Stop Digital Madness! Professor Reilly responds
Dear Mr. Holt:
Enclosed is my rebuttal to J. Gordon Holt's "Stop Digital Madness" ("As We See It," Vol.8 No.8). In spite of being tremendously upset and angered by the manner in which my work was reported, by the errors contained therein, and by the shoddy reproduction of the graphs, I have kept my rebuttal at an unemotional, professional level.
To place an antidigital label upon me is entirely without foundation. In trying to restore my music system to listenable quality, after going through a large number of components and finding that no one—dealer, manufacturer, etc.—could provide a system that remained listenable over a sustained period of time, I stumbled upon the possible digital connection wherein ultrasound apparently played a key role. Approximately 20,000 hours have been consumed in trying to reach the truth and in exploring other possibilities (footnote 1).
I prefer to accept what happened as misjudgment that I was part of the rabid antidigital crusade. For me, as for you, well-reproduced music, however it is achieved—digital or otherwise—is the important issue. I cannot emphasize strongly enough my horror upon finding what the ultrasound from digital recordings was apparently doing. Silence at this point would have been unconscionable.
Sincerely yours, Judith G. Reilly
Judith Reilly's Response to "Stop Digital Madness"
The appearance of the issue of Stereophile containing "Stop Digital Madness" at the [January 1986] Consumer Electronics Show in Las Vegas served as a catalyst for many constructive conversations. Much interest was aroused concerning my studies on turntable speed variations and their possible link with ultrasound from digital recordings. However, this Stereophile article contains some serious misinterpretations of my research and the theories developed therefrom, which must be addressed.
For the record, ultrasound, which we believe plays such a key role in turntable speed variations, was postulated as a result of sonic degrade (footnote 2) in several of the preamps I owned. The idea of spurious, above-band signals associated with the digitization process was advanced at the time. This work was indeed first mentioned in Fanfare, January/February 1984, by Neil Levenson. Spuriae in the form of ultrasound were subsequently identified by others. These were reported by Levenson in Fanfare, May/June 1984. Since then, independent tests have confirmed that ultrasonic information accompanies digitized music.
As an attempt to control variables, all of the components in my system were replaced, thereby removing several layers of sonic degrade. However, there remained a smear or blurry quality that overlaid the music, particularly classical piano or operatic vocal. After several months of trying a fair cross-sampling of quality cartridges, I eliminated them as the probable source of smear. This left only the turntable, which had played some digital records. Since speed variations seemed to be the only way to account for musical smear, I then began to measure turntable speeds and found corroboration with the hypothesis. Several hundred 'tables were studied, some of them for months.
After thousands of hours of measurements, patterns in the data became evident, from which we could draw conclusions. Tables which had played no or few hours of digital recordings did not exhibit the type of short-term speed inconstancy observed with 'tables which had. If there were cooling currents or if the temperature was reduced (usually to 65-67 degrees F for direct drives, or to the low '60s for belt drives), the speed fluctuations would generally abate. These correlations led us to question the mechanism involved.
A literature search revealed that ultrasonic waves in the frequency range common to the spuriae from a digital recording do create microcracks in crystalline materials, including metals. To the doubters let me suggest reading W.P. Mason's "Acoustical Properties of Solids" contained in New Directions in Physical Acoustics (North-Holland Publishing Co., 1976). The manner in which ultrasound creates microcracks can be readily understood with dislocation theory from the field of metallurgy. Crystals of all kinds almost always contain imperfections called dislocations.
(There are many books on this topic including A.H. Cottrell, Theoretical Structural Metallurgy (St. Martin's Press, 1959), and W.P. Mason, Physical Acoustics and the Properties of Solids (D. Van Nostrand, 1958). When crystals are stressed, say by ultrasound, these imperfections move and pile up at metallic grain boundaries, ultimately leading to the development of microcracks. This behavior of dislocations has been widely documented using very high magnification electron microscopes. (A number of books by Gareth Thomas, as well as a myriad others, contain micrographs of dislocations engaging in these various activities.)
Even though the momentary amplitude of the ultrasound reaching the shaft must be small, ultrasonic pulses are always arriving at the shaft, causing a cumulative effect. The literature on dislocations repeatedly states that very small stresses ultimately will produce these effects. Under such stress atomic and molecular bonds break consecutively rather than simultaneously, thus requiring less energy. In the nondestructive testing with ultrasound referred to in the article, short-duration, much more energetic signals are used. Detection of larger flaws, not tiny microcracks, is the objective. Consequently, the brief exposure to ultrasound in this technique minimizes large-scale microcracking.
Footnote 1: We hope that all 20,000 hours were not spent by Prof. Reilly alone; that amount of time represents 9.6 years of solid 40-hour weeks in addition to her full-time job at Quinsigamond Community College.—Larry Archibald
Footnote 2: "Degrade" does not appear as a noun in any of our dictionaries, but in keeping with Prof. Reilly's request that we not change any of her words we are leaving it as is.—J. Gordon Holt