The Colossus of Codes
This month my concluding report on the 79th Audio Engineering Society convention will look at some new products I found particularly interesting.
At every AES convention, there is always something going on "behind the scenes," or at least some activity not on the exhibit floors. This time, in a suite on one of the upper floors of the New York Hilton, my dear friend and colleague John Eargle was demonstrating a fascinating new four-channel digital recorder.
The recorder, Colossus, is a product of a new company called By the Numbers. One of the principals of this company is Brad Miller, founder of Mobile Fidelity. You may recall that Mr. Miller became well known for his wonderfully exciting and highly realistic recordings of steam locomotives, many of which were issued by Mobile Fidelity. It was his desire for a compact multi-channel digital recorder that prompted him to enlist the aid of Lou Dorren. Back in the quadraphonic era, Dorren invented a format for quadraphonic FM that was widely acknowledged as the best sys tern for quadraphonic broadcasting.
Lou Dorren has been working in the computer field since that time, and his Colossus digital audio recorder apparently is derived from some of his experience there. For the Colossus, Dorren developed a proprietary code which is especially tailored to the video recording medium and which makes no use whatever of data reduction. Dorren claims his system makes more efficient use of the video medium than current video-based PCM digital recorders.
The Colossus works with profession al U-Matic video decks, and Dorren anticipates it will be able to interface with consumer videocassette recorders as well. The Colossus provides four digital channels with 16-bit resolution at a sampling rate of 50 kHz; band width is in excess of 20 kHz. A special model for scientific application is now being developed that has four channels, 12-bit resolution and a bandwidth of more than 100 kHz. This unit is in tended for such use as the recording of sounds made by whales and dolphins, which requires wide bandwidth but not wide dynamic range. An 8- to 12-channel recorder with 16-bit resolution and 20-kHz bandwidth is also envisioned. This would combine the video format for archival program storage with a high-density hard disk to facilitate overdubbing and insert recording.
Eargle was demonstrating the basic four-channel Colossus recorder through four JBL loudspeakers, with two stereo amplifiers. The indefatigable Brad Miller had taken the Colossus to the National Air Races at Reno, Nevada, and the sound of P-51 Mustangs screaming around the pylons was hair raising, an impressive testimonial to the capabilities of this digital recorder.
(Incidentally, Jack Renner of Telarc is currently testing one.) This four-channel recorder from By the Numbers ( P.O. Box 8359, Incline Village, Nev. 89450) is expected to cost $15,000, and production models are scheduled for delivery this spring.
Is it a direct attempt to resuscitate the moribund quadraphonic sound? Probably not, but with the availability of both CD and DAT formats, some form of multi-channel sound might eventually emerge. One obvious use for the Colossus four-channel recorder immediately comes to mind; the unit would be ideal for recording the four signals from the B-format outputs of the Calrec Soundfield microphone for Ambisonic recordings. Those signals are front/ back (called "X" in the Ambisonic sys tern), left/right ("Y"), omnidirectional ("W"), and vertical ("Z") for height information. If all these channels were properly reproduced, with an overhead loudspeaker included for the ht dimension, you would have full "periphonic" sound. Without the Z channel, the system can be reproduced as full horizontal surround or matrixed into a stereo signal which can be decoded into a surround signal. My opinion is that this type of recording will ultimately become very popular.
Dr. Stanley Lipshitz, that dynamic doyen of digital technology from the University of Waterloo in Ontario, has been avidly recording Ambisonic sound with the Calrec Soundfield microphone; at the Colossus demonstration, he evinced great interest in this system.
At the Meyer Sound Laboratories exhibit, John Meyer showed his CP-10 complementary-phase parametric equalizer. John, a brilliant young engineer with some highly original ideas about new technology, was my systems engineer at Crystal Clear. He designed the special low-noise preamplifiers and power supplies that made it possible to use the Bruel & Kjaer 4133 instrument microphone for music recording. (Nowadays, Mel & Kjaer produces several condenser microphones for recording, but back then we had to roll our own!) Meyer also designed high-intensity, very low-frequency speakers for playing the Apocalypse Now movie soundtrack.
The CP-10 provides five bands of parametric equalization, plus high- and low-cut shelving filters, for each channel. The unit employs complementary-phase circuitry, assuring minimum phase distortion even at extreme cut or boost settings. Under operating conditions, its dynamic range is more than 110 dB.
But what's really important about the CP-10 is its use as part of a unique measurement and compensation sys tern for sound equalization in concert halls, studios, and listening rooms. The system is called SIMCAD, an acronym for "source independent measurement computer aided devices." Meyer has devised a method of using a Hewlett-Packard 3582A dual-channel FFT spectrum analyzer, plus some proprietary equipment, to accurately measure the amplitude and phase response of sound systems in concert halls, using music as the test signal-and with an audience present! The advantages of this system are considerable. Obviously, a concert hall has a different acoustic character when an audience is present than it has when empty. However, since most sound-system/concert-haft measuring devices use test signals (such as swept sine waves or random noise) that audiences would find objection able, these measurements are per formed in empty halls.
In the SIMCAD system, once the sampling microphones have been selected, the analyzer is set for dual-channel operation in the "transfer function" mode using rms averaging. Two hundred averages can be obtained in a couple of minutes. When about half the audience has filed into the hall, a cassette tape of music is played as a test signal. Amplitude and phase measurements are used for preliminary set tings of the CP-10 to eliminate or re duce room resonances. When the hall is full and the music performance begins, the speed of measurement permits further fine-tuning of the CP-10.
Meyer tells me that slight tuning adjustments can even be made throughout a concert to compensate for changes in temperature and humidity.
With the SIMCAD system and the CP-10 equalizer, Meyer has been able to compensate and apply anti-resonance correction to many notoriously over-reverberant performing arenas and concert halls. He has accompanied the great Luciano Pavarotti on his U.S. concert tour, and Pavarotti has been delighted by the improved clarity and naturalness with which the system reproduces his performances. In an upcoming 1986 Pavarotti recital in Madison Square Garden in New York-a place that seats more than 20,000 people--Meyer will again be on hand with his SIMCAD system.
While the SIMCAD system has its major use in the concert hall, it can do some wonderful things in studios and in listening rooms as well. Meyer points out that after analysis, the anti-resonance equalization, plus the ability to suppress up to 40 mS of early reflections, can provide a superior listening environment, with the impression of a much larger acoustic space.
If one were to buy the $3,500 CP-10 from a Meyer dealer or sound contractor, the contractor would bring the HP FFT analyzer and other equipment to the customer, analyze the room, and set the complementary equalization to complete the process. Not inexpensive nor easy, but from all reports well worth the effort and expense!
(adapted from Audio magazine, Feb. 1986; Bert Whyte)
= = = =