out that the reflectivity of a metal can be changed by putting one or two clear layers next to the metal, and that the degree to which the reflectivity is changed is a function of the index or indices of refraction of those one or two layers of clear material. Thus, if the index of refraction of the plastic is slightly changed by the first strong hit of laser light, then the reflectivity of the metal would be different for subsequent hits of light (presumably the reflectivity would be lowered in this case, since we hear the sonic quality degrading for subsequent hits). Eventually the clear plastic, thanks to its molecular memory, would restore itself to its original molecular arrangement and original index of refraction, at which time that section of CD would sound fresh again.
      Once again, we see here that the CD medium is vulnerable to external analog influences. Again we see further evidence that the quality of the eye pattern, here determined by amount of reflectivity, does indeed influence the final sonic quality at the output of the CD player. And, once again, we see here that the digital medium's vulnerability is eerily reminiscent of the analog vinyl LP. As you know, the groove of a soft vinyl LP is physically distorted and fatigued by the first passage of the hard stylus. If you play the same section of an LP a second time within a short time period, it will not sound as good, because the groove wall is still distorted and fatigued. And it takes a number of hours for the molecular memory of plastic to bounce back and restore the groove back to its original physical geometry and original physical strength, after which the LP sounds once again almost as good for a subsequent playing as it did for the first.

-- Multiple Instances of One CD

      The focus of this article is on ways you can improve the sound from each single CD in your collection, so the examples just above, showing evidence that eye pattern quality affects sound, pertain to changes in the sound of a single CD. But perhaps the strongest evidence, that eye pattern quality does affect sound, comes not from comparing the sound of a single CD before and after, but rather from comparing the sound of multiple, alternative CD copies of what should be the same "digital" data and therefore should presumably sound identical. The fascinating finding is that they in fact don't sound identical, and in many cases the only possible (or most plausible) cause is a difference in the quality of the eye pattern.
      Our first example of this sort pertains to multiple instances of exactly the same CD. As we discussed in IAR issue 79, we compared two instances of a first edition classical CD from a major label, and they sounded very different in sonic quality. What could explain these sonic differences? CDs are plastic discs, pressed in hot molding machines, just as vinyl LPs are. The pits and pit edges that will create the analog eye pattern are mechanically molded or stamped into the plastic by a stamper mold. Since the title was a small run classical CD, the mother for any plural CD stampers was presumably the same. But stampers wear over time, as they press out hundreds of plastic CDs. The stamper that pressed out one instance of our classical CD could have been more worn than the other, or perhaps both instances of our CD title came from the same stamper, but one instance came from later in the life of that stamper, so it was more worn. A more stamper could well stamp out pits with edges that are more gradual, more sloppy, less sharply and cleanly defined. Alternatively, the thin flash of aluminum that does the reflecting might have been better quality or better applied on one CD instance than on the other. Or perhaps the polycarbonate substrate (body) in one CD instance was from the bottom of the plastic material barrel, so it did not transmit reflected light as well. Or, although both CDs were brand new, perhaps one had a thicker coating of mold release compound than the other, and thus did not reflect as strongly.
      In any case, the two CD instances had to be physically different in some way, since they sounded different. We just noted three possible physical differences above: poorer (less sharply defined) pit edges due to a worn stamper, poorer reflectivity due to inferior aluminum coating, and/or poorer reflectivity due to contaminated disc surface. Note that all these possibilities affect the quality of the analog eye pattern. So, whichever of these possibilities is the actual cause, they all support the same finding that's at issue here: the quality of the eye pattern does indeed affect the final sonic quality you hear.
      The popular misbelief is that multiple instances of the same CD should sound identical, because CDs contain digital data, and these multiple instances of the same CD contain identical digital data, and bits is bits, so they must sound the same. But now we know that the data are read from a CD in analog form, not digital form. And we know that analog factors like sharpness of molded pit edges and degree of reflectivity affect the quality of the analog waveform that is read from the CD. So in point of fact a CD contains analog information (including sharpness of pit edges and degree of reflectivity), not digital information. True, that analog information indirectly represents the amplitudes of music signal samples in digital format. But that analog information is still a very concrete, real intermediary between the abstract digital representation of music on the disc and the later digital interpretation and recreation of that digital information inside your CD player, which then subsequently re-converts its freshly created digital data back to analog.
      In a very real sense, the music signal, only abstractly represented as digital on the CD, goes through a digital-to-analog conversion as the concrete information is actually read off the CD, then an analog-to-digital conversion within your CD player, as your player interprets the incoming analog information to freshly create a digital bit stream, and then later another digital-to-analog conversion in your CD player's DAC chip, so that your CD player can output an analog music signal to your amplifier. Now, we know that each conversion of a signal, digital-to-analog or vice versa, brings with it the potential for degradation of the music, since each physically real conversion is less than perfect; so the mere fact that we now realize there are two more conversions than popular misbelief supposes is bad news, suggesting that laser media are less perfect than previously thought.
      But the real kicker here is that, thanks to this extra, heretofore unsuspected double conversion, your music signal is actually in analog form for an intermediate stage in the process. And analog signals are notoriously susceptible to being degraded by external analog influences. So, while your music is in analog form at this intermediate stage, its sonic quality is vulnerable to external analog variables, variables like degree of pit edge sharpness and degree of disc reflectivity, both of which affect the analog degree of quality of the analog eye pattern waveform actually read off the disc.
      Of course, the whole reason for going to a purportedly digital medium like CD, instead of an analog medium like vinyl LP, was because digital is supposedly robust and immune from external influences. But, by converting your music to analog at this intermediate stage in the chain, laser media like CD actually can make your music vulnerable to those same analog corruptions and variables we experienced with analog vinyl LP. One wonders what the design engineers were thinking (or if they were thinking), when they designed a new digital system to have similar analog vulnerability as a purely analog system. Hello? Is anybody in there?
      We all fondly remember how, in the old days, we'd compare two pressings of the same analog vinyl LP, and find they sounded different because one came from a more worn out stamper, or one had more mold release gunk, or was made from inferior reprocessed material (vinyl vs. now polycarbonate or reflective aluminum) instead of virgin materials. So it is bitterly nostalgic to find that CDs are just as vulnerable to these same analog influences, thanks to the extra conversions which make the signal analog for a while.

-- Burning a CD Copy

      Our second example pertains to bit-for-bit digital copies of a CD. We can make a freshly burned CD copy of a pressed and molded commercial CD. We can verify that the copy contains, bit-for-bit, representations of exactly the same digital data as the original (save perhaps for say one error per minute, which could not affect continuously ongoing audio quality). Yet, when we play both original and copy on the same CD player, they sound different. How can this be? It can't be different, if the process were entirely truly digital (and many engineers, believing the process to be wholly digital, think that those people who hear any differences are suffering from delusion). But the sound could very well be different, if the process has an analog intermediary stage, a stage where external analog influences could affect the music signal.
      This second example is especially powerful evidence in our discussion here because, when we freshly burn a CD copy, we independently know that there are some differences between the copy and the original. We know that they are the same digitally. But we also know that they are different in an analog manner (for instance, the reflective layer is different, and the pit edges are made differently). Because the differences between original and copy are only analog, and not digital, we can zero in, on our newly discovered analog intermediate stage in signal processing, as being the only point where the signal could possibly be vulnerable to external analog influences. And what form does the music signal have at this analog stage, the only possible point of influence? It is in the form of the analog eye pattern, as read from the CD. So the analog eye pattern logically must be the point at which external analog influences are able to wreak the sonic differences we in fact observe between original and copy. Thus, we have further logical evidence that the quality of the eye pattern does indeed make a sonic difference (the eye pattern is logically the only point in the chain where a quality difference from an analog external influence could affect the sonic outcome).
      Then, this logically derived evidence is further buttressed by empirical knowledge and evidence. As noted, we independently know that there are physical differences between the original and copy CD, and that these differences are all analog, not digital. But we also empirically know that the specific nature of these physical differences (different pit edges and different reflective layer) will affect the quality of the analog eye pattern. So, when we hear sonic differences between the original and copy, our empirical knowledge that the physical differences between original and copy will affect eye pattern quality is further evidence that eye pattern quality is in this case responsible for the sonic differences we hear, and therefore that eye pattern quality does indeed affect sonic quality.
      In other words, we have two kinds of support here evidencing the thesis that analog eye pattern quality affects sonic quality. There's the logical support, coming from the fact that the analog eye pattern stage is logically the only stage that could be affected by analog influences. And then there's also the empirical support, coming from our independent knowledge that the physical differences between the different sounding discs will affect eye pattern quality.
      A fascinating footnote adds even further evidence. Thus far, in this example, we've spoken only about there being a sonic difference between original and copy. The fascinating footnote is that the copy sounds not merely different than the original, but also better than the original. How on earth can a copy sound better than the original? Haven't we been rigorously taught that a copy, no matter how nearly perfect, can at best only hope to approach the quality of the original from beneath, but never surpass it?
      Well, here it turns out that we independently know a good reason why the copy could sound better than the original. The original has pit edges stamped out in a molding process, and by a worn mold at that. These molded plastic pit edges are likely to be dull and rounded corners, rather than sharp corners. And even the rounding of the edge is likely to be somewhat wavy, sloppy, and irregular, rather than clean and straight. All these properties of the molded type of pit edge conspire to produce an inferior quality eye pattern. On the other hand, the copy has pit edges freshly engraved by a laser, one of the sharpest cutting tools we have. So the pit edges on the copy are likely to be much sharper, and also much cleaner and straighter, thus producing a superior quality eye pattern when played back by your CD player.
      Note that the copy is bit for bit virtually identical to the original, in the digital data represented on the disc. But the actual manner in which that data is read is as an analog signal. And, even though the two discs might be digitally the same, they are not analog-ly the same. When it comes to what you hear as music, what counts of course is the signal picked up by your CD player at the actual real time of playback. The CD copy puts out a better analog pit edge signal than the CD original, so what guides and determines the quality of the music you hear when you play back the copy is the quality of the analog pit edge signal on the copy, not the quality of the analog pit edge signal on the original. You don't listen to the music while your CD burner (say your home PC) copied the disc; you only listen to it later. Your CD burner first read the original analog pit edges, then interpreted them into digital form, and then laser cut a fresh CD with better quality analog pit edges, which in turn give you a better quality eye pattern at time of actual music listening.
      Your CD burner executed an analog-to-digital-to-analog conversion, thereby converting an inferior analog pit edge signal into a superior one. If the new analog pit edges were created as an analog-to-analog copy, then the truism would probably still hold, that the copy cannot be any better than the original. But here the analog pit edges represent digital representations of music amplitude. So, by converting the poor quality analog pit edges to the digital form they're supposed to represent, your CD burner then has a chance to freshly create better quality pit edges as better analog representations of the digital representation (note our intentional double cascading of the concept of representation).
      By making a laser cut copy, you are trading one analog pit edge for a better one, effectively trading one instance of an analog medium for another better one. That's again just like the good old days of analog vinyl LPs, when you would trade a vinyl LP from mass pressing run, made on cheap reground vinyl, for an LP made on small custom run from premium virgin vinyl, which gave you a sharper, cleaner groove for a higher quality analog input signal. This means that you can profitably copy all of your CD library onto cheap CD blanks, play the copies in you CD player, and hear better music.
      This additional fact, that the copy not only sounds different but also actually sounds better than the original CD, is dramatic strong further evidence of our finding in this chapter, that quality of pit edge, which is critical to determining the quality of the eye pattern, is important to the ultimate sound you hear. Laser media, in spite of their touted purported digital nature, must indeed be vulnerable to the analog quality of this analog eye pattern signal, which depends upon the analog quality of the analog pit edges.
      It's also worth noting here (in passing at this point) a fact which will be amplified upon below. The CD burner presumably was able to read all the amplitude data correctly from the original CD, even with its inferior quality pit edges (that's why it was able to make a correct bit-for-bit copy of the digital data represented on the original CD by the analog pit edges). We can deduce from this that the reason, why music from your CD player is vulnerable to quality differences in the input analog eye pattern, cannot be that poor quality eye patterns somehow cause your CD player to get the amplitude data wrong for its re-creation of your music. In other words, we have abundant anecdotal evidence that a poor quality eye pattern audibly degrades the final music signal, but the reason why this is so cannot be that the amplitude axis of your music signal, as created by your CD player, is somehow wrong due to misinterpretation of this poor quality analog input signal. A poor quality analog eye pattern input signal does not, according to our reasoning here, somehow corrupt the amplitude axis of the music signal you hear (unless the CD is so dirty that your CD player's error correction fails and your CD player has to go into interpolation mode). Well, the music waveform you hear, as created by your CD player, has only two axes. One axis is amplitude and the other axis is time. So, if the amplitude axis cannot be corrupted (except by a very dirty CD), and we know from listening that something is getting corrupted, then it logically follows that it must be the time axis that is getting corrupted. Therefore, this peculiar phenomenon, that a CD copy sounds different from and better than the original, gives us logical evidence to deduce here that the mechanism of CD's analog-like vulnerability to its analog eye pattern input signal must have something to do with corruption of the music's time axis, as the CD player creates that axis for the music you hear.

-- Listening to a Computer Hard Disk

      There's even more evidence about the sonic importance of analog eye pattern quality, relating to sonic quality improvements realizable by copying CDs. Peter McGrath, noted golden eared audio pro who works at the cutting edge of high quality sound, has found that you can get even better sound by copying a CD not to another freshly burned CD, but rather to a computer hard disk, and then playing back your music, as you listen to it in real time, from your hard disk.
      Now, why should this technique yield sound that is any different, let alone better? Again, proponents of the bits-is-bits school would scoff at such a proposition, since the digital amplitude data on the hard disk is virtually exactly the same as on the original CD. Even those of us who understand the previously explained phenomenon, about sharper pit edges on a laser burned CD copy sounding better than the molded pit edges on the original CD, might have trouble with this. After all, when you burned your CD copy on your home PC, the digital data from the original CD was temporarily stored on your PC's hard disk before the CD copy was burned. So how could this intermediate stage possibly sound any better?
      The answer to this question is fascinating, and furnishes important further evidence that a poor quality analog eye pattern from a CD does indeed affect the final music you hear. Remember first that what counts in music playback is the nature of the chain in real time, at the time you're actually listening to the music. When McGrath plays back the digital data from the hard disk in real time, what's happening at that time? What is the quality of the analog pit edges at that time? What is the quality of the analog eye pattern at that time? The answer is that there are no pit edges, there is no eye pattern. Playing back the music directly from hard disk gets rid of the pit edges and eye pattern entirely. There are no pit edges and no analog eye pattern that could have analog vulnerability to external influences.

(Continued on page 58)