An undated essay by the late Owsley “Bear” Stanley, sound man and LSD manufacturer for The Grateful Dead and the rest of the 1960s San Francisco counterculture scene. Taken from his website…
We commonly hear the remark that the digital sound on CD’s is inferior, or “inaccurate” to the sound of an analog vinyl LP which is made from a purely analog master tape.
What is said is partially true, the unplayed pressing of a “converted master”, that is, an acetate master cut on a mastering lathe, which is plated and that plating used to press the vinyl, and assuming that the vinyl is first quality virgin material, is very close to the master tape (note the phrase “very close”= not the same). Due to the mechanical reality of the process of making the disc, there are artifacts which aren’t in the original. That said, the real problem rears its ugly head: there isn’t any stylus which can accurately trace the grooves in the plastic record. The cutter uses a stylus which has very sharp corners (not surprising, since its job is to cut the plastic master), and therefore creates a groove which only a like-shaped stylus can trace perfectly. Unfortunately such a shape would simply reform (overcut) the groove into a straight furrow with no audio information remaining after its passing.
So you have a choice of two traditional stylus shapes to use for recovering the audio information from the grooves. One of these has a conical shape, and is usually called “spherical”. after the shape of the tip. This shape cannot come very close to following the movements of the cutter at any but the lowest frequencies. The other shape is a stylus which has an elliptical cross section, used with the major axis placed across the groove in an attempt to follow the cutter a bit more closely, but still quite inaccurate at the higher frequencies as well. Worse yet, both styli cause serious damage to the surface of the plastic inside the grooves. The friction of the stylus in the groove, exacerbated by the downward pressure required to keep it in the groove melts the plastic and so destroys the information on the sides of the groove. The damage is so severe (I’ve examined a lot of records under the microscope in the days when I produced the Old and in the Way LP) that you can only play the record once with any sort of fidelity with the elliptical point, and no more than 3 times with a spherical/conical.
In absolute terms the reproduction of even the best set-up has differences with the original recorded tape as much or more so than digital, only of a different kind, and somewhat “sweeter” in the ear — but inaccurate nontheless. Perhaps these people would be happier with a cassette made directly from the analog master, if such exists. In that case be sure that the cassette is either metal or genuine CrO2 tape, as the ferric formulations including high bias types won’t hold the highs for more than a few months.
The above comments about tapes are not to be extended to the digital tape format known as DAT. The recording of digital information on a magnetic medium is not a permanent way to store information. The information on a DAT tape is very short wavelength square waves, and the tape has a serious problem with self-erasure. The effect is somewhat more subtle than the more obvious loss of highs with the analog tapes, but is more difficult to deal with since it is a broad-spectrum type of information damage. The rule with DAT is to always re-copy them every 8 to 10 months to ensure that dropouts and data errors are minimised. Since this is obviously not going to be practical for most people the only way to store any significant amount of digital music is by burning it onto a CD-R disk.
So what good is vinyl? Something that you can have sitting in your living room as a curio to amuse and impress your friends? Turntables for playing LP’s that are any good cost a small fortune, both due to the mechanical difficulties of servoing that much mass and making it silky smooth, (a necessity to prevent rumble) but also due to the fact that only a very few extreme fanatics want them. The pickups are a real-world compromise in their mechanics, and very few are any good, and those few set you back big bucks, as well as introducing additional changes in the sound versus the tape originals. In addition, the resurgence in vinyl pressings seems almost exclusively confined to the Rap/Techno market and is driven by the DJ’s in that genre.
I agree that the sample rate chosen for encoding CD’s is far too low for the best fidelity, and I sincerely hope that a data rate of 200 to 500k will emerge with the introduction of the DVD, but at least the CD will always play the same each and every time you put it on the turntable. Why else do you think we use it? I can hear a lot of things in the music that has been digitized at 44.1k, don’t get me wrong, and in fact I am sure that I hear all of the artifacts (such as anti-aliasing) that the most vocal critics of the digital media are voicing. All things in the real world require some compromise.
Perhaps eventually some clever engineer will come up with a crossed-laser, non-contact pickup head for reading the vinyl grooves without causing meltdown. Not an impossible task, given the state of the art, but to take advantage of this you will need records that have never been played even once with a conventional pickup.