Jim Thiel: A Coherent Source Page 5

Atkinson: However, you've more than doubled the moving mass.

Thiel: Yes, the sensitivity of the unit would, everything else being equal, be quite low. We've had to compensate for that by utilizing two very large magnets to achieve the sensitivity we need. For obvious reasons, the cost is high—the diaphragm is expensive, and the large magnet system is quite expensive. But you end up with very good results. The styrofoam both damps the resonances [that metal cones can have] and adds a lot of stiffness to the composite system so that the resonances are moved well above the main operating range of the driver.

Atkinson: So even though you're using first-order crossovers with their low rejection of out-of-band artifacts, at least near crossover, your metal cones are well behaved for a couple of octaves out of band.

Thiel: It's very important that the diaphragm response is well controlled to a much higher frequency than the crossover frequency.

Atkinson: The woofer in the CS6 uses a metal cone—is that stiffened in any way?

Thiel: No, it's not really needed. First, there's no necessity of making the cone flare shallower than that required for optimum strength. And second of all, even without any additional treatment, the internal resonance of that cone is approximately three octaves above the crossover frequency of the driver. So we can achieve essentially perfect results from the CS6 woofer with more-or-less conventional technology.

Atkinson: Will the coaxial mid/HF drive-unit appear in other Thiel designs?

Thiel: We are developing a CS7.2, which will replace the CS7. The '7.2 incorporates some things we've learned from and developed for the CS6. The drivers have all been redesigned from scratch, and we will be making them in-house. The '7.2 uses the same tweeter moving system that we developed for the CS6. All CS7s, by the way, will be retro-upgradeable to the '7.2 by replacing the drivers and replacing the crossover network. But there are no changes in the cabinet or the baffle, and the upgrade can be done by shipping the speakers back to the factory. The CS7.2, however, will be considerably more expensive than the CS7.

Atkinson: What goals regarding both loudspeaker and drive-unit design have you still to achieve?

Thiel: One thing that would be ideal is for all the drivers in the three- or four-way system to be coincident. So you would have not only a coaxial unit, you would have a tri-axial unit, or four drivers that are mounted coaxially. That would be an improvement, if you could pull off such a thing. There are always improvements to make, with better diaphragm materials, lower-distortion, higher-quality components, better cabinet construction methods. We just have to see what new ideas we come up with in the future.

Atkinson: Something that you obviously take very seriously is measurements. What measurements do you consider relate best to the perceived sound quality of a speaker?

Thiel: I believe if you have experience, you can relate most of the measurements that are taken to one or another aspect of the listening experience. Most obviously, the frequency response relates to the overall spectral balance of the loudspeaker. It also relates, if you know how to interpret it, to colorations that the speaker may have due to various resonances. Frequency response is the most useful tool. But many other measurements relate very well to different aspects of speaker performance. For example, cumulative decay spectra relate very well as a tool for objectifying the audible resonant behavior of speakers. The dispersion characteristics of a product will relate to the imaging performance of the product. Distortion measurements relate to how clean the product sounds.

But on the other hand, there are some sonic aspects of a product that we don't really have very good measurements to interpret. Particularly the ability of a speaker to resolve low-level detail. There are some measurements that relate to it . . . the cumulative decay spectra will relate to some degree because, if a speaker has resonances, they will tend to mask subtle musical details.

Atkinson: A criticism of high-end audio that's sometimes made is that using listening as an evaluation tool cannot be taken seriously because it can't be calibrated, it can't be consistent. Yet all successful high-end audio manufacturers, in my experience, do careful, consistent listening. How do you carry out your critical listening? Say a supplier comes along with a new component that you are interested in trying, and there would be a significant cost saving that you can pass on to your customer by using it. How do you assess whether using it would be good for you or not?

X