As We See It

This is quite a hypothesis, and certainly one that will gladden the heart of every digiphobe! But even assuming that Prof. Reilly's measurements and measuring techniques are beyond reproach, she is going to have a very hard time selling her hypothesis to the rest of the scientific community: on the face of it, the relationship she proposes is absurd.

Basically, what she asks us to believe is that "ultrasonic spikes" from digitally mastered records somehow reach the turntable's hardened steel bearing at sufficient intensity to crack them. Personally, I do not believe it.

Before I passed final judgment, however, I took the trouble to phone the good Professor. She sounded personable, intelligent, and just as puzzled as I about her own findings. I raised some specific questions; here are her responses:

Do you have objective evidence of the existence of these spikes in the playback signal from digitally mastered audio recordings? "Yes; we have oscilloscope traces."

How large are the spikes, relative to maximum recorded program level? "The spikes were 80mV, in the midst of program material that was only 20-30mV."

Are they random or cyclical in occurrence? "Cyclical."

They occur regularly? "Yes."

At what frequency? "Between 25 and 60kHz."

From what signal sources have you observed these spikes? "All digitally mastered recordings."

Audio compact cassettes too? "Yes." (footnote 1)

And CDs, of course. "Oh, yes. And analog mastered discs that had been fed through a digital delay line." (footnote 2)

How consistent was the relationship between playing digital discs and the increase in platter speed variation? "100% consistent. They all showed it."

Prof. Reilly went on to say that several turntable manufacturers were investigating her findings "with great interest." Which ones, I inquired? "Oh, there's no point in my telling you their names. They admit privately that I'm on to something, but they wouldn't admit it publicly." (footnote 3) (And they didn't. We mentioned Professor Reilly's claims to two of them; both expressed dismay at Professor Reilly's findings, and could find nothing in the performance of their turntables, whether of analog or digital background, to confirm her findings.)

One manufacturer was named during our conversation: Mark Levinson Audio Systems, who supposedly repaired a preamp of Professor Reilly's after it had been "damaged" by receiving the input from digitally mastered analog records. I called them the minute I got through talking to her.

According to Mark Levinson Audio Systems engineer Sid Chatterjee, the "repair" had been a routine updating, done under Prof. Reilly's strict instruction that no digital recordings be played even in the same room where her component was being "repaired."

Professor Reilly subsequently sent us a packet of material showing many speed-constancy measurements, microscope shots of microcracks on what were identified as platter spindles, and some poor Xeroxes of oscilloscope traces showing the large "ultrasonic spikes" protruding from what appeared to be musical program material.

Now, hard evidence is hard evidence, and Prof. Reilly appears to have more than enough to support her claim that something narsty is afoot. When honest research turns up an inexplicable relationship, it cannot be dismissed out of hand. If such a relationship does exist, I cannot hazard a guess as to what its cause might be; I must say I am very dubious about all this, for several reasons.

The mechanism described by Prof. Reilly violates everything we know (or, to be charitable, think we know) about physics in general and audio in particular.

Ultrasonic spikes have been observed by other researchers. The output from a wideband phono cartridge exhibits these spikes when the stylus mistracks; they have been measured at amplitudes as great as +50 dB at frequencies as high as 60kHz. They are not, however, part of the disc content—they are the result of mistracking during playback, impact pulses caused when a stylus momentarily loses contact with the groove wall, then slams back into it. There have not, however, been any independent confirmations of spikes (whether from mistracking or some other cause) from digitally mastered LPs and not from analog mastered LPs.

---

Footnote 1: Audio compact cassettes (the common, garden variety sold at K-Mart) are doing extraordinarily well to squeeze out 15kHz at -10dB level, let alone 40-60kHz at +50dB! Tape saturation at high frequencies ensures that no ultrasonic pulses such as those described by Prof. Reilly can possibly be gotten from the playback of a cassette, no matter how it was originally mastered.—J.Gordon Holt

Footnote 2: One of the best-kept secrets in audio has been the fact that some disc-cutting systems convert an analog signal into digital, put it into a holding buffer (to delay it so the automatic groove-pitch/depth control can anticipate it), then convert it back to analog. Thus, many analog-mastered discs on the market have actually been digitally processed along the way.—J.Gordon Holt

Footnote 3: In a subsequent conversation, Professor Reilly averred to me that Linn had talked to her about the problem, and was taking steps to address it, but didn't want their name associated with her findings. VPI was said to have "come to her" with problems they had found with their own products; Harry Weisfeld of VPI, however, stated unequivocally that he'd heard of Professor Reilly but never talked to her on any subject. Professor Reilly said that SOTA wanted to have nothing to do with her; this turned out to be true, but SOTA is trying to duplicate her findings, so far with no "success."—Larry Archibald