|
|
| |
|
Live
Sound: Part 1
|
|
by
Al Delaney
|
| |
The
sense of hearing is our brain's response to nerve
impulses originating in our ears. Those impulses
are generated by a series of events beginning
with the movement of the small diaphragm we commonly
call the eardrum. The eardrum is set in motion
by fluctuations in the surrounding air pressure.
Those fluctuations propagate through the air,
spreading out and away from their source, as a
series of compression waves. Sound is defined
as those air pressure fluctuations capable of
producing the sense of hearing-the pressure waves
have to meet certain parameters. Acoustics is
the science of sound and the interaction of sound
with its environment. In this first article I'll
cover the very basics of Acoustics.
Sound
Waves
As
alluded to above, to be perceived as sound,
the pressure fluctuations have to fit a specific
set of parameters. They can't be too great or
too little and they can't occur too frequently
or too infrequently. If they're too great they'll
damage the eardrum, if they're too little they
won't be perceived. If they occur too frequently
the eardrum won't have time to respond to the
first pressure change before it's hit with the
next, and if they occur too infrequently the
brain "kind of" moves on before recognizing
the event as part of a pattern.
If
someone stuck a tuning fork and you were able
to measure and record the pressure changes caused
by its vibrations, you could plot that data
as a function of time.
|
|
|
| The
resulting graph, if it were a very good tuning fork,
would look like the plot of a sine wave or a cosine
wave (you remember these from High School, don't
you). It would be the plot of a pure tone. If the
tuning fork were for the same note as the heavy
E-string on a bass guitar and the surrounding ambient
air pressure were 760 mmHg (1 Atmosphere), its plot
would look like this: |
|
|
Here,
the distance above (compression) and below (rarefaction)
the ordinate measures the deviation from ambient
pressure caused by the sound wave. The greater
the deviation, the louder the sound will be. This
"loudness" is measured in decibels (dB).
We'll talk more about that a bit later.
On
a graph like the one above it's easy to identify
the smallest section of the plot that could
be repeatedly copied and laid end-to-end to
reconstruct the whole plot. Measuring how much
of the ordinate is spanned by one of those copies
gives the period (T) of the sound wave. The
wave frequency (n) is then 1/T. The unit used
for frequency is Hertz (Hz). At five repetitions
per second n = 5 Hz, 10 repetitions per second
n = 10 Hz and so on.
Related
to frequency is wavelength (l). Sound waves
propagate through air at a certain speed (v).
That speed is primarily determined by air temperature.
At room temperature (72o F) sound waves propagate
at approximately 1,132 feet per second; it's
1,165 ft/s at 100o F and 1,087 ft/s at 0o F.
Wavelength is related to frequency by the speed
of sound. The governing equation is l = v/n
that is wavelength is equal to the speed of
sound divided by the wave's frequency. For the
pure note depicted on our graph n = 83 Hz so
at room temperature l = 13.6 feet.
Another
way of looking at it is like this: Strike the
tuning fork again. Wait until the sound waves
make it across the room and wave a magic wand
to stop the action (kind of like the drawing
at the top of the page). Now measure and plot
the air pressure as a function of distance from
the tuning fork. Your plot would look very much
like the Pressure/Time plot above, but instead
it would be a Pressure/Distance plot with the
pressure extremes dampening with each cycle.
In this plot wavelength could be measured directly
from the graph.
Complex
Sound Waves
Instruments
don't produce pure tones. Music would be very
boring if they did. Instead each instrument
has its own unique timbre. This timbre is the
result a totality of vibratory affects within
the instrument. Below is a plot of the sound
wave produced when I plucked the E-string on
my guitar.
|
|
|
| You
can see the wave has the same frequency as the pure
tone wave of the first graph (and consequently has
the same wavelength) but nothing much else looks
the same. This is because when the string of a guitar
is plunked a variety of oscillatory effects within
the guitar contribute to the sound, not just the
string. The resulting combined effects produce a
complex wave pattern. That pattern is the result
of different pure tones, having different frequencies,
amplitude and phase shifts (a delayed start time)
adding together-they either constructively interfere
with each other or destructively interfere. To make
it a bit more clear, look at the two graphs below.
The first contains two wave plots. The second is
a plot of the resulting addition of the two waves.
|
|
|
| |
|
|
| In
an extreme example of constructive interference,
two identical waves meet and are perfectly in phase
(all their valleys and peaks match). Then all the
pressure fluctuations will double. An extreme example
of destructive interference would be when those
two waves were perfectly out of phase (peak meets
valley and valley meets peak). Then the waves would
cancel each other out-no sound.
The
plot of a complex sound wave is the plot of a
special kind of mathematical function. It's a
type of function that can be exactly represented
by the summation of different sine and cosine
functions. This is called its Fourier Representation.
Synthesizers use Fourier Representation to great
affect. They don't produce sounds like individual
instruments do; instead they produce the combination
of pure tones needed to mimic the instruments.
Once
you have the Fourier Representation of a sound
wave, you can create another very useful plot-that
of its Fourier Transform. This would be a plot
of the decibel levels needed at each frequency
to reproduce the sound. In the introduction to
this series of articles, I mentioned that at the
heart of the sound system I was learning is the
Yamaha DM 1000 Digital Mixer. I was somewhat surprised
that this system doesn't have the ability to perform
discrete Fast Fourier Transforms. It sure would
make EQ'ing a lot easier.
What
Do You Hear?
The
table below lists frequencies for some common
musical elements (and some other things). It also
lists their wavelengths. I often find the wavelength
to be the more useful number when trying to understand
how a sound will "play" in a room but
of course frequency is more useful when hunting
an annoying sound during EQing.
|
|
Element
|
Frequency
(Hz)
|
Wavelength
(Feet)
|
| Lowest
perceived |
20 |
56.6 |
| A0,
1st piano key |
27.5 |
41.2 |
| Kick
Drum |
63 |
18 |
| E2,
open 1st Bass string |
82.4 |
13.7 |
| E3,
open 1st Guitar string |
164.8 |
6.9 |
| C4,
Middle C on Piano |
261.3 |
4.3 |
| C4-A5,
Soprano voice |
261.3-880 |
4.3-1.3 |
| Snare
Drum |
1K |
1.1 |
| C8,
88th piano key |
4186 |
0.27 |
| Symbols |
~5K |
0.23 |
| Highest
Perceived |
20K |
0.057 |
|
|
Loudness
"Loudness"
is the result of an almost instantaneous difference
in pressure that impacts the eardrum. (Relatively
slow changes in pressure, like when we travel
up a mountain, have no affect because our sinuous
systems work to equalize pressure on both sides
of our eardrums.) The governing equation is:
|
|
|
| (You
remember logarithms from high school, don't you?)
Here the |
|
|
| is
the root-mean-square of the acoustic pressure where
the acoustic pressure is just the difference between
the ambient pressure and the increase (or decrease)
of pressure due to the sound wave. Root-mean-square
is just a mathematical way of including the positive
affect of a negative value-the decrease in pressure
cause by part of the sound wave. |
The 
|
|
is the ambient pressure value. Pressure can bemeasured
in mm Hg (millimeters of mercury), atm (atmospheres),
or some other unit-in the first and second graph,
I chose mm Hg. In any case it represents a force
applied to a unit area. The unit for loudness is
the decibel (dB). Below is a table listing some
typical dB levels. |
|
Loudness
(dB)
|
Example
|
| 160 |
Ruptured
Eardrum |
| 140 |
Pain
Threshold |
| 120 |
Hearing
Loss with prolonged exposure |
| 100 |
Automobile |
| 80 |
Loud
Radio |
| 60 |
Conversational
speech (not my mother in-law) |
| 40 |
Average
living room (not mine) |
| 20 |
Quiet
room |
| 0 |
All
quiet |
|
| An
interesting note, it's not uncommon for weather
conditions to cause local air pressure changes great
enough that if they occurred instantaneously they
would rupture the eardrum. At 760 mm Hg, the air
pressure on a nice day at sea level, it would only
take a quick 22.8 mm Hg change to damage the eardrum.
The
Environment
One
of the first mistakes I made when mixing a live
show was to rely on the dB meters to balance the
channels. What I didn't realize was that the size
and shape of our heads, the placement of our ears
and the structure of our ear canals all conspire
to selectively amplify a certain range of frequencies.
Below is a plot of average frequency sensitivity.
Notice that our maximum sensitivity is at about
2k-3k Hz.
|
|
|
| |
| The
above example serves to stress the importance of
using your own sense of hearing when determining
what sounds good and what doesn't. That said, the
characteristics of sound waves that interacted with
the environment, the human head in the above case,
are reflection, refraction and dampening. I'll explain
these characteristics separately but realize, they
do intimately interact to produce a cumulative affect.
Reflection:
There is a degree to which sound waves reflect
from surfaces. They reflect better off some surfaces
than others and in fact certain frequencies will
reflect off a surface that absorbs other frequencies
but, as a rule the softer the surface the less
the reflection. In most cases you would like to
minimize sound reflection as much as possible.
Venues
with concrete floors and brick walls will tend
to sound like echo chambers. Ones with carpeted
floors and wood or textured walls will have a
better sound. Also, it's easier to reduce sound
reflection from directional instruments like horns,
guitars and vocals than it is from drums or even
piano. We use a shield around the drummer but
even that doesn't totally eliminate the problem.
Each
venue will also be susceptible to a specific frequency
of standing wave. This occurs when a wave front
reflects off the back wall and constructively
interferes with the rest of the wave. This is
usually more of a problem in small venues. The
offending wavelength is easily identified if you
know the length of the room.
Refraction:
Previously I spoke about constructive and destructive
interference. Well those are aspects of refraction.
If you have ever seen a wave tank in a high school
science class you'll understand what I mean (boy,
you really had to pay attention in high school
to get much out of this article). In one experiment
they generated some waves and sent them through
a couple of small slits. After the waves passed
through the slits, becoming two wave fronts, they
fanned out and interfered with each other. In
another experiment they would place different
size objects in the wave path. If the object was
large relative to the wavelength you would see
a lot of the wave reflected. But if the object
was small there wouldn't be any noticeable break
in the wave front. Sound waves react the same
way; they'll interfere with each other and long
wavelength sounds will not be impeded by small
objects in the wave path while short wavelength
sounds will be reflected.
One
of the effects of refraction can be heard if you've
ever in the parking lot of a football stadium
when the band is playing. Because of the bass
drum's long wavelength it will clearly be heard
from the parking lot but you won't pick up the
higher pitched instruments until your inside the
stadium and have a direct line of sight to them.
Dampening:
This is a decrease in the amplitude of sound waves-a
reduction in loudness. I touched on dampening
in the section on reflection when I spoke of surfaces
absorbing sound waves-it is hard to separate these
concepts. Sound waves also dampen as a natural
process of propagation through air--as a portion
of their mechanical energy is constantly being
converted into heat energy. There is also a non-energy
related reason why sound "gets quieter"
further from its source. Sound wave fronts spread
out in an arc from their source as such someone
standing close will experience a larger percent
of the wave front than someone standing further
back.
Some
Observations From A Sound Rookie
The
board mix isn't and shouldn't be the room mix.
We've put together a system that allows us to
broadcast live shows in real time over our web
site from any remote location. When we first did
this, I thought it would be best to take the signal
direct from the venue's mixer. I've found that
that's not a good idea, especially if there's
a lot of horns, drums or piano in the mix. The
reason is that a good sound engineer will be aware
of and account for the sound coming off the stage,
not just what's coming out of the speakers.
Stereo
is usually not a good thing. When mixing for a
venue you want everyone in the place to hear the
same thing, as much as possible. If the mix is
separated you'll only be catering to those lucky
enough to be standing in the center of the room.
Bodies absorb sound. This is an easy one. Bodies
will change the acoustics of a room, so the mix
for an empty room will be different than a mix
for a full room. But of course if the room's empty,
who cares?
No
matter how good you are, you won't be able to
cover the room perfectly. It just can't happen.
There will always be some quiet spots or even
dead spots somewhere in the room or some areas,
usually close to the stage, where the mix isn't
even.
|
|
|
|