Bewaring of the Green
May 15 - June 15, 1990


The latest CD craze reflects badly on its adherents

by David Ranada

"The New, Confounding Green Marker Craze" was first nationally reported in the March 1990 issue of the monthly CD newsletter ICE (International CD Exchange) -- "It's being claimed that the sound of CDs can be significantly improved simply by marking the inner and outer edges of the disc with a green felt marking pen." The reaction within the audio community so far has been mixed, so we asked technical editor David Ranada to look into the matter and give us the final word on this whole controversy. His report follows:

A recent fetish sweeping through some audio circles may be leading many of its practitioners to sow unwittingly the seeds of destruction for their CD collections. These people -- many of whom ought to know better -- are applying permanent ink markers to their CDs' outer rims and the inside edges of the holes. Claiming that this process improves CD sound quality, they ignore the most basic principles behind a scientifically controlled listening test (the only kind of listening test producing conclusions that can be generalized to all listeners), and furthermore show a striking capacity for self-delusion. At the very least, the proposed physical means by which the treatment is supposed to operate displays a profound misunderstanding of the physical and mathematical principles behind digital audio and the compact disc system. And at its worst, it is possible that this procedure could permanently -- and fatally -- damage the very compact discs it is designed to improve.

I decided to give it a try anyway. It seems that the preferred color of the marker is green. But I used paint instead: Pactra Acrylic Enamel A46 bought at a hobby shop and applies with a small brush. And the color I chose was not green, but flat black, for reasons that will be explained later.

In several before-and-after, uncontrolled listening tests -- procedurally similar to those described in anecdotal accounts of the process -- I heard none of the various (sometimes inconsistent) effects attributed to it. Nor could I hear any sonic differences (let alone "improvements") when I compared two pressings of the same disc in two different CD players with instantaneous, level-matched switching between them. (There were very slight pre-existing sonic differences between the two players that I used, but these are audible only with "pink noise" test signals, and the sonic nature of the differences can be safely attributed to the known frequency response deviations of the two players. These differences are small enough not to be audible with music on treated or untreated discs.)

If indeed there are any sonic effects from the procedure, they cannot by any stretch of poetic license be described as "obvious" (or even "miniscule"), and could probably only be scientifically established by a series of double-blind controlled listening tests that I have neither the time nor inclination to perform. (A "double-blind" test would be one in which neither the person controlling the test nor the listener knows which disc is being played. The controls include the use of CD players with identical -- to within 0.1 db -- frequency responses, matched levels, and rapid, fraction-of-a-second switchover between discs.)

The burden of proof of the "green marker treatment" rests with its proponents, none of whom have reported the results of a double-blind controlled listening test establishing the objective reality of any improvement. There is, insread, every reason to believe that such a result would never be obtained, since the only explanation I've seen of how the treatment is supposed to work cannot be responsible for its claimed effectiveness.

It goes like this: The laser beam from the player, besides reflecting off the aluminized playing surface, gets scattered by the playback process and bounces around the inside of the disc. Eventually, this reflected light bounces off the edges of the disc and is reflected back into the disc, finally going back up into the laser pickup mechanism where somehow it mixes up the true digital-audio data, a sort of optical "multi-path" effect leading to diminished sound quality. The improvement comes from the light-absorbing power of the ink or paint, which is supposed to prevent the edge reflections from occurring. There are several important facts of life (laws of physics) that preclude this mechanism from operating. Each of the steps in the explanation is vulnerable to attack.

Light scattering is indeed fundamental to the operation of the CD system, since the thickness of the "pits" on a CD has been chosen to interact with the "color" of the playback laser and to scatter the laser light when it falls on one of them. The laser pickup therefore "sees" the digital audio data as a very rapid series of bright (light reflected) and dark (light scattered) flashes.

But for the green marker treatment to work, substantial amounts of unwanted laser light would have to be reflected back up through the pickup in untreated discs, a prospect I find very unlikely due to the light wavelength involved, the angles of reflection, and the number of reflections that would have to occur. Furthermore, the velocity of light is so high that before the pickup is finished scanning even one data pit, the light scattered at the beginning of that pit will have bounced off the edges and will have re-entered the laser mechanism -- where the deleterious effects would be immediately obvious in the pickup output. If multiple reflection pickup re-entry were truly a problem, practically every bit of data would be altered and the CD system wouldn't work at all!

Nonetheless, I decided to test for the presence of edge reflections in the pickup output. I painted only one half the circumference of a disc, and observed the direct output of the laser pickup on an oscilloscope (the so-called "eye pattern"). If there were a major effect on the laser output, it would have been visible as a once-per-revolution change in the eye pattern. I saw none. (Changes in eye pattern due to the processes used at different CD mastering and pressing plants are far more obvious than any possible changes induced by edge reflection. I haven't heard of any cults growing around the sonic quality of the pressings of a specific CD manufacturer, as there used to be with the LP, and there shouldn't be any.)

Even if there were strong edge reflections, there is no guarantee that coating a CD's edges with a randomly selected paint or ink would reduce or eliminate the reflections. The initial color proposed for the treatment was green, on the mistaken assumption that since the wavelength of the laser light used in CD players is in the infrared region of the spectrum, a green pigment would absorb it. Maybe so, but substantial infrared absorption is not an inevitable consequence of something appearing green. That's why I used black paint in my tests, on the assumption that an opaque black pigment is far more likely to be infrared-absorbing than semi-transparent marker ink. It certainly disrupts the laser beam more when painted on the surface of the disc.

Suppose substantial amounts of re-reflected light did get mixed into the true audio data in the laser pickup, without somehow producing an obviously disrupted eye pattern. The sensing system would faithfully pass such confused data to the player's error-correction system, and the scrambled information would signal its presence by an increase in the number of detected data errors. (The rate of undetected errors in the CD system is already so low that any reduction of it would have essentially no audible consequence.)

To check this possibility, I hooked up my trusty disc-error counter to my equally trusty first-generation CD player. I ran before-and-after correctable-error counts thirty times for a one-minute, twenty-second track near the edge of a 71-minute CD (surely the effect would be most obvious near a disc edge, where the returning reflections would be at their strongest). The average number of fully correctable data errors before treatment: 899.23, plus or minus 4.06. After treatment, the number rose to 909.13, plus or minus 5.34. This is essentially identical behavior, even though it looks like the error count increased. There were no uncorrectable errors (data interpolations) -- either before or after painting the edges.

Anybody who understands how the CD system operates knows that if the data reaching a player's digital-to-analog converters (DACs) is fully and successfully corrected, the digital portion of the entire CD signal chain is operating perfectly -- the data reaching the player's DACs is truly, exactly, precisely, and genuinely identical to the data generated by the encoding analog-to-digital converters. Just how well the DACs do the difficult job of turning the data into audio signals can indeed have audible consequences. But if the data entering the DACs is the same both before and after the treatment (as the lack of any data interpolations here shows), the DACs have no way of "knowing" that the disc has been treated, and the audio signal produced by them will be the same. If you hear a change, then, it is probably because you are imagining it.

This whole episode is discouraging to one who has spent the last several years trying to explain digital-audio principles to the listening public. So I'm proposing a counter-fetish with real teeth to it. It is conceivable -- and I checked this out with an engineer friend of mine who works at a midwestern CD factory -- that the solvents, surfactants, dyes, pigments, and other chemicals contained in paints and inks could dissolve or otherwise erode the clear lacquer that not only protects the "back" side of the CD from scratches but also prevents the thin aluminum layer beneath from oxidizing (one variety of "laser rot"). Discs with aluminization extending fully to the inner or outer edges must have the protective lacquer applied to their edges as well, so that oxidation cannot begin from the side. But if you paint over the edges of such a CD with the harsh chemicals in a marker pen, you may be inadvertently stripping away the disc's own protection, exposing the aluminum layer to oxygen and humidity and possible long-term irreversible deterioration.

It's possible that the chemicals used in common permanent markers won't do this, but the typical listener has no way of knowing if a given treatment is CD-safe. Determining that would require a battery of cycled-temperature/humidity accelerated-aging tests. Remember, it was only a few years ago that certain overseas CD plants ran into "laser rot" problems, caused by the supposedly safe inks they used in printing their CD labels. Why take chances with your investment in CDs? The long-term risks are not worth a nonexistent sonic return.

David Ranada, formerly Technical Editor of Stereo Review and High Fidelity, still contributes to Musical America and is a technical consultant to Curtco Publishing.