Interviews

Deutsch: Do you use your own measuring systems exclusively? Or do you also use some of the commercially available systems?

Tchilinguirian: We've got MLSSA, Brüel & Kjaer, and LMS, so we've got virtually every commercial system.

Deutsch: Which do you find most useful?

Tchilinguirian: Ours! I'm not saying this because I'm biased, but it's very quick. When you can look at all that information simultaneously, it saves time in having to go and turn the speaker and measure, turn and measure, etc. Our measurement system has eight channels, so you can look at eight curves simultaneously. And at the stage where we're developing the crossover and drivers—looking at their off-axis performance—it's probably the most useful equipment we have.

Deutsch: You mentioned correct phase response as one of the design goals. A couple of years ago, in an interview published in Stereophile (footnote 2), Kevin Voecks [then with Snell, now with Revel—Ed.] said that, although phase response is measurable and people can discriminate phase with tests involving sinewaves, it's really not very important with music signals. His view is that you should aim for flat amplitude response, and if that means having drivers that are out of phase, no big deal. Would you agree with that?

Tchilinguirian: It may not be the most important thing.

Deutsch: Would you give up phase response if it gave you better amplitude response?

Tchilinguirian: It depends on what else was going on...why the phase response is out. If the phase between two drivers is incorrect, you've got to see it on the amplitude response—they're not adding correctly. Whether it's on- or off-axis vertically, you have to see that.

Deutsch: Isn't it possible, even if the amplitude response is flat, that the drivers are adding correctly in the amplitude domain but may be out of phase with each other?

Tchilinguirian: Then they've got to start subtracting above or below the crossover point. And maybe you can pick a point in space where they're all adding, but there will be problems elsewhere. For example, if done correctly, a first-order crossover will give you good phase response and flat amplitude response on-axis. But measure it in the vertical plane, and you're going to see some problems.

Deutsch: You're going to get lobing effects.

Tchilinguirian: Right. With the Veritas, you can measure on-axis and look at the individual driver output as well as the overall response, and you can see that they're adding very well, even the woofer and tweeter. You can then measure 15 degrees up and 15 degrees down, and there's very little lobing. I've measured a number of speakers that have huge dips as you measure up and down. Typically, first-order networks give you that. That's one of the reasons we selected the fourth-order slopes on the Veritas. Another reason was to try to get the tweeter output and the midrange output below the crossover point down as quickly as possible. That increases the dynamic range and reduces distortion. So every time we think of a solution to a problem, we look at what else it will do—what are the other things that will have negative effects?

We want to come up with a solution that has the fewest possible tradeoffs. We're not going to try to get perfect phase relationships if it's going to really degrade the vertical dispersion.

Deutsch: What tradeoffs have you had to make in developing the Veritas?

Tchilinguirian: The biggest thing is size. I'd love to develop a $6000 loudspeaker that would go down to 17Hz, at good levels. Unfortunately, it would be very large. There are other ways you could do that—still, its cabinet would be much larger than that of the Veritas. And there may be some things in the future with Energy—at a higher price, of course—broadening the spectrum right down to the lowest frequencies. But for now we wanted a loudspeaker that would cover a very large bandwidth while still keeping its size as small as possible.

Deutsch: So one of your concerns was how it would look in the average living or listening room?

---

Footnote 2: Vol.13 No.3, March 1990, p.100.—John Atkinson