Edward Tatnall Canby - The Saturday Review

Perserving Sound in Time and Space - Phonographic Theory

Sound WavesThe recording of sound involves, whatever the method, the actual preservation of a pattern of vibrations in time - a process that challenges the imagination.

Sound itself, as we usually encounter it, is a matter of vibrations in the elastic and flexible medium, air, transmitted directly to our elastic and flexible mediums, and, thence, via some highly complex and not well-understood means, to our brains in the form of nerve messages.

To record sound we must capture the air-vibration pattern. To do so, moreover, we must substitute for the elusive dimension of time another dimension: length. The completed recording must exist simultaneously in a physical sense, throughout its time span; so time is eliminated, to be reintroduced in the playing of the record. This despite the adage that time "waits for no man;" time is irreversible, and actually cannot be reproduced at all.

The Complaint Medium

Recording processes are more numerous than you might think. Vertical and lateral grooves, on disks and cylinders, are merely a beginning. Several types of light-ray recording are of importance, notably in the sound moving-picture field. The latest recording method of all-magnetic recording - has been put to practical use in this country only since about 1948, though the principle was discovered and applied, after a fashion, near the turn of the century.

We may make an important distinction at once between mechanical recording methods, involving actual physical vibrations recorded in terms of motion itself (a phonograph needle is such a vibrating part), and those recordings which do not involve any physical vibration. All recordings that are made and played by means of at least two steps where actual motion transmits the vibratory energy that is the pattern of sound itself - cutting stylus and playing needle. On the other hand, light-ray recording by means of photographic film may be accomplished without moving parts (except for the inevitable film motion that represents the passage of time itself), the sound vibration pattern being recorded by means of light-intensity changes, (Another method, variable area recording, leaves the light intensity the same but varies the area of light transmitted through the film. Results are virtually the same). Magnetic tape, similarly, is a medium where no motion other than the forward motion of the tape - the marking off of time - takes place; the actual recording is achieved through magnetization of the coating on the tape, and the playing back responds only to this purely magnetic pattern. Such recordings are free of the complex problems involved in forcing a physical object such as a needle, in spite of its own weight and mass, its own natural vibration speed or resonance, to vibrate in conformity with an imposed pattern.

That merits some further thinking - it is a very important point. You may compare for yourself a transmission medium such as a pipe and water, involving a material with mass and weight, to a transmission of energy via the electrical medium. To move water, with all the problems we must cope of physical mass - notably the force of gravity. Water flows down; it must be forced up. We cannot rush water. Its speed is very limited. It is clumsy. But to move energy from place to place in the form of electricity is another thing altogether, for electricity is free from the impediments of physical mass. It goes uphill as easily as down; its speed is that of lightning. Even electricity has, so to speak, a mind of its own; it will not always conform exactly to imposed patterns without offering resistance, or contributing unwanted patterns of its own. But compared to a physically massive medium, it is miraculously useful to us because of this very freedom from mass, coupled with its virtually infinite capacity to transmit energy in quantity along prescribed metallic paths, and to transmit patterns of unimaginably numerous sorts (TV pictures, for example), including the patterns of sound waves.

The problem, again, in recording sound is that we must transmit a vibration pattern from one "place" to another with a minimum of tampering, of distortion. You can see that this means that the transmitting medium, what ever it may be, must impose an absolute minimum of its own "ideas" upon that pattern. It must be the Perfect Servant, entirely ready to respond with absolute faithfulness to any outside pattern we may impose upon it. These stringent conditions are met almost ideally in those elements which have no substance in the usual sense - light, electricity, magnetic force. A vibrating needle, even though it be no more than a semi-microscopic point of jewel, still has very much of a mind of its own, especially when you ask it to vibrate at perhaps 12,000-odd times a second. The supporting needle shank, the set-screws, the phonograph pickup arms, and the rest of the mass to which it must be attached, make its behavior pattern, its motion in space, all the more obstinately unpliant. Hence the extreme desirability of a recording that is entirely outside the "physical" area, from the time the original sound reaches the microphone until reproduced sound leaves a loudspeaker.

Practical limitations and conveniences, however, make this a somewhat unmanageable ideal so far as home reproduction of music is concerned; we still have the phonograph needle and its mechanical motion, and by the looks of it we will have the needle with us for some time to come.

For that matter, practical recording began long before electricity or any of the other related media could be used at all. The original process was all-mechanical, involving physical motion from beginning to end, and continued so from the beginning (1877) right through until the mid-twenties, when electrical circuits born of radio were introduced. "Electrical" recording followed. The recording itself was still, of course, mechanical by reason of the cutting needle, but the pattern was transmitted to the needle electrically. Somewhat later there was "electrical" reproduction of records in the home, the actual re-creating of sound from the record being a matter of mechanical motion, as it still is today. It was the transference of the vibration pattern from source to cutting needle and from playing needle to the loudspeaker that became electrical, and with enormous advantage, too, since in addition to its faithfulness and ease of handling, an electrical pattern can be amplified (made larger) to any degree imaginable, and yet retain its essential shape.

Recording of Time

How can one record the element of time, then reproduce it? Not a simple problem at all. Time is usually translated into motion of some sort; when that motion is repeated, the time pattern may be reactivated. A recording medium must capture time, and with extreme accuracy. It must also capture, within or upon that time pattern, the specific and complex pattern of waves or vibrations that is sound itself.

The original concept of recording, as envisioned by Edison, involved the capturing of time in a moving groove, to be traced like a railroad track by a stylus. That conception is still with us today in the disk record. To avoid spreading time over miles and miles of groove length, that groove was from the beginning reduced to a circular path, as one confines hundreds of yards of string to a ball and miles of wire to a reel. Originally the groove made a circular "spiral" path around a cylinder. Only later did the spiral wind in upon itself on the surface of a disk.

It's a good idea, at this stage, to fasten upon this conception of the recording of time itself, and separate it from the recording of sound patterns. Set up any sort of recording machine and run it without sound, and you have in effect a pure (or semi-pure) recording of silent time itself. A "blank" LP record will contain a perfect spiral groove a half mile long that, given the proper mechanical equipment, will support a following stylus-needle for a rigidly accurate slice of time, in utter silence. So also with a half mile magnetic tape, wound upon a plastic reel; and so with the Edison cylinder, the Berliner flat disk, the moving-picture film, and the rest. Time recording comes first. Time is limited by length, extended by more length-one dimension in terms of another.

Lateral and Vertical Recording

Introduce sound itself upon this rigidly controlled time recording, and we have a new and more diverse element. The grooves of all types of groove recording are basically the same. But with sound waves to be translated into actual physical motion of the stylus which cuts the groove, divergencies of method appear at once.

In one major type of recording, that which we still find on our phonograph disks, LP and otherwise, sound vibrations are converted into side-to-side vibrations of the groove-cutting stylus. Given our predetermined pattern, this side-to-side motion is "recorded" in the groove shape itself, as a side-to-side waver, snake-like. The half-mile groove is no longer perfectly even but preserves in its sinuosity the pattern of the stylus movement. In the other type of groove recording, the original method (still used in specialized radio transcriptions), the vibrations occur up and down in the stylus, vertically, rather than side-to-side laterally. The groove that is cut takes on a pattern of "hills and dales," of deeper and shallower places without disturbance of its sideward track.

The Two Silent Dimensions

Take a good look at these grooves. They are most decidedly not sound itself, nor even the actual pattern or shape of a real sound wave. Sound, traveling in air, is neither a vertical nor a lateral wave but a compression wave, a longitudinal pattern somewhat like a wave that might travel from end to end of a long coiled spring. (You have read of the "shock wave" that moves outward from the exploding atomic bomb. It is an expanding wave of compression in the elastic air, not unlike the far tinier compression waves of sound that radiate in an expanding circle from a source of sound vibration.) Compression sound waves, in the recording process, are first converted at the microphone into electrical current carrying the same pattern, then into side-to-side (or up-and-down) mechanical motions of a stylus. At this point, to be sure, compression waves of actual sound may be generated in the vicinity of a recording stylus and we can thus occasionally hear wisps of actual sound from the recording machine or the playback phonograph needle ("needle talk"). But the pattern cut into a record is not sound.

What is it? Think again. A combination of accurately measured time in terms of length, of inches of groove per second, and of accurately measured sound patterns in terms of width, of sidewise displacement of the groove from its exact spiral track. A record is totally silent. There is no motion upon it. The grooves lie still. There is no energy. There is nothing but shape. The passage of time that is essential for recording is not visible in any way, nor indeed does it exist upon a record as such. We have here a kind of static mechanical memory which, given the correct energizing in terms of time versus groove length and side-to-side deviation, will produce not sound, but a physical vibration not unlike that which occasioned the shape of the groove in the first place. Converted into electrical patterns, this information must be re-converted in a loudspeaker into compression air waves before we may rightly call it sound and hear it.

Such a common place in our lives is this record that we are not easily fired in imagination by the wonder of this arresting and preserving of time itself in purely dimensional terms. Lengthwise and side-to-side no more. But within those two dimensions any sound may be delineated, to the last overtone of time complexity.

Excerps from the Saturday Review Home Book of Recorded Music and Sound Reproduction
by Edward Tatnall Canby, 1952 - Prentice-Hall, Inc. New York