From our own article on diffusion, I hope this helps for the basics -
is a very misunderstood concept in the audio field. What exactly is diffusion? According
to Webster, there are several definitions which relate to various fields. Let’s take a
look at the generic definition and one more audio related.
movement of a substance from an area of high concentration to an area of low
Diffusion: In audio, the scattering of sound waves, reducing the
sense of localization.
Those 2 definitions may seem to be very different but
they really aren’t. If we think of a series of sound waves bouncing off of a wall,
they’re concentrated and moving in the same direction at the same time. The reflections
off of the wall will come back to the listener at approximately the same time and from
approximately the same direction. The ‘concentration’ is still in tact in the spatial
and time domains.
Now, if we introduce a diffuser onto that flat surface and we
generate the same sound waves, what happens? Most people would say that we are scattering
the sound in other directions rather than in the directly reflected path. This is true,
but is not the whole story. A good diffuser not only changes the reflections to propagate
in different directions, but does so EVENLY in all directions. Just splitting a wave to go
into 2 directions doesn’t do enough. We’ve effectively changed the concentration of
the waves in a certain portion of the room. We’ve also effectively taken the intensity
of the initial wavefront and split it among the various reflections so that each one is
not only coming from a different direction, but each is also weaker and harder to
distinguish but we’ve lost no energy in the process.
A proper diffuser also
impacts the time domain. If we have the same wave and it gets reflected into multiple
directions evenly, the length of the path the various waves must take to reach the
listener also changes. In the world of sound, distance is time. For ease, we can say that
1 foot equals approximately 1ms of time. So, if we’ve changed the path into say 8
different paths (in reality, there are many paths but we’ll look at 8 just for ease of
understanding), each one ideally with have a different path, which means that they’ll
arrive at 8 different times. Again, we’re diluting the concentration of sound but this
time in the time domain instead of spatially. In addition, these different path lengths
cause differences in the number of reflections and the amount of air the wave passes
through which will cause each reflection to have a different intensity (more dilution in
yet another domain – intensity).
So, now we have 8 reflections that have been
changed and are all different in 3 domains – time, direction, and intensity. This makes
it much more difficult for our ears and brain to determine exactly where the sound is
coming from. This fits perfectly the definition of a lack of localization. The net result
of this is that we trick our brain into thinking the room is larger that what it is and
yielding a more spacious sound.
There are a great may myths about ‘home
brew’ ways to provide diffusion. Most do not work at all and many work poorly or only
over a very narrow range of frequencies. Let’s take a look at one – a bookcase with
books set at randomly varying depths.
First of all, books, if anything will be
more absorptive than reflective at all but the highest frequencies. Second, random depths
do not generate random reflections over a predictable and controllable area. The width,
height, spacing, and pattern of the wells of a diffuser are carefully calculated to make
sure they generate a smooth and even scattering of the waves over optimally a 180 degree
angle. Third, without the careful calculated spacing, we can cause frequency related
aberrations due to constructive and destructive wave interactions from various
reflections. We’re in effect getting very little of the benefit of a diffuser while
causing more issues in the frequency response.
Another myth is using CD/DVD
cases or LPs in the same sort of random manner. In addition to the issues presented above,
the cases themselves are so narrow that the frequencies that would be affected would be
only in the highest ranges. Again, the width, depth, and pattern of the wells and peaks of
a diffuser not only generate the proper diffuse pattern, but also determine the frequency
range over which a diffuser is effective.
Frequently Asked Questions:
Q: Is it true that I have to sit far away from diffusion to make it work?
yes and no actually. You need a certain amount of distance between the listening position
and the diffuser to allow it to do its job. The closer you sit, the less chance the waves
have to move farther from you and the less difference there is in the timing that they
reach your ears. They’ll still have some effect but not as much.
Q: Is it
true that you need to have a large room for diffusion to be effective?
A: Again, yes
and no (seeing a pattern here? Every room is different). Normally, in a smaller space, one
needs proportionately more absorption for bass control than in a larger room. This leaves
less space for diffusion – not to mention the issues raised in the first question.
However, diffusion can still be effective up high in the room to assist in killing slap
echo and controlling the ‘zing’ in a room – especially a tracking room.
Q: What is the best place for diffusors?
A: This really depends on the situation
and what you’re using the room for. Diffusion can be used up high in a room to kill slap
echo, in the rear of the room along the side walls, in place of absorption on the ceiling,
etc. The most common place to see diffusion in a studio environment is on the rear wall.
However, many times, you’ll benefit more from absorption there to kill a strong bass
null off the back wall.
Q: How would these help in a live tracking room?
A: In a live room, we want it to be as the name implies, relatively live. But, we still
need to get it under control in terms of decay times. We also want to get that ‘big’
sound. Diffusion can help to eliminate echo and produce a nice large sound without killing
too much of the liveliness. It does so by spreading the sound, causing it to move through
more air and bounce off more surfaces which decreases the intensity of the sound. For
decay time control, diffusion is normally used in conjunction with broadband absorption.
David Shevyn General Manager GIK Acoustics Europe