Jim Thiel: A Coherent Source Page 6

Thiel: I view the ear as a measurement tool. And you can use it to evaluate the differences between the sound of loudspeaker designs just as usefully as you can use a computer to evaluate objective differences. Take the different materials that can be used for woofer chassis, for example. We use cast aluminum chassis in all of our woofers. I know of no measurement that can tell you that this driver basket will sound better than that one. But in fact, it's very easily audible. Cast aluminum sounds much better than stamped steel. Now maybe it's because aluminum's nonmagnetic and there's less interaction with the magnet system; maybe it's because its resonant behavior is different. It doesn't really matter to me necessarily what the reasons are. The only thing I have to do is to determine for myself if the performance is better and make a decision, based on performance and cost, whether I want to incorporate it into my design.

I have an advantage over users of loudspeakers, or even reviewers of loudspeakers, because I've been able to build up a lot of experience that can correlate measurements with listening. I can make a change to the crossover circuitry and measure the effect that has on performance, then go and listen to what effect it has on the subjective performance. And I've done that tens of thousands of times over the years. You can build up quite an understanding of the correlations between measurements and subjective impressions.

Atkinson: How does it make you feel when you have to deal with issues for which you don't have any engineering tools?

Thiel: I enjoy it. If I could do all of the engineering work with computers and calculations, I would not find it as enjoyable. I like that there's also an intuitive aspect, where you have to incorporate the knowledge that you've gained from experience in addition to what you objectively know. Just because you're doing a subjective evaluation does not mean it's incorrect or different for each person.

My task as a loudspeaker engineer/designer is not to prove to skeptics that these things matter. I don't care if they believe it or not. And I'm not restricted to only using tools that I can prove to them are correct. I only have to prove to myself that these differences are important.

Atkinson: And your chief financial officer.

Thiel: And my customers! [laughs]

Atkinson: What are your priorities in sound quality? Are you trying to get high sensitivity above everything else? Or do you start by trying to get the best stereo imaging? Or are you going for tonal neutrality? Which is most important?

Thiel: I like to think that there's not one that's most important, that there are four broad performance areas that are all important: the tonal fidelity, the spatial fidelity, the transient fidelity, and the dynamic fidelity. There are examples on the market of speakers that are very good in one or two of these areas but poor in other areas, and I don't think that any of those products are truly good products for most people. If a product has a serious limitation in even one of those areas, it makes it very difficult for most people to truly enjoy the musical performance of the product, I feel.

I'd rather have a product that was merely "good" in all respects than one that was great in one respect and poor in another respect. (Of course, I'd prefer to have a product that's great in every respect.) For example, I'm not willing to forgive tonal-response errors in a product just because it images very, very well. Imaging in itself is not a sufficient reason to like a product. And the same would be true of every other performance aspect. I think a speaker needs to be very good in all respects, or at least equally good in all respects.

Atkinson: One thing you find with loudspeakers that are overdeveloped in one area but lacking in others is that they start to drastically restrict the types of music that their owner plays.

Thiel: Our goal is the opposite—we want to open the customer up to more musical enjoyment of more types of music.

Atkinson: But as a designer, you don't have an unlimited budget for any of your models, even the most expensive. How do you compromise what you are trying to achieve with the CS6 in order to get a much-more-affordable speaker like the little CS.5?

Thiel: The first thing we'll sacrifice is quantity rather than quality. The first thing I trade off is how loudly will the speaker play, and how big of a room is it intended to play in. The next thing I will trade off is bass extension. I'll be willing to trade off a half-octave, or an octave, or one and a half octaves of deep bass. And at this point, I don't feel that I have yet sacrificed anything in terms of the sound quality. You can go quite far in reducing costs by trading off quantity.

Atkinson: Presumably because there's not a linear relation between bass extension and cost. It's probably more like a square law. Halve the low-frequency extension and you reduce the cost by a factor of four.