Paul Hales: A Passion for Speakers Page 5
Deutsch: Do you mean on-axis frequency response?
Hales: Yes. The off-axis frequency response is also hugely important, but, generally speaking, the first thing I do is get the on-axis frequency response correct. I have enough experience now that, with given drivers and given baffles and box geometry, I can anticipate what I'm going to get. I can tell by looking at a driver's natural response how much I'll be able to manipulate it to achieve a flat characteristic. And one of the criteria for the Revelation series' driver-selection process was a very smooth, natural driver response. That's how I achieved the less than ±1dB variation.
Deutsch: Have you had a situation where a combination that worked on paper somehow just didn't sound right?
Hales: Yes. Absolutely. That happens all the time. But every once in a while you get a combination, or you'll get a driver that should perform to a certain standard and it just doesn't. I had that in the Revelation product. I had a very well-known, highly respected tweeter that looked fantastic on paper. When I tried to implement it in the system, I got very nice measured results. Sonically, it was clearly inaccurate. I had to abandon that one and use something else.
Deutsch: How do you make that decision?
Hales: I firmly believe you have to have more than one set of ears. My good friend Larry, a retailer, and I spend a huge number of hours listening. And a lot of time modifying and tweaking. I find if you don't have at least one more opinion—and sometimes it takes two or three opinions—it's very easy to get confused and to get focused on one aspect of performance. You unintentionally ignore something else that may be important.
Deutsch: Do you do blind comparisons?
Hales: No blind listening at all, actually. Sometimes it's so late at night that I'm blind, but it's not intentional...We have a couple dozen CDs that we use every time and that we feel very intimate with, that we know what we're supposed to achieve with in terms of soundstaging, tonal balance, bass response, and so on.
Deutsch: Yesterday we were listening to a prototype of the Revelation Two, and you said, "Yes, that sounds good, but there's something still not quite right."
Hales: Compared to what I know the recordings to sound like, the Revelation Two prototype sounded real good, but was not quite getting a couple of important aspects of performance that I know could be there. We're not getting a layering in the soundstage—everything sounded kind of flat. I know what to do to get back the layering. I have a "layering" knob—I turn it up to 10.
Deutsch: Why don't you turn all the knobs to 10?
Hales: Because when you turn this knob to 10, if you're not paying attention, another setting goes down to 5. They're all interrelated.
Deutsch: Like a video monitor with interacting brightness and contrast controls?
Hales: That's right. You can't turn up the layering knob without it also affecting the timbre knob. You don't want to screw up the timbre in an effort to get the layering, so you might have to turn this knob two notches, but back this one off a notch in order to retain a certain quality. Because there's a huge number of variables, it's very difficult getting it all just right. That's why it takes a long time.
Deutsch: Are you talking about tweaking the crossover?
Hales: With my computer tools and the measurement techniques that I've developed, I can measure the drivers, design a crossover, and get a flat-measuring loudspeaker in half a day. From then on it's a subtle process, based mostly on subjective evaluation, but also objective. We'll do this for up to 12 hours in a row. Because once you get started, you can't really stop and expect to pick up the next day right where you left off. It takes a great deal of time to get tuned in to exactly what's happening from a sonic standpoint as well as an engineering standpoint. The Revelation Three was one of the first models to have a fantastic transfer function.