Interviews

Henry Azima: Absolutely. In fact, I was talking to another West German reviewer, Franz Schuller, who has reviewed 5000 CDs and has seen every plant in the world; he is probably the world authority on digitally recorded sound. The stories he was telling me were awful. It appears that while many recordings actually released on CD may have started with 15 to 16-bit resolution, when the data were finally put on the disc, the resolution was probably no better than 10-bit. They use third-generation tapes, they use bad printing, they fool around with the mixing, and they just destroy the sound. Franz was certain in his own mind that probably only 1% of the CDs you could buy, as of the end of 1986, are true 15 to 16-bit. The rest are far below.

John Atkinson: That's the way I understand it. You may start with a 16-bit master, but the minute you do any operations on the data, if you use a digital mixer, even if you do something as fundamental to recording as adding analog HF equalization or boosting gain, you reduce the effective resolution. In the digital domain, digital mixers, filters, and equalizers are based on the fact that the signal has mathematical operations performed on it, with the result that the resultant data consist of considerably more than 16 bits; when these long words are truncated back to 16-bit, unless done with care, large amounts of quantization distortion can be introduced.

Henry Azima: It depends how accurate these mathematical manipulations are. Like, if the accuracy of the signal being mixed with the audio stream is only 12-bit, then the end result cannot be better than the 12-bit.

John Atkinson: It means that an engineer designing playback machines is, to some extent, hung on the quality of disc.

Henry Azima: We are at the mercy of what's on the disc!

John Atkinson: Let's talk about Mission's current CD machines, the PCM 7000 and 4000. Almost three years ago Philips announced their second-generation chip set, based on a dual 16-bit DAC IC with 4x oversampling, but they apparently had tremendous problems in getting them into production compared with the tried and tested 14-bit technology. What does Mission do to ensure that the chips you get from Philips are good enough?

Henry Azima: We know that we get preselected chips because these chips go into all their machines. In fact, I think they grade them themselves for their cheaper machines and use the worst ones; we are assured by our contract with them that we get, preselected, the best available chips. Personally, I think we are far better off than with the 14-bit chip set.

In particular, the new digital filter which works with the 16-bit DAC is far superior to the old one. The impulse response, which is convoluted to get the low-pass filtering, is far more accurate; it gives just 0.02dB amplitude ripple in the passband, which is far superior than the 0.5dB we have with the 14-bit. It uses more coefficients and it's truncated better. The result is a far superior digital filter.

John Atkinson: Do you think that Philips' way of using oversampling is fundamentally more accurate than the general Japanese way (until this summer) of using straightforward 16-bit DACs running at 44.1kHz?

Henry Azima: I am totally convinced. There are two or three advantages, most important of which is the fact that you can get the out-of-band rejection without the problems associated with high-order analog filters. There are no problems with phase shift, because the filtering is done by a time-delay method. It's not at the mercy of the components or of temperature change, it's a very mathematical and very accurate way of really achieving the desired response. And with the new chip set, we get at least 60dB attenuation out-of-band; the previous set gave only 50 and better.

John Atkinson: Assuming that the political problems are resolved, does Mission intend to produce an R-DAT recorder?

Henry Azima: Mission is now looking very closely at producing one; I think we'll be probably one of the first manufacturers in the UK to make an R-DAT machine. At the moment, we are looking at many different sources in Japan from which to source our transport, and we have to compare and see which one would actually work out better.

Obviously, many people are afraid of R-DAT. It represents a big problem for the recording industry, because if the R-DAT actually comes to the market, people can produce cheap, cheap, cheap, copies of prerecorded music in Taiwan and Hong Kong which would completely destroy the royalties to the musicians. It would be extremely hard to control that. Very, very hard.

And I understand that the Japanese weren't ready to go into the market with R-DAT, in the sense that they try not to upset the market by making the man in the street misbelieve what he has been told. However, one day he's told CD is the best; the next day, he's told it's R-DAT, the reason being the fact that the Japanese factories which used to make video tape decks have no work. They need to make something to keep people in their jobs. The result is that people get really skeptical; it can actually terribly upset the marketplace.

But the fact of the matter is that these two media, CD and R-DAT, can exist side by side, as the turntable did with the cassette deck.

The problem, I think, is that the claims made for this new machine—that it is as good as CD and can do the same job—are not true. There are two or three parameters involved here: one is that the R-DAT, being a tape medium, will be at the mercy both of tape stretching, and of dropouts of the magnetic information. Second, because it's based on mechanical friction, it could actually deteriorate. There may be error detection and correction mechanisms inside, but it won't be the same. Whereas the CD, if you take care of it, can actually be a perpetual medium. Number three is that CD is true random access, whereas the tape is not. These important features, I think, will distinguish the two media.

I think there is no doubt, however, that R-DAT will replace the analog Compact Cassette within the next five to six years.