Chapter XIII

The Recording of Sound

Word count: 4100

Cost Levels of Sound Reproduction

∣📄 p.177 The forward drive of technology offers the prospect of a number of new possibilities in the transmission of texts. Television may render long-distance reading a possibility. The transmission of facsimile texts by radio is on the verge of practical realization. The direct recording of sound is already in use for research purposes, and is definitely superior to written reproduction in certain fields of linguistics and music.

In general, the technology and cost relationships encountered in the reproduction of sound follow the pattern of those encountered in the reproduction of texts.

In disc recording of sound, for instance, the disc area is most economically used if it is made to rotate at constant linear velocity. The ordinary phonograph runs at a constant velocity in revolutions per minute. This means that the grooves in the short circumference near the inside of the disc are moving at a speed of one and one-half feet a second, which is adequate for the reproduction of sound frequencies up to 5,000. But the same number of revolutions per minute (seventy-eight is the international standard) would then carry the outside grooves of a large record at two or three times this speed. It takes two or three inches of groove at the outer edge of the record to reproduce the same amount of sound that is reproduced nearer the center on one inch. By varying the speed of the record in revolutions per minute, the velocity of the groove can be made constant. ∥ This economizes record space but increases apparatus cost. The relationship is precisely the one that is encountered in microphotography, where an increase in cost of apparatus yields an economy in cost of film.

The relationship of first cost to running cost is encountered in sound reproduction as in text reproduction. A set of master discs costs about $7.50 for one side, $15.00 for two sides of a record. The cost of making pressings from these is from $.55 to $.75 each, of fifty; or more are needed. But an inferior record can be made, one at a time, by “dubbing”—that is, playing the master record and recording it on a new disc. The dubbed records cost about $1.00 each. The cost structure here is like the comparison of photo-offset and photostat, and curiously enough, the two cost lines cross each other at about the same place—twenty-five copies.

As might be expected, the cost per word of sound recording is higher than the cost per word of text reproduction, but the spread is not so great as might be imagined, Fifty 12″ discs would record 100,000 words. The filing space necessary to hold them would be a little more than a cubic foot. These records, in an edition of fifty, would cost from $2500 to $3000. A very expensive printing job, with copious illustrations, might well run up a cost very close to this.

The figures upon which these conclusions are based were furnished by Professor Miles L. Hanley, who has written the following description of the technique of sound reproduction.

Technique of Phonographic Recording

By Miles L. Hanley

Phonographic recording is rapidly becoming an indispensable tool for research. Conventional transcriptions cannot adequately represent the unusual intervals of folk music and many of the important char∥ acteristics of speech, such as voice quality, intonation, distribution of stress, speed and rhythm of utterance. Written records, however good, must contain a personal and subjective element. Without some ∣📄 p.178 objective check, their use requires an act of faith—faith that the field worker has heard accurately, that he has been able to translate what he heard into symbols, and that we know just what he means by his symbols. Phonograph records not only provide such a check, but they constitute a perfect archive of material which may be consulted and studied in the future.

After the invention of the phonograph in 1877, its obvious advantages to the linguist, the folklorist, the ballad collector and the anthropologist were promptly recognized by European scholars. The Vienna Phonogramm-Archiv was founded by Siegmund Exner in 1899, and the Musée Phonétique by Auzulay in 1900. Since then many other collections have been established: a second one in Paris, two in Berlin, and others in Cologne, Lubeck, Frankfurt, Leyden, Oslo, and Zurich. In America, although our scientists have made great advances in the mechanical technique of recording, our scholars have been slow to see the possible uses of this tool. It is true that some folk song and American Indian material has been collected in the past, but mostly by the unsatisfactory methods of cylinder recording or magnetic recording. So far we have nothing in America that could be called even the beginning of a central Phonogramm-Archiv, but we do have some important collections and many more are in prospect. A few of these are: Professor ∥ Greet’s American Speech records which have been made at Columbia during the past ten years, the folk songs collected by John A. Lomax, the Gullah negro records made by Dr. Lorenzo D. Turner, and the records of New England speech made for the Linguistic Atlas of America (United States and Canada).¹

Sound Recording Methods²

In all kinds of electrical recording, sound waves (periodic variations in pressure) produce mechanical movements—usually in a sensitive diaphragm. These movements are translated into corresponding electrical vibrations or variations which (usually after amplification by vacuum tubes) produce mechanical movements in a second diaphragm or a moving armature of some type. Electrical variations may be translated into light variations (sound on film) or into magnetic variations (Telegraphone, Blattnerphone). In the familiar phonographic technique, the sound vibrations are translated into mechanical movements of a stylus which cuts or imbeds a groove in a disc, cylinder, or strip of motion picture film.³ The groove may be of constant depth, recording lateral swings of the stylus (lateral cut, “snake track,” Berlinerschrift) or of varying depth (hill and dale, Edisonschrift). Hill and dale recording is used in dictating machines and in some very high-quality studio recording, ∣📄 p.179 but lateral cut records are most practical for common use at the present time.

The recording methods referred to in the preceding paragraph may be classified as (1) Photographic, (2) Magnetic, (3) Mechanical. Since the particular kind of mechanical recording known as instantaneous recording on disc will be used in most research projects, the first and second will be dismissed with a few brief remarks.⁴

1. Photographic. In this method, electrical variations produce corresponding light variations, which in turn produce either a varying amount of exposure (variable density, Western Electric) or varying area (variable width, R.C.A.) on a photo- graphic film. Equipment for this method can be easily portable and can be used under extreme climatic conditions, but the record cannot be played back at once. According to Dr. H. A. Frederick, duplicates, even commercial ones, show considerable variations from print to print. Such duplicate records have thus far failed to show as great uniformity as disc record pressings. The individual film records are expensive, delicate, bulky, and difficult
to handle compared with disks.⁵ In contradiction to this statement it has been pointed out that duplicate prints produced by the large film companies are now fairly constant and that variations are mainly in the range above four thousand cycles. The fact remains that it would be difficult for the ordinary research worker to get copies ∥ of any such quality. The excellence of film recording is unquestioned, but the expense of this method is prohibitive for most research projects.⁶

2. Magnetic. In this method, electrical variations produce variations in the degree to which a magnetic wire (Telegraphone) or steel tape (Blattnerphone) is magnetized across its axis. This method requires special equipment and is not much more satisfactory than cylinder recording. Duplicates cannot be made, though of course it is possible, with special equipment, to dub the record on disc or film. The great disadvantage of the magnetic method is that, with the alloys now available, a wire or tape soft enough to record sensitively loses its magnetism in a short time.⁷

3. Mechanical. Mechanical methods include cylinder recording, scratching on motion picture film, studio recording on wax, and the various instantaneous disc methods.

Cylinder recording may safely be excluded. It is almost as expensive as disc recording, and much less satisfactory. The records are fragile, insensitive to high frequencies, difficult and expensive to duplicate, and quickly worn out by use.⁸

Recording by means of grooves scratched on motion picture film has not, so far as I know, been tried very extensively in this country. The material is not expensive, since discarded picture films may be used, but duplication is difficult ∣📄 p.180 and special apparatus is needed both for recording and playing. I have heard that there are difficulties in keeping the films flat and constant in velocity. The recording medium, celluloid or cellulose acetate, is also open to objections which will be discussed under disc recording.⁹

Since most of us require records that can be made in the field and paid for from limited funds, studio recording on wax is usually out of the question. But if the material can be recorded in the studio and expense is) no object, records can (be made of a perfection that would have seemed miraculous five years ago. In studio re- cording, the groove is cut in a soft “wax” (really a mixture of metallic soaps). “The surface of the wax, after being thus engraved, is rendered electrically conducting. This can be done in a number of ways, such as dusting with graphite or bronze powder, sputtering with gold or silver in a cathode chamber, or by the chemical precipitation of silver.”¹⁰ The wax is then electroplated with copper and a “master” made. From this “master” or negative, records may be pressed, though in commercial practice, where many pressings are to be made, a “mother” is electroplated from the “master” and from that, one or more “stampers,” that the master negative may not be worn out. The layman who does not understand the delicacy and difficulty of recording on a studio wax should see the process, or read a detailed account (e.g., that in Brown’s Talking Pictures, cited above).

What might be called “semi-studio” recording has been tried by the Lautabteilung of the National Library of Prussia and by the Royal Irish Academy.¹¹ Portable apparatus for recording on studio wax exists, but it is much more cumbersome than apparatus for instantaneous recording and is open to many serious practical difficulties.

The remainder of this. discussion will be devoted to instantaneous recording on disc. There are three kinds of methods: ∥ (1) those using a pregrooved record; (2) cutting; (3) imbedding.

(1) Pregrooved records have been widely used in home recording. There are two methods: In one the stylus point drops into the center of the pregroove, and produces a groove within a groove; in the other the stylus point fills the pregroove and produces undulations in it. The discs are expensive, but when a small amount of recording is to be done and high quality is not required, this method offers great advantages, since the ordinary pick-up arm is used and no feed for the cutting head is required. Very good records may be produced with this method by using a special turntable and powerful motor, but in this case the added cost of discs and special equipment is such that one might as well have provided a turntable feed in the first place.

(2) Instantaneous recording by cutting is similar to wax recording in that an actual chip is removed from the discs. Compositions of gelatin, celluloid, or cellulose acetate are generally used. Gelatin is sensitive to moisture and deteriorates with age. Celluloid is inflammable and cellulose acetate less satisfactory in quality than celluloid. “Metallophone” discs, made in Germany, produce excellent results, but the price is very high, especially for the twelve-inch discs. A new principle is employed in the Draloton disc, also made in Germany. The recording is done on a soft “bakelitartiges” material, which is afterward hardened by baking. The baking process, which does not injure the original record, makes it playable with a steel needle. But the Draloton disc requires considerable skill on the part of the operator, must be kept fairly cold, and deteriorates rapidly with age. For these reasons it is not a suitable material for field recording. The Presto Recording Company of New York markets a disc somewhat similar to the “Metallophone,” which they ∣📄 p.[180a]

∣📄p.[180b] (blank) ∣📄 p.181 claim to be even better. The records, however, are just about as expensive. A steel needle of a special type is used in reproduction. Under ideal conditions, the special discs referred to give considerable better results than aluminum, but since most of them are sensitive to heat or atmospheric conditions, or deteriorate rapidly with age, they are, on the whole, less suitable for field work than aluminum. The Marguerite Company of London uses a similar disc which has a smoother surface than the Presto, but is otherwise open to the same objections and requires a special trailing needle. An excellent disc of similar type has recently been perfected by the Sound Specialties Company of Waterbury, Connecticut. The disc is coated with a kind of cellulose acetate lacquer which hardens slowly with time. It may be played with any ordinary steel needle.

An important objection to the cutting method is that no provision is made for removing the long spiral chip, which tends to curl around the stylus. In studio recording the chip is withdrawn by a suction pipe as fast as it is formed. Both the styluses and cutting media available at present tend to be erratic. The same objections apply to records scratched on either nitrate or safety film.

(3) In the imbedding method, the stylus is fed across the disc at a uniform rate and presses a groove into the disc, which has been treated with some lubricant to reduce surface noise. Machine oil or axle grease is satisfactory, but the cleanest and most permanent lubricants are waxes, e.g., paraffin. The ordinary aluminum disc has a hard surface and requires a heavy weight on the recording head. It is obvious that additional sensitivity could be secured by the use of a softer medium and a lighter weight cutting head. For this purpose, discs softened by annealing have often been tried, with varying results. If annealed too rapidly, discs tend to curve and to record unevenly. Discs annealed slowly in oil, however, keep their ∥ original stiffness and produce better records. Apparently they will stand as many replayings with fiber needles as the harder discs.¹² A special sapphire play-back needle may be used on aluminum discs and gives excellent quality without injuring the records. Such a play-back needle, how- ever, must be made to fit exactly the groove in which it is to be used. Consequently the records made by different machines will require different play-back needles.

Field Recording. For field recording, prime requisites are portability, ruggedness, economy and simplicity of operation. Other requirements will vary with individual projects. Almost all methods known to me have been mentioned, but in my opinion, the most suitable for most projects is disc recording on annealed aluminum.¹³ It allows for satisfactory quality with simple operation, relatively low cost of equipment and materials, ease of storage, simple and accurate duplication. Single copies may be re-recorded on aluminum at slight cost, but if more and better copies are required, a metal master from which standard type records may be pressed may now be made by direct galvanic processes at a cost of under ten dollars for the making of a twelve-inch master and several pressings.¹⁴ Portable outfits ranging in cost from $225.00 to $1,100.00 are now on the market. The finest portable equipment I have seen is that provided by Harvard University for Professor Milman Parry’s expedition to Yugo-Slavia.

Power Supply. When dependable alternating current is available, it is, of course, the most convenient source of power. Otherwise batteries must be used. Some weight may be saved by the use of a powerful spring motor for operating the turntables, but in any case batteries are needed for the amplifiers. When working in New England I used a combination set, with which I could record on D.C. or A.C. ∣📄 p.182

Microphones. The familiar one-button carbon microphone is inadequate for good recording, since it performs accurately only within a very small frequency range; but the best quality two-button carbon microphone, for example, the Western Electric 600A, gives a fairly uniform response. A good carbon microphone is satisfactory when the speaker stands fairly close to it so that small amplification is required; otherwise the unavoidable carbon hiss will produce a swishing noise in the record. The dynamic, crystal, ribbon, and condenser types require more amplification than the carbon microphone, but have no extraneous noise. The condenser is excellent in quality, but since it is sensitive to moisture and requires particular care, it is unsatisfactory for field recording. I have had no field experience with the new crystal microphone, but have heard very good reports of it. The choice between the dynamic and ribbon or velocity type depends on the nature of recording to be done. The ribbon or velocity microphone is highly directional, that is, picks up the sound coming from directly in front of it or directly behind it. This feature is often an advantage, since background noise is largely eliminated. The dynamic and crystal microphones are not so directional and are suitable for several speakers. Until recently the expense of good dynamic microphones has been prohibitive, but excellent ones may now be obtained at a cost of under $100.00. For most purposes the dynamic microphone is to be recommended.¹⁵

For aluminum recording it is best to use a pre-amplifier, for it is then possible to strengthen the frequency response at the point where aluminum is poor, namely ∥ above 3500, with a consequent gain in clarity.

The present discussion has been largely concerned with recording in the field, but there are many other practical uses of recording equipment. In laboratory experiments dealing with sound, it is possible to record simultaneously the sounds studied.¹⁶ It is very easy to make excellent records of radio programs either as received or as sent out from the studio. Many important broadcasts are now so preserved. There are many obvious educational applications, among them teaching the pronunciation of foreign languages, testing musical aptitudes, correcting speech defects, adding sound effects to silent films, and, thanks to the process of hearing by bone conduction, in teaching the deaf.

Instantaneous recording with the latest type of discs opens up many hitherto unrecognized possibilities. Recording done with two turntables may be perfectly continuous, so that material in any quantity is recorded as it is produced. With a good transcribing instrument, such as is now available, one may play such records, localizing exactly the material of interest to him, and re-record on a new disc. Rerecording in the past has commonly meant serious loss in quality, but with the new coated discs, excerpts are almost indistinguishable from the original. This fact is of great advantage in dealing with folklore, ballad material, or anything produced by a naive speaker, who cannot be expected to give a rehearsed and timed performance. The material may be edited and even rearranged.

Notes