CIRCULAR PHOTO-ELECTRIC SOUND TRACK JOINT CLOSURE

Photo-optical discs containing concentric variable area sound tracks are the result of modulated light exposing rotating film. The initial point of exposure represents the start or “head” of a given track; after one full revolution, the trace of exposure will again meet its own “head” and be terminated, thus forming the end or “tail” of a given track. The point at which the track’s head and tail meet is the “joint”, or joint area.

In attempting to transcribe continuous, sustained tones (either pure or complex in waveform nature), irregularities have always occurred at the joint. Such irregularities will manifest themselves as random waveforms, and will, in turn, effect reproduction so as to produce audible noises at the joint area.

Initially, attempts were made to produce discs with joints that were made by simply allowing the two ends of the track to meet (the result of carefully aligned instantaneous “start” and “stop” points). A “butt splice”, so to speak.

The result was joints that produced loud audible “clicks” or “pops”; something quite similar to a broken phonograph record’s skipping needle. Yet, silent joints could be produced via quiescent (unmodulated) tracks. It was therefore obvious that the problem was within the modulation itself, rather than being a result of the “mechanics” of closing the track across the joint area.

The problem posed by the actual modulation itself is the result of a universal characteristic of unstable complex waveforms: they are constantly in a state of flux. As pure as they may sound to the ear, they are, in reality, constantly shifting and changing throughout an infinite array of possible wave shape combinations. Their final form is the resonate sum of a complexity of constantly shifting overtones, a fluctuating fundamental, and inherent ambient noises. Electronically stabilizing” all these effects would be a classic case of “Winning The Battle & Losing The War”, for it is these same instabilities that make music sound like music.

The “butt splice” method was, in effect, an attempt to simply “start” with some portion of whatever wave shape was available at that instant in time, continue for one full circle on the film, then abruptly terminate with the “other side” of a wave that exactly matched the first! The “odds” of this happening rank right along with the probability of not only locating two identical snowflakes, but finding them sitting perfectly aligned, one atop the other!

Of course, considerations were made toward this end. A device was planned which would utilize spinning magnetic heads to scan the audio information piece by piece while still on magnetic tape, in the hope of finding two matching “pairs” that would coincide on the disc. After due consideration, the idea was abandoned. Tests had shown that (1.) even if such “pairs” could be found, they could only be aligned on the disc by real-time alterations in velocity, which would destroy tuning; and (2.) extensive examination of actual sound track waveforms presented no evidence that such “pairs” actually exist at all.

Thus, the “butt splice” system was abandoned. It had proved unworkable due to, for the most part, the same reasons for which it is considered undesirable in the practice of splicing magnetic tape. Magnetic tape is best served by a “diagonal” splice — one that gradually crosses the playback head’s gap so as to avoid instantaneous transient peaks, if not artificially created waveforms.

The analogy is not altogether correct, in that magnetic tape is (by nature) an electromagnetic media, and therefore algebraically (rather than geometrically) additive. Optical sound tracks are not algebraically additive. Therefore, a pure diagonal cross section of a sound track, regardless of angle (duration), would only combine with an adjacent opposing diagonal to form a totally non-symetrical bilateral structure across the splice. Such lack of symmetry would result in extreme distortion throughout the duration of the joint area.

Of course, a number of other methods were attempted. Velocity “rate-changing” machines were tried; discs were split and re-joined with mirror images of themselves; attempts were even made at graphically “drawing” good joints. These “artists renditions” were particularly humorous; they may have looked OK, but they sounded awful! The folly of this technique is understandable -aesthetically, the tracks simply look like a little artistic “touch-up” would work. In reality, however, the task of “drawing” a playable sound track by hand ranks in difficulty with the task of “carving” a playable phonograph record by hand!

The most successful method by far (and the method currently in use at this writing) is that of forming the joint area across two sides of a bilateral track with a bilateral diagonal junction.

The joint itself is an overlay (or, photographically, a “double exposure”). It is created by a gradual increase in both bias voltage and modulation amplitude as applied to the optical recorder’s sound head (specifically, the light valve assembly). Both bias and amplitude are increased by electronic logic circuitry at carefully controlled rates. They must be by complimentary, since the value of full (100%) modulation is geometrically a function of bias; amplitude in excess of 100% modulation will result in extreme harmonic distortion, since the light valve’s mechanical ribbons will physically meet. Such “ribbon clash” is also highly injurious to the light valve.

As this information is applied to the valve, the ribbons start to open (from their “closed” position; a result of applied bias in the quiescent mode) to begin a modulated “taper”. The moving film will record the taper (starting symetrically from the track’s theoretical center line) until, a few milliseconds later, the track will have reached its full width (as full bias and amplitude are achieved). Normal recording will then continue around the circular track.

As the “head” of the track is again approached (after nearly one full revolution of the film), the logic circuits will be triggered to close the joint. This is accomplished by a reversal of the “open” procedure.

The “open” and “close” mode tapers will be overlapped to such a degree as to nearly align with each other at their maximum width points. Therefore, the bulk of the actual “tapers” will be lost in the double-exposure through cancellation by the rest of the track.

However, it must be remembered that these photographic tapered tracings are being modulated. As previously discussed, the specific wave shapes resulting from this modulation will be different for the beginning of the track than the end. Therefore, when superimposed against each other, the wave shape traces will not match. Instead, they will combine geometrically to form a synthetic, unnatural shape. It is this wave shape which reproduces as “joint noise” on existing photo-optical discs.

Three critical factors should be noted regarding “noise” created by joint areas (or photo-optical splices) produced in this manner:

1. It is not actually “noise” at all, but rather a perfectly accurate reproduction of a specific wave shape;

2. Since this specific wave shape is the net result of the overlay of two random signals, and the overlay itself is random in regard to the phase relationship between the two, it will never happen exactly the same way twice;

3. The “sound” of the joint area is only significant in context; that is, the effect of the joint must be considered in relation to the information that surrounds it.

As a result of these factors, we may satisfactorily explain the seemingly unpredictable nature of the joint area. Joints often make low pitched “thumping” sounds. This is because the overlay has occurred in such a way that the combined wave forms obliterated each other, leaving little other than a great “swell” in the track which will reproduce as a single massive low oscillation.

At other times, the joint produces a dull “thud”. Usually the result of obliteration of fine structure (high frequencies) — a good example of “in context” examination. The “sound” heard is not really a “sound” at all, but rather the lack of sound! As the joint area passes, the high frequencies momentarily drop out, leaving a gap which the ear interprets (in context) as an objectionable noise.

Other joint areas produce high pitched “chirps”. These are generally the result of mid-range overlays which fall out of phase with each other. Their combined effect will be an artificial harmonic pitched one octave above either if examined independently. A short burst of this harmonic (as the joint passes) will sound like a brief “chirping” or “pinging” noise.

Many other effects are possible, as are a number of combinations of the above mentioned ones. This is why existing discs present such a variety of joint area “noises”.

Then, there are occasional joints that make no “noise” at all. These are sometimes (but rarely) the result of all the variables and random factors occurring in such a way that a wave shape is produced which is neither uncomplimentary to its surrounding structures nor badly deformed within itself. In other words — pure luck!

But quite often, joints will not produce objectionable sounds simply because they occur within tracks that, by nature of the modulation itself, share the joints own characteristics. In other words, the joint isn’t any less “noisy” — it’s simply that everything around it is more noisy! It’s the “examination in context” business again.

Simple verification of this can be found within the 40 (at this writing) existing Optigan® Music Program Discs. Of those 40, thirty-six contain “rhythmic” tracks (musical sound tracks where various instruments are playing in rhythmic movement, as opposed to a “keyboard” track in which a single tone is sustained throughout to form an “endless” loop). Rhythmic tracks are produced by exactly the same photo-optical transference as are sustained tracks. But — they have no uniform content. One may be a strumming banjo; another, a beating drum. But none are comprised entirely of constant, sustained tones. Therefore, they have no constant reference against which the ear can make an “in context” examination of the joint area.

The result is obvious. Of the 36 discs, each of which has 20 such “rhythmic” channels (720 total tracks), no objectionable joint noises occur!

But what of those occasional tracks which are sustained tones, and yet still lack any audible irregularity across the joint?

Again, they may be the result of simple blind luck. But occasionally, an entire disc will show consistantly “quiet” joints. Obviously not luck. Instead, it is, again, a function of the modulation (keep in mind — an “entire disc” is made up of a group of tones all from the same source). The harmonic nature of the modulation itself is the common denominator. In this case, the modulation itself acually “masks” the joint!

This phenomenon is most common in exceptionally complex waveforms containing an abundance of unstable harmonic overtones. The result will be “natural” distortion — distortion which is readily accepted by the human ear in that it is native to the modulation itself. A good example is the combined sound of a group of vocalists. Consider the variables involved when a group of people all sing a sustained tone (“ah”, for instance) all at once:

1. They are not in musical unison at all. The men may be two octaves below the women.

2. The total sound is no more “pure” than the sum of its components. And in this case, its component elements are the individual vibrations of everyone’s larynxes. Hardly a stable source, by technical standards! Not to mention staggered interuptions for breathing, swallowing, etc.

3. Even individually, the human voice produces a tone which is “rich” in natural distortion. In a group, the distortion is cumulative. In an extreme example, such as a person with a sore throat or partial laryngitis, the “hoarse”, raspy effect will be quite pronounced.

Therefore, an optical sound track modulated by a group of vocalists will stand a good chance of displaying a “quiet” joint — not because the joint is any more perfect than any other joint, but simply because it fits better into its surroundings!

Thus far, this paper has dealt with the specifics of the photo-optical disc’s joint area as effected by related acoustic elements. However, the complexities of photo-optical junctions are also subject to photographic criteria. The utilization of an “overlay” technique, though desirable from an acoustic standpoint, forces consideration of the logarithmic shift in resulting density caused by increased exposure. Ignoring this aspect would result in track deterioration (distortion) either by latent image loss (a factor to be dealt with in the recording of any sound track due to the Ribbon Velocity Effect, but even more so across a joint area), or by Photographic Contamination by Proximity (“Silver Growth”).

As you can see, there’s a little more to producing “quiet” joints than meets the eye.

Over the years, I have listened to much commentary and heard many suggestions on how to solve the problem. The Number One misconception, by far, has been the erroneous assumption drawn by many people that the joint is an entity unto itself. Over-simplification of the problem might lead one to believe that the joint is a “thing”, and that it makes noise, and that all we have to do is isolate it and fix it and it won’t make noise any more. Nothing could be further from the truth.

Circular opto-electric photographic sound tracks currently represent a rather new media. Scientifically, they are still in their infancy. Information on them is scarce, and even what is available is ofttimes untried if not completely erroneous. No currently proposed systems promise an amalgamous joint closure.

However, I remain confident that it can be done. I do not claim to have all the answers. But I will lay claim to a certain degree of expertise in the field of circular sound tracks. To date, I have produced over 22,000 of them.

It is with this in mind that I present this paper, on the basic assumption that a solution can only come out of a thorough comprehension of the problem.

Mike LeDoux

2/10/76