Vinyl Stereo Encoding -- Not That Anybody Cares

Audio on vinyl is having a resurgence among audiophiles and millennials, despite the fact that CD audio, and higher resolution audio on DVDs and Blu Ray is orders of magnitude better in every measurable way. It is also more convenient, generally skip-proof (in memory and solid state drives - and if you don't like MP3 compressed, there's always lossless).

So why is vinyl still popular? Well, it's fun! I enjoy listening to my old records. I bought them in the 1960s, '70s and '80s, and never replaced them with CDs. Many were never even available in digital formats, except the ones I transcribe myself. And frankly, most of them still sound as good as the day I bought them. I always played them on decent equipment. But there's more to it than that.

By the time digital audio hit the scene, vinyl (or analog) recording state-of-the-art had reached a very high level of sophistication. Nothing like the 96 dB dynamic range, wider bandwidth, ruler-flat frequency response, and vanishingly low distortion of digital, but dynamic range on vinyl is in the high 70s, save for the intrusive clicks and pops from damage or defects in the vinyl itself. The human ear is quite tolerant of distortion and frequency response errors, so we have that going for us.

While I never bought into the lie that vinyl sounds better than digital (even moderately compressed MP3s), vinyl does sound pretty darned good. Some records sound almost as good as a CD. The trouble is, some CDs sound really bad, and some vinyl sounds really bad, making comparison hopelessly subjective. It depends on how much care went into the engineering and manufacturing.

Which brings me to today's topic: How is two channel stereo encoded into one groove of a vinyl record? Well, after some really horrible ideas involving dual tonearms (which would have been completely incompatible with mono, and would have taken up twice as much space, and would have had intolerable phasing problems at high frequencies), they finally settled on the +/-45 system.

There are several misconceptions about the +/-45 degree encoding scheme, and I intend to clear those up today. To begin with, we need to understand that the cross section of a record groove is a 'v' shape, with the walls separated by 90 degrees. In other words, the outer groove wall is +45 degrees from a vertical line bisecting the groove, and the inner wall is -45 degrees, for a total of 90 degrees.

+/-45 Degree cutting head (upside down)

The stereo encoding is sometimes described as, the outer wall is modulated with the right channel, and the inner wall is modulated by the left channel. Well, sort of. But that is more of a side-effect than a specification. Another way of stating this is that the right channel is modulated on a plane -45 degrees from vertical, and the left channel is modulated on a plane +45 degrees from vertical. It is possible to separate the two channels on playback, because vectors that are separated by 90 degrees are orthogonal -- independent.

Neumann VMS 80 Record Cutting Lathe
(If you look closely, you can see the cutting head just 
above the right side of the turntable.) (Click to embiggen.)

So it seems this orientation will modulate the groove walls independently with each channel, but what about mono? A mono cut means that both left and right channels are identical, and a +/-45 degree cutting stylus will move laterally, in the same plane as the record. Both groove walls move together in the same direction. With a real stereo program, the cutting stylus moves in all directions. If you were to look at an image traced out over time, it would look like a Brillo pad. So the groove walls are not really independent.

In fact, a better way to visualize this, is to imagine L+R being recorded laterally, and L-R being recorded vertically. If both channels are equal, L-R goes to zero, and you have a mono record. I am not aware of any record cutters that work this way, but actually, it might make the stylus motion and groove excursions easier to control. Incidentally, L+R/L-R (also called mid-side, or M/S) is totally compatible with +/-45. The vector sum of the stylus motion is identical with either matrix.

Stereo phono pickup (cartridge)

The main reason that phono pickups (cartridges) are all +/-45 designs, is that they can decode the left & right channels directly, whereas a M/S cartridge would require additional decoding. Not that it's difficult; stereo phono preamps could have had this capability since day one, if anybody had thought of it. But it would have added a few cents to the cost of each unit sold. So +/-45 it is. But any record cutter could switch to M/S at any time (even on the same record!), and still be totally compatible with all +/-45 pickups (decoders). As I say, the vector sum of the stylus motion is identical.

The fact is, there is good reason to want to record lateral and vertical components separately. Lateral recording is more resistant to distortion, and can be recorded "hotter" than vertical can. So limiters could handle each plane separately, possibly having less audible impact on the sound quality of the end product. It might also be advantageous to engineer a cutting head optimized for separate lateral and vertical excursions, vs. two identical left and right coils, which could be a compromise for vertical and lateral cuts.

Not that anybody cares about this anymore. I don't think there are many new breakthroughs in record cutter designs, these days. The payback would be marginal, and maybe what we have is good enough for the market share that vinyl has, in the big picture.