Forget Size, SHAPE Matters

When it comes to building an acoustically sensitive room, like a mix room or a live recording space, I always like to be brought in to consult at the very beginning of the project. This allows me to give input regarding site location (like avoiding a location next to a cement plant or rail road tracks), and check the site out for other issues that may impact how well the facility works.

But the most important reason for me to join a project at the same time as the architect is because the shape of the room, and the relationship of the three dimensions in particular, is the single most important factor in attaining a good sonic quality when the room is completed.

I like to say that once you get the shape of the room right, you almost have to work hard to make it sound bad.

It’s Hip To Be Square
First, I start with a rectangular room. Parallel surfaces, right angles, the whole thing.

Some people recoil at the thought of this, citing the “rule” that you don’t want parallel walls (they cause flutter echoes). But what you gain from having a rectangular room with parallel walls far outweighs what you sacrifice.

In every enclosed space there is a specific resonant frequency for every dimension. (These resonances are also known as “room modes” or “eigentones”.) So in a room with, say, non-parallel side walls (“splayed walls”), every different dimension between these two walls has a different resonant frequency. That means that as you walk from the wider end to the narrower end, all the resonant frequencies will shift up from their lowest at the wide end to their highest at the narrow end.

The Damage Done
In practice, this does two things.

First, it adds a very substantial level of variability to the room, so that the sound quality or timbre will change as you move around the room. Second, it makes calculating the frequency and location of the room modes nearly impossible, so the devices used to address the room resonances won’t be very effective.

On the other hand, in a rectangular room, it is very easy to calculate the frequency and location of the room resonances. They are predictable, their frequencies and locations are easy to calculate, and the frequencies don’t shift as you move around the room.

What you get in return for this is greater sonic consistency and substantially less variation in timbre as you move around the room. This makes it a lot easier to do your mixing, playing, or listening. And you are much more likely to be confident that what you hear is what’s “on tape”. If you have clients, working with them will be easier because everyone is hearing much closer to the same thing.

Meanwhile it’s easy to deal with any flutter echo with simple wall treatments like diffusors or absorbers. But the coloration, increased variation and unpredictability that come from shifting dimensions can’t be fixed. Band aided, maybe, but not fixed. Not with EQ, not with absorption, not with anything. The variation will always be there.

Things Take Shape
Now that we have chosen a rectangular room, the question becomes “what should the dimensions be?” Designing a good sounding room doesn’t have to be a hit-or-miss affair; it requires close examination of the room resonances. For good acoustics, selecting the width, length and height is a lot more than making the best use of the available space and designing a nice interior layout.

To end up with a good sounding room, the most important thing is to adjust the three dimensions during the design phase so that ALL the room resonances are distributed as evenly as possible in the frequency .

This is done by selecting three room dimensions that work with the space available, calculating all the resonances for each dimension, and then tabulating them by ascending frequency. The goal is to get the differences between the frequencies as even as possible. Ratios that are whole numbers should be avoided, as they are exactly what you’re trying to get away from.

Sometimes two or more resonances will pile up around one frequency, which happens with whole number ratios. This should be avoided, because this creates coloration in the room timbre. Likewise, if there is a large gap on both sides of a resonance you’ll end up with coloration there as well.

Your first calculation will have some obvious problems, so the next step is to change one of the dimensions and perform another calculation. And again. And yet again. Keep repeating the calculation until you have dimensions that provide as even a distribution as you can get.

There will be some limitations on what dimensions you can use, such as a room height that can’t change or a room length that is set in stone. You also don’t want to use dimensions that will make your room uncomfortably narrow, leave you with an oppressively low ceiling, or otherwise make the room uncomfortable to be in. While this limits the solutions available to you, it also limits the number of calculations you have to make, so it’s easier on you.

So now you’ve settled on a rectangular room, which your builder and accountant will like, and you’ve selected dimensions that have the smoothest distribution of resonances. You’ve just gone most of the way to getting a room that will sound good.

As I said, once you get the dimensions right, you almost have to work to make the room sound bad.

Last Thoughts
I’ve tried using dimensions from a successful room and scaling them for another room of a different size. But when I ran calculations with these new dimensions, I found it didn’t produce the even distribution of resonances that I thought it would.

I found two reasons for this.

First, going larger or smaller changes the distribution of resonances so that some may get too far apart, and some may get too close. Second, calculating the resonant frequencies involves a non-linear mathematical function that renders a linear scaling inappropriate.

So yes, you have to do the entire calculation exercise with every different room.

But the results are well worth the effort.