In general, how to get good dispersion?

[cuibono]

I'm contemplating a new speaker design, and want to take driver dispersion into more consideration .. something along the lines of Linkwitz's Orions.

One of the design criterion for the Orions was wide, smooth dispersion, and to me, the speaker shows its value. I understand the basic principle of why drivers start to beam - as frequencies increase, their wavelength decreases, and past a point about where the driver diameter is half the frequencies' wavelength, the speaker will become more directional .... So the question is, what can we do about it? I am particularly concerned about the upper midrange, where the mid driver would hand over to the tweeter. Lets pretend the crossover is somewhere in the 1k to 3k region.

What can be done to keep dispersion even in this region? Beyond a few, complicated horn design, I don't know of anything that can be done to 'standard' drivers to affect their dispersion. Supposedly, one would have better dispersion with a narrower midrange (I think), but Linkwitz ended up using an 8" driver for his mid (I guess he needed the extra surface area to allow a lower cross over point to the woofer..) He also magnet mounted the mid to avoid cabinet difraction, which seems related.

What can we do to control dispersion?

Anyone know anything about the 'acoustic aperature' on some Paradigm speakers?

[badman]

The easiest way to keep a smooth transition in terms of dispersion is to use a large multiway where no driver is operating much into directional territory, so like a 12 to 200HZ, a 6.5 to 1kHz, a small (think 2-3") mid 1k-3.5k or thereabouts and a small tweeter for the top 2 octaves.

But dispersion is a complicated issue, and there are problems with going massively multiway, complexity and cost being large among them. A good 2 way crossover is hard enough, imagine 5 way!

[Keith Taylor]

Cuibono, I think you are asking the most important question in speaker design. I would be inclined to rephrase it to "how to bring directivity under control". As to how directivity should vary with frequency is a different and much debated subject. What is not in dispute is that the off axis responses need to be smooth. With the currently available drivers we have devices that cannot change their acoustic size with frequency, so we have to resort to measures that impose a directivity on them. Dipole or cardiode operation and horns/waveguides are among the few options if you want to control directivity both horizontally and vertically.

Line arrays and related ideas such as MTM can make the horizontal directivity broad, if not even, but the vertical directivity means that you are in the "near field" over a lot of the frequency spectrum. It seems line array enthusiasts that do not get to hear a lot of live music can live with the unreality of being in the far field of the woofer(s) and the near field of the mids and upwards. I hope I don't start a flame war here, but never having heard a line array, not being golden eared, but hearing live music three times a week I came across a line array in an audio shop that had a long ribbon tweeter and a vertical array of mids. Moving around the room with the bass falling off at 6 Db per doubling of distance and the rest at 3 Db? was something I found very unnatural.

Being an owner of the Linkwitz Orions, I don't believe Linkwitz is sugesting directivity needs to be broad, just even. The metal cone mid driver is an interesting case in that conventional wisdom? has it that it is too big to sucessfully transition to a 1 inch dome tweeter. Linkwitz points out that the dipole operation of the mid gives it a wider pattern at fXO than if it were in a box.

One of the more interesting ideas for controlling directivity was described by a guy called Sandrick in a series of articles in Speaker Builder magazine a few years ago. He was building a speaker to enter in an AES sponsored competition, one of the rules being that it had to use passive crossovers. The design was all dipole, but instead of a pair of mids above and below a ribbon tweeter he had one mid above the tweeter and two on each side. He covered the two side mids with thick foam rubber to create a "natural" low pass filter. Thus the mid source changed its size with frequency in the horizontal plane.

Bye, Keith