Resolution v2.7 1003
26/9/03
9:50 AM
Page 60
sweet spot
An insight into subwoofers
The fast spread of discrete multichannel audio logically reinforces the use and development of subwoofer speakers. While subwoofers are
extensively used in professional, consumer, home theatre, car audio and movie theatre markets, the history and development of their
design and technology is quite poorly documented.
CHRIS ANET, GENELEC
I
N THE MID-1950S various manufacturers of hi-fi
speakers started to introduce fairly compact sealed
cabinet enclosures that had long-throw type
woofers. In the early 1970s Jensen had a talented
engineer, James Novak, who came up with equations
describing a bass reflex speaker design. Other
researchers, such as Thiele, Small, Ashley and
Benson, continued with Novak’s ideas and refined the
mathematical analysis and synthesis for closed,
vented and passive radiator speaker systems. Since
then, several software packages have evolved to make
the system design even easier.
The ability to reproduce extended low frequencies
was of great interest to the movie industry. In the
1950s Cinerama – then Cinema-scope – and Todd-AO
Studios first brought multichannel sound to the large
cinema screen. A 6-track 70mm Todd-AO system
premiered in 1955 and in 1976 an adaptation of this
system was developed that featured a centre mono
channel with dialogue mixes only, leaving the left and
right channels with a new role as ‘bass extension’
channels. These two channels provided low frequency
effects made to enhance the cinema experience. The
first movie to feature this ‘bass extension’ was Star
Wars in 1977.
In 1992 Dolby Digital was introduced with the
cinema release of Batman Returns, which made a very
clear step forward in the use of the LFE or ‘.1’ channel
in audio, video and movie productions.
If subwoofers had become an important addition to
standard stereophonic reproduction systems during
the late 1970s, today, with the wide acceptance of
multichannel audio, they have become an essential
device in any type of reproduction system, whatever
its cost and size might be. In most cases, the
subwoofer reproduces a significant bandwidth of the
LFE channel, and also the very low frequencies of
some or all main channels, with the help of a bass
management system.
There are a few facts about low frequency
reproduction and hearing that have to be considered.
The location of a subwoofer with respect to room walls
will strongly affect how the subwoofer excites the low
frequency acoustics of the room. Also, relative to a free
field acoustical space, once a subwoofer is placed on a
floor and against a wall, there is a 12dB level increase
(this amplitude gain of 6dB per wall occurs usually
below a few hundred Hz, where the radiation is fully
omnidirectional), and once it is placed in a corner, the
increase is 18dB. This change in the available sound
pressure, due to power doubling at each halving the
radiation space, is beneficial to subwoofer design as it
enables reproduction of low frequencies with less
distortion.
At low frequencies the room absorption is usually
less effective and because the room modes are sparsely
located, the separate standing waves are easily
audible. A measurement of a subwoofer in a certain
position will reveal the performance of the subwoofer
in that location only. When the position changes, so
will the acoustics of the room, and hence the
60
subwoofer performance might suddenly be very
different from the initial position. The variations in
room acoustic behaviour create large variations in
frequency response, which strongly colour the
perceived bass quality.
At very low frequencies the auditory sensation of
sound changes and the pitch recognition disappears
and becomes a feeling of pressure. Researchers
Guttman and Pruzansky [1] initially established that
this change occurs at about 19Hz. Further research by
Buck [2] showed that the average lower limit of pitch
was perceived at 12Hz with 120dB SPL of pure tone
resolution
signals. The ability to hear pitch changes depends on
SPL (9.6Hz at 123dB, 7Hz at 129dB). These results
are in direct relation to the work of Fletcher and
Munson[3] in 1933, and Robinson and Dadson[4] in
1956 (see Figure 1) about the minimum threshold of
hearing. It shows clearly that our hearing is not very
sensitive to low frequencies reproduced at low level.
As the level increases the audibility improves, but is far
weaker than in the midrange band. It also shows that,
to generate any audible sound, the subwoofer has to
have a certain minimum output capability.
The most essential property and requirement for a
October 2003
Resolution v2.7 1003
26/9/03
9:59 AM
Page 62
sweet spot
subwoofer is its capacity to move air. In the final
technical specification this converts to the
maximum driver excursion and effective
piston area.
To achieve a constant sound pressure level
while decreasing frequency, the necessary air
volume displacement is proportional to
1/(frequency)2. In other words, the air volume
displacement increases by 12dB per octave
with decreasing frequency. This relationship
does not depend on the type of enclosures
used. So, if we need for example 50cm3 of
volume displacement at 100Hz for 110dB
SPL, we need 1250cm3 at 20Hz for the same
SPL. Additionally each doubling of the SPL
(+6dB) needs twice the volume displacement.
Hence the constraints on the driver excursion
and diaphragm, as well as on its suspension,
are significant when you want to achieve
high sound pressure level as well as low
frequency extension.
Most subwoofers are built around closed,
vented, band-pass or transmission line box
principles. The most common of these are
closed and vented boxes. Again, independent
of the type, the efficiency is proportional to the
size of the box. The smaller the box, the lower
the efficiency and consequently, the higher the
necessary amplifier power for a certain SPL
output. Hence, you can see very small
enclosures with 1000W amplifiers with, in fact,
a final SPL output identical to the output of a larger
box with a 100W amplifier.
Comparison of the performance of a closed box and
a vented box of the same internal volume will reveal a
number of things. The efficiency of a closed box is
lower than that of a vented box and to get the same
efficiency the low frequency cut-off of a closed box will
be higher. The driver displacement capacity must be
higher in a closed box and, due to the required longer
excursion of the driver in a closed box, the distortion
is in practice often higher. Finally, the impulse
response of a closed box is often better than that of a
vented box.
There are constraints in vented subwoofer design.
Because of the performance benefits, since the late
1970s many manufacturers have developed ‘vented’
or ‘reflex’ enclosures. The vent – or reflex port – is an
integral and necessary part of the design and research
to improve the vent performance and efficiency has
been pursued over the years.
Looking at the acoustical phenomena and their
mechanical analogies for a reflex enclosure, the vent
and the box form a ‘mass-spring’ resonant circuit.
The ‘spring’ is the air inside the box and the ‘mass’ is
the acoustical mass of the air in the vent. This
acoustical mass is proportional to the ratio of the vent
length to its cross sectional area. The same acoustical
mass can be achieved with very different dimensions;
a small vent area with a short tube is equivalent to a
large vent area with a long tube. However, not any
combination is valid.
At the vent-enclosure resonant frequency most of
the radiation comes out of the vent. If we have 110dB
SPL coming from the driver, the same is coming from
the vent but at a lower frequency. This implies that
there is a lot of air movement in and out of the
enclosure through the vent. Because of that, the next
step is to look at the air flow properties inside the vent.
To minimise the speed of the air flow the cross sectional
area of the vent should be large. As a consequence (to
keep the correct acoustical mass) the vent becomes
62
often larger than the driver is, or there will be
several passive radiators for one active
driver.
Coming back to the vented design, there
are two main problems in the vent: its
possible length and turbulence. To avoid
turbulence in the vent, the air flow should
be laminar. To keep this flow laminar, the
vent shape must be smooth without sharp
angles, the air speed in the tube must be
low enough and the vent ends have to be
acoustically properly terminated.
To design a subwoofer that incorporates
all the above crucial acoustical parameters,
Genelec developed the LSETM series
subwoofers. The enclosure shape is made
from a strip of sheet metal that is wrapped
to a spiral. The vent is formed between the
rounds of the spiral. The length of the tube
can be as long as needed, matching the
SPL output and the air flow required. The
vent is located at the perimeter of the
enclosure and has no acute angles. The
radius of the spiral being variable, the vent
height also changes along the tube length
avoiding any tube whistling and chuffing.
The vent ends are also carefully formed to
provide a proper termination and a smooth
flow transition between the various air
volumes – inside the enclosure, in the vent
and in the room.
The minimised losses in the reflex tube and
its smooth shape provide clarity and definition
with low distortion. So, despite a fairly small enclosure
size, you can achieve very low frequency reproduction
at high level. ■
References
Fig. 1
long. When the vent length grows, the enclosure
dimensions soon start limiting the maximum length
that will fit inside a normal enclosure.
In conventional designs, the vent is sometimes bent
to gain additional length in the speaker enclosure but
acute angles in the vent shape will cause increased air
turbulence. Air turbulence means losses that increase
rapidly with sound pressure level and cause audible
noise, compression and distortion.
One elegant solution to this problem was published
by Harry F. Olson. He substituted a long vent with a
‘drone cone’, a heavy speaker cone without voice coil
and magnet (hence it is called a passive radiator). The
acoustical mass of the passive driver diaphragm is
made equal to the acoustical mass of the column of air
in a theoretical long vent, and in general, the system
performance is equivalent to a vented enclosure.
Passive radiators do not take much space in the
enclosure, but without proper design their performance
can limit the maximum system output. Above a
certain sound pressure their displacement becomes
non-linear and the dynamic range is limited. Naturally
this should not happen before the driver starts
reaching its limits. Therefore the passive radiator is
resolution
[1] N. Guttman and S. Pruzansky, Lower Limits of
Pitch and Musical Pitch, J. Speech Hear. Res.,
vol. 5, pp. 207-214 (1962 Sept.)
[2] M. Buck, Perceptual Attributes of Supraliminal
Low-Frequency Sound and Infrasound, presented
at the 66th Convention of the AES, preprint 1663
(1980)
[3] H. Fletcher and W.A. Munson, Loudness, Its
Definitions, Measurement and Calculation, J.
Acoust. Soc. Am., vol. 5, pp. 82-118 (1933 Oct.)
[4] Moore, An Introduction to the Psychology of
Hearing, 3rd edition, Academic Press Limited,
1989. ‘Equiloudness curve’ redrawn from
Robinson and Dadson (1956), by permission of
the authors and the Director of the National
Physics Laboratory, UK.
Ilpo Martikainen, Laminar Spiral Enclosure (LSETM),
what it is and what it does? – Product launch, Genelec
Oy, (February 2002).
Martinsound Reports, Secrets of doing surround
sound on your existing console, copyright
Martinsound, Inc. (1990 – 1999).
Louis D. Fielder and Eric M. Benjamin, Subwoofer
Performance for Accurate Reproduction of Music, J.
Audio Eng. Soc., Vol. 36, No. 6 (1988 June).
Contact
GENELEC, FINLAND:
Website: www.genelec.com
October 2003