The rhythm method
There are several methods available for positioning loudspeakers, ranging from the arbitrary to the rigorous. One of the best we've had some success with in the past is the WASP method. How it works is intuitive - it relies on 'voicing' along with a helper and the ability to deal with feeling self-conscious. Musing on just why it works led to much idle wondering, after which cogitation the rhythm method was hatched: the Twisted Pair's take on speaker positioning, with a new ! low fat ! spreadsheet. You can have this bit of hackery served
stuffed or
zipped for Excel 5/Office 95 or later.
location, location, location
When the wavelength of a sound is smaller than the room in which the you are hearing it, the sound propagates in the room as discrete waves. If the wavelength (or multiple wavelengths) exactly coincides with one or more of the room dimensions, then a standing wave results.
Because standing waves represent fundamental modes of resonance in the room they tend to dominate the tonality of the space, since it takes little energy to excite a resonance. UK bathrooms are testament to this, with their hard reflective surfaces (tiles) and compact dimensions they tend to have modes which are reinforce the male singing voice: hence the army of bathtub singers.
Art Ludwig's Sound Page has an excellent introduction of the basic physics of sound and sound propagation, with lots of pretty pictures.
When a standing wave exists the node (null, point of zero air velocity) represents total cancellation, and the regularly spaced anti nodes are points of maximal vibration. Because the nodes are points of cancellation no amount of energy poured in at this point at this frequency will be audible since it just cancels out; another way of looking at is to consider that the impedance at the driving point is very high, meaning very little power can be transferred. Result, you can't hear that frequency anywhere in the room because there's (theoretically) zero acoustic energy transferred. This is why room correction EQ is very limited in scope because you can't fill a vacuum.
An example. Suppose a speaker is sat 1/3rd of the way along a room of length L metres; at the third harmonic of (L/340*3)hertz, and multiples thereof, this position will be a node meaning total cancellation of that frequency in that direction. A frequency sweep would show notches in the sound level at multiples of this frequency
|
Of course in the orthogonal direction the same thing happens at a different set of frequencies; and the same between floor and ceiling and all major subdivisions of the space, for example large items of furniture. All these notched responses are additive and set the basic acoustic field in which you are trying to listen to music, a field which is subdivided in three dimensions into a fine mesh of points of cancellation - the image left shows just one frequency in two dimensions. Put your acoustic source in one of the nodes and you are losing output to cancellation and interference effects. The result is that your expensive system sounds broken, especially if the room is nearly square: in which case modes across will be lumped together and the ripples in response caused much deeper. Loudspeaker positioning is the art of placing the transducers, so far as is possible, in one of the islands left over.
|
The rhythm method, our contribution to this morass, is to identify the principal axial modes which dominate this effect, and then use a simple function based on the golden mean to identify positions where the speaker's distance to a room boundary never coincides with the harmonic series of possible standing waves in the room. There's nothing magical about the golden mean, it's just a handy sorta transcendental number.
what to do & how to do it
Open the spreadsheet and have a play. In essence, all you need to enter are the dimensions of your target room - if it's not rectangular, or you have a large opening into an adjacent space, ignore that fact for now; we'll examine these effects on other pages. All you have to do is:
Fire up the spreadsheet and fill the target room dimensions in the yellow (metric) or blue (imperial) boxes.
Push the big button and peruse the results. The principal axial modes are shown, along with some suggested speaker positions for you to try.
Ignore the strange looks you get from your family as you mark out the floor with masking tape and spend the evening jumping up and down between pushing speakers around and listening to them.
Enjoy.
The fine tuning
If you decide to give the suggested speaker locations a go, bear in mind the following:
For our purposes speakers fall into three basic groups: free space (most monitors and floorstanding boxes); wall mount (Linn Kans, Naim Intro - DBLs); dipoles - Quad ESL, Martin Logan, Magneplanar. 'Omni-directional' speakers should be treated in the freespace category
Freespace speakers heed both dimensions.
Wallmount speakers need only heed the side wall distance, because the other dimension is a given
Speakers with a dipolar characteristic can ignore the side wall distance, because dipoles have negligible output 90 degrees off-axis. ESLs can even work pushed up against the side walls.
setup distances need to be accurate - aim to be within 10mm, measured to the centreline of the bass driver flush with the baffle. Yes you can hear the difference, an instance when a miss really is as good as a mile given how close some modes can be located.
If, in one dimension, the listening room does share a large opening with an adjacent space also try setting length/width to the combined total, i.e. dimension + distance to far wall visible through opening. This can give better results since most of the energy, and the most widely spaced modes, are in the bass that is have long wavelengths.
The positions the spreadsheet suggests are a best guess and can always be adjusted to taste. It just gives you somewhere to start.
If you have any comments, results - positive or negative - or problems with the spreadsheet please
let us know.
© the twisted pair 2001
20.12.01 Page added, downloads made available