(Continued from previous page)
WSR Reber: Now, in the Casablanca II we use here in our reference system, for example, you offer a Standard set of D-to-A converters...

Cardas: We offer four quality levels of converter in Casa II. The least costly, Standard, is single-ended. The others offer balanced.

WSR Reber: What's the benefit of going from single-ended RCA to a balanced XLR format?

Cardas: In a true differential balanced circuit, like all balanced Casablanca converters, the active signal is essentially doubled. Both the positive and negative are processed through complete and identical circuits.

Sinclair: You have the same signal in negative and in positive. We do this from the digital input to the analog output. This cancels spurious noise and distortion in both the digital and analog domains.

Cardas: Common mode rejection is the technical term.

Sinclair: In a Theta system, we do this all the way through to the output of our amplifiers. From our flagship Citadel to the Intrepid, all Theta amps are balanced input to output. They've done this in recording studios for years, but only recently in consumer A/V.

WSR Reber: What's the benefit in sound quality?

Sinclair: There's more harmonic right-ness, more of a "you are there" feeling. But it has to be done right. There are products with XLR connectors that are not truly balanced differential circuits; they just have a phase-splitting IC at the end. You must truly double the circuit.

Cardas: You have more control in a fully differential balanced circuit. The circuits see if errors have been created and eliminate them.

WSR Reber: So the presence of an XLR output doesn't guarantee it's a true differential balanced circuit, as you described?

Cardas & Sinclair: Correct.

WSR Reber: And while a consumer wouldn't necessarily know this, I'm sure manufacturers advertise balanced, whether it is true differential balanced?

Cardas: Yes they do.

Sinclair: You don't pick up extra noise in the cables, if that would otherwise be a problem, but none of the other advantages of balanced circuits are there, and you've added extra circuitry, which is likely to give its own coloration and noise to the sound.

WSR Reber: While we're on D-to-A con-version, what is the sound quality difference between the linear multi-bit DACs used by Theta Digital and Sigma/Delta 1-bit DACs?

Sinclair: While Theta uses a variant of a sigma/delta in its least expensive D-to-A boards, and multi-bit DACs in its more expensive boards, I would not draw a conclusion on relative sound quality about this. There are good and poor sounding DACs of both types and of all bit depths. If I were to worry about a specific part in the conversion chain, it would be the digital filter. In our Superior and Xtreme DAC's we use programmable filters. They're much longer, more complete, more intelligent filters than what's available in a stock filter chip.

WSR Reber: What are they filtering?

Sinclair: They're filtering the spurious noise created by the digital signal. Are Widescreen Review readers familiar with the Nyquist Theorem?

WSR Reber: Many, but not all.

Sinclair: Beyond its most basic function of keeping the spurious noise residing above the pass band from affecting the audible signal, you have to think about the digital filter as a reconstruction filter. It's about putting the signal together again. A really intelligent filter adapts the signal to the waveform coming through. Simple ready-made filters do not do this.

WSR Reber: So in going from Standard to Xtreme, you basically get a more sophisticated filter design and a higher bit rate?

Sinclair: And a better DAC. The higher bit rate isn't really the issue.

WSR Reber: Then to say 18 bits versus 20 bits, or 24 bits, doesn't mean anything?

Cardas: It's an easy thing to put on a checklist, but it doesn't really say anything about the sound. From a practical standpoint, if you want to play 24-bit/96 kHz material with a Casablanca, you need at least a Superior DAC, but that's not the reason one sounds better than the other.

Sinclair: There are really lousy sounding 24-bit DACs, and pretty damn good 16-bit DACs. It's the same kind of pitfall as buying an amp on power output alone. The 20-bit DAC in our Superior DAC boards will resolve as many bits as the 24-bit DACs in our new Xtreme boards. Now, the Xtreme sounds better, but has the same resolving power--which is what you figure to get with more bits, but you don't necessarily. Because something says it has 24 bits doesn't mean it has 24 bits of resolving power, it just means it operates at 24 bits.

WSR Reber: The Xtreme DACs, which we installed in our Casablanca II today, operate at 24 bits. The reason they sound better is not because of the increase in bits-but what? Is it the filtering?

Sinclair: The DAC is a better DAC than the previous ones. That it operates at 24 bits doesn't hurt, but the point is that it's a truly great 24-bit DAC. Also, the digital filter is better than the one on the Superior board. And the DAC and the first amplifier stage are on a Teflon board.

WSR Reber: So, we've gone through these stages, and we're at the back-to-analog phase. What happens now?

Sinclair: At the end, there's a current-to--voltage stage and a buffered output, and then you're looking into the amplifier.

Cardas: Through very high-quality cables.

Sinclair: Can you tell she grew up in the cable business?!

Cardas: Some other elements should also be noted. For example, our power supply. Using multiple transformers allows each processing stage to have its own clean, unfettered source of power. We do this in the lower priced Casa Nova processor as well as in the Casablanca. These are linear transformers, not little switchers which have their own radiation problems.

Once in a while I look at a competitor's product. We open it up and find they have one unshielded switching power supply. This is something we never do.

Sinclair: Gary, the next time you have a competing product, see what's in there. You'll be shocked.

Cardas: Each step sees its own clean power, and each step sees, theoretically, limitless power. So it can draw whatever it needs without interfering with whatever else is going on elsewhere in the processor.

WSR Reber: So the design process is a science and an art that started back at the inception of digital, and you've learned, through your experiences and experimenting and testing, which approaches work better than others?

Sinclair: Precisely. While a lot of what we do is science, a lot is a body of 'witchcraft' that has been built up over the years, where we've found something that works, but can't explain why. Most of this knowledge we don't talk about. We also still have Dave Kerstetter, our original head engineer who started Theta's first project 15 years ago.

WSR Reber: One thing consumers may notice is that some processors, preamps, and DVD players are made of flimsy plastic. They feel very light, and connectors sort of wobble when you attach a big cable. They have poor build quality. Where does that come into play in terms of absolute performance?

Sinclair: Some of it is performance-related, and some affects longevity. The kind of products we're postulating are expensive no matter what we do, so it seems necessary to build them robustly, so end users can keep them for many, many years.

Cardas: Part of the weight issue is the transformers we talked about. The transformer brings heft, and the more transformers, the heavier the weight. We get that weight through legitimate means, by actually having more transformers. Part of it is also related to resonance. You have to have enough mass to keep resonance's down. Then there is the shielding issue. We want to keep outside radiation from getting in, and inside radiation from getting out.

Sinclair: It's really marvelous that the large, multinational companies can bring out moderate-cost equipment that serves most people well. The people who are interested in Theta equipment are perfectionists who want to stay at the leading edge for many years--that's why they buy our stuff. It's a different philosophy.

WSR Reber: The amplification components in an audio system raise the low-level signal from the source components to a level loudspeakers can convert to acoustical energy. The preamplifier stages in surround processors or controllers provide switching, buffering, and volume control for six or more channels, along with some low-level signal amplification. The power amplifier receives the preamps' elevated signal and further "amplifies" it to a level that can drive loudspeakers. An accurate amplifier delivers an accurate replica of the pre-amplifier output to the loudspeakers. The amplification process occurs continuously as the signal passes from one stage to another, until it is sufficient to drive loudspeakers. The performance of the loudspeakers depends on the condition of the electrical signal they receive, after numerous stages of amplification through all the other components in the system. What is it about your amplifier approach that distinguishes Theta Digital?

Sinclair: Two things. As we mentioned with low-level amplification, we believe in differential balance, and our amplifiers are balanced input to output. This means there are really two channels of amplification for every channel. One channel for each phase angle, each one looking at the other, and doing differential correction, or common mode rejection, of any distortions that are there. The other key point is we believe in as little negative feedback as possible. None of our amplifiers use any overall negative feedback.

There are two reasons feedback is common in audio amplifiers. It's hard to stabilize an amp, to keep it from oscillating and blowing itself up without overall feedback. Also, while overall negative feedback can radically lower the measured distortion, we found that whenever overall negative feedback was used, it could be heard in a predictable way, in a more audible way than you could hear measured distortion. We worked hard to create stable designs that did not need overall feedback. To understand this you have to understand how feedback works.

Page Number: 1 | 2 | 3 | 4 | 5