage looking towards the balcony,
which could be used for placing a chorus of up to 300 for recording. This image
shows Anatole Fistoulari conducting the LPO with Mado Robin.
Photo: Courtesy Decca Music Group Ltd.
Domed ceiling areas can cause greater focussing of sound as they
are curved in all directions, and it is usually a good idea to avoid
setting the players up under a clearly defined domed area. Low
ceilings will send very early reflections back to the players and will
impart a rather boxed-in feeling to the recording, so any high ceiling
would be preferred.
The effect of arched ceilings that are often found in older concert
halls and churches is dependent on their materials; those that are
made of a combination of ornate wood and plaster can work really
well as they disperse the sound energy in a complex way that
enhances the reverb time. Both Kingsway Hall (now demolished) and
St Eustache Church, Montreal (used for many Decca Orchestre
Symphonique de Montréal (OSM) recordings) have ornate arched
ceilings. Where an arched ceiling is undecorated and made of hard
stone, the effect is not nearly as pleasant.
Figure 1.1b
Figure 1.1b A view of Kingsway Hall towards the stage with the balcony behind the
camera. This image shows Sir Georg Solti conducting the LPO and Vladimir
1
Ashkenazy playing the piano for Bartók sessions.
Photo: Courtesy Decca Music Group Ltd.
Figure 1.2 shows an image from St Eustache Church, with the
seating removed and flooring in place for recording.
Figure 1.2
Figure 1.2 St Eustache Church and the Orchestre Symphonique de Montréal
Photo: Courtesy Decca Music Group Ltd.
Another interesting feature of churches are side chapels; if large
enough, these areas can have their own distinctive reverb which only
becomes audible when the players are playing loudly, and so their
effect is unpredictable and intermittent. To locate the source of the
problem you will have to go and walk around the church while the
performers are playing.

1.1.4 THE CONTENTS OF THE SPACE

Soft furnishings, curtains, carpets, and padded audience chairs will
absorb HF and mid frequency (MF) reflections and shorten the reverb
time. The presence of an audience seated on hard chairs will shorten
the reverb time in comparison with the unoccupied hard chairs
(which is something to remember when moving from rehearsal to
concert). Likewise, areas of wall surface that are uneven, such as
areas of ornate wooden decoration or cases full of books, will
disperse and absorb reflections and are very good for taming echoes.
When running a recording session, two of the simplest things you can
do to change the acoustics is to add or remove any upholstered
audience chairs and open or close any large curtains.

1.2 What to be aware of when looking at a venue

When you first visit a venue that you are considering using (or having
to use) for a recording, it would be useful to be able to get a feel for
whether the space is going to produce difficulties or whether it is
going to give you a usable reverb signature. Background noise levels
must also be taken into account, and Chapter 2 contains further
advice on visiting a potential venue.
Things to pay attention to immediately are:

Traffic noise and aircraft noise. (Is there a quieter time of day?
Can you record at night?)

Buzzing lights. (Can you get them fixed? Work without them?)

Air conditioning noise.

Noises and cracking sounds from the heating system pipes

(many venues have old heating systems that have to be run for
quite a while before the temperature settles down and the noises
stop).

Broken panes of glass that might let birds in. Noisy birds within
a church roof space are a common recording problem.

Identify anywhere that that has some degree of acoustic isolation

from the main space that you might be able to use as a control
room.

In terms of assessing the acoustic suitability of the space, there are a

few things you can do.
Firstly, visual clues:
What’s on the floor? The surface most likely to produce good
results for classical recording is a wooden floor over a solid
foundation, with a natural finish such as spirit varnish, wax or
oil. This will give some nice reflections and slightly more
warmth as some of the HF is absorbed. A shiny polyurethane
varnish or a stone floor will give a more brittle sound as it
reflects more HF. Carpeted floor will absorb a lot of HF and MF
and will be less enjoyable for the players if it covers the whole
floor, although a small area of carpet around the performer can
be a help if the floor reflections are very bright.

Is it a sprung floor? If so, it could act as a bass trap and remove

some of the LF reverb spectrum, depending on its exact
properties.

What’s on the walls and other surfaces? Stone will give brighter
reflections, and too much HF can make the strings in particular
rather tizzy and over-bright. Wooden areas are good, especially if
they are rough or have a lot of architectural detail. This will
disperse sound and mitigate against strong, distinctive single
reflections. A minimalist glass and steel building with little in
the way of architectural decoration inside will have a lot of
distinctive HF reflections and very little to disperse sound
energy, and as a result will probably sound coloured and
horrible.

Is there an arched ceiling? Again, if this contains a lot of wooden

beaming, or panelling, it is likely to be beneficial, but if it is very
unadorned stone, it will create a lot of HF reverb.
 Is there a dome? Are the players planning on sitting right
underneath it? See if you can talk them out of this.

Is the ceiling low? Be aware that a space that is too small in

volume can become acoustically saturated quickly if the source
is large and/or loud such as a brass band or orchestra. Playing
something loud in too small a room is likely to be unpleasant for
the players, and the recording will also suffer as the sound will
not feel open with its full dynamic range but compressed and
closed in.

Are there any side chapels the size of small rooms that are
enclosed on three sides? These will have their own reverb
characteristics that might be triggered by loud sounds in the main
space.

Then, some aural clues:

Clapping at various locations within the building will give you

an idea about the timing and quality of early reflections. Are
there any distinctive echoes where the sound seems to slap
straight back at you? Are there any areas of flutter echoes? You
will also get a sense of what the HF decay sounds like, and
whether the reverb is bright.

Stand next to a wall and talk or sing at it – this will give you
some useful information about what sort of frequencies the walls
are reflecting. Hard plaster will reflect differently to older
horsehair-type plaster.
 Making a variety of different frequency sounds around the
building – hooting, squawking, and so forth; use your chest voice
and head voice and see how the building responds to a variety of
stimuli.

To check for LF standing waves, you will need to be producing a

substantial volume of lower-pitched sounds. See if you can set
any off within the building. If you are recording something with
organ, see if you can ask the organist to play some lower notes to
check for any unusual acoustic behaviour.

Stand at the position within the building where the performers

are likely to be situated. Do sounds that are emitted here set off
any unpleasant echoes, flutters, or standing waves? Do they
activate any side chapel reverb if loud enough?

Ask someone to walk down into the building and talk back to
you from various places. They should do this without shouting or
raising their voice so that the room is not excited to any great
extent. The intelligibility of the speech will give you a helpful
indication of the nature of the early reflections, and how much
clarity you are going to get from the musicians at various
distances. If you find that speech intelligibility is compromised
at 3–4 m away (9′10″ to 13′), you should not consider using the
venue at all. If it is becoming difficult to understand at 9 m (30′)
away, the recording might be more challenging but achievable; 9
m is roughly the depth of an orchestra, and the back players
should at least be able to understand the conductor.

1.3 What can you do to help with poor acoustics?

Without the ability to completely remodel the space you have to work
in, there are still some things that you can do to help where the
acoustics are less than ideal. Below are some of the most common
problems, and some suggestions for helping with them.

1.3.1 THE SPACE IS BOXY SOUNDING WITH A

SHORT REVERBERATION TIME
Short reverb time is one thing; this can be helped by the addition of
artificial reverb (see section 17.5). The boxy quality is down to the
room being a little small, possibly with a low ceiling, and simple in
shape, and the early reflections are what gives this quality to the
recorded sound. The amount of room colouration that you are picking
up needs to be reduced, and this is best done by using curtains
everywhere possible to reduce reflections as much as possible, and
then adding in artificial reverb later on. Adding reverb to a very dry
sound is discussed in section 17.5.7; it can be difficult to do
realistically and might involve layering more than one reverb. If
adding curtains to the space is not possible, moving microphones a
little closer to the musicians and using more directional microphones
will reduce the proportion of reverb to direct sound. The danger of
this approach is obtaining too close and localised sounds from the
players, so should be used with care. Again, the resulting recording
will need artificial reverb.

1.3.2 THE SPACE HAS AN EXTREMELY LONG

REVERB TIME
This would be typical of a large cathedral, especially if the surfaces
are stone and marble. Reverb times in the range of 6–12 seconds are
not unusual, but make the recording of non-liturgical music quite
difficult. Given the size of the space, there is likely to be little you
can do to significantly reduce the reverb time, although keeping as
many padded chairs out as possible is sensible. Wooden pews and
chairs will not have the same beneficial effect. As in section 1.3.1,
getting closer with microphones and using directional microphones
so that the direct sound dominates the signal will be the best way
forwards. The reverb tail will still be there, but at a lower level, and
careful addition of some artificial reverb of a shorter length to fill in
will help improve the reverb tail to make it feel more natural. Figure
1.3 is an illustration of what you might try to achieve. (See also
Chapter 17.)

1.3.3 THE SPACE HAS PROMINENT FLUTTER

ECHOES
There are two approaches. One would be to avoid placing the players
between the walls that are causing the problem; this might of course
be impossible. The second approach will be to treat the walls by
hanging something absorbent, such as curtains, over at least one of
them to prevent the repeated bouncing back and forth of the sound.
Figure 1.3
Figure 1.3 Shows a loud sound with a long, fairly low level reverb tail, and how shorter
artificial reverb can be used to fill in the gap

1.3.4 THE FLOOR IS TOO REFLECTIVE

If the immediate floor reflections are very bright, then placing the
players on a small piece of carpet (about 1.5 m (5′) square for an
individual player) will help this. Annoying floor reflections are much
more of a problem with chamber music where there are very few
players and a large expanse of shiny floor. Where there is an
orchestra, there are more bodies in the room that will help to absorb
some of the immediate sound and break up floor reflections. This
effect can be partially replicated when performing chamber music by
adding randomly distributed upholstered chairs around the players to
reduce and disrupt floor reflections. They can also be used to good
effect in the corners of the room to break up general early reflections;
the addition of a lot of upholstered chairs will reduce the overall
reverb time.

1.3.5 THERE ARE DISTINCTIVE LOW TO MID

STANDING WAVES
LF standing waves are not something that can easily be prevented
without big architectural alterations as LF energy is not absorbed by
materials such as curtains and carpets. If it is possible to avoid
placing your microphones where the particular frequencies are
exaggerated or diminished, you should do so. Sometimes, moving the
room set up around by 90° can help with standing waves.
Any decisions about changing aspects of the performance space
should be made with the involvement of the musicians who will need
to know what you are doing and why and be given a chance to try out
the results. If they strongly dislike the effect on their playing of you
having curtained off the entire room, you will have to accept a
compromise position. Acoustics affects both the performance and the
recording, and we need both for a successful outcome.

Note
1 BARTÓK/Piano Concertos 1 & 2/LPO/Solti/Ashkenazy/DECCA (1980) LP SXL
6937.