American WideScreen Museum Home Page

The second of three FANTASOUND articles provided by John Schmul is an optimistic appraisal of what can be done with directional sound reproduction in the future. It was presented at the 1942 Spring meeting of the Society of Motion Picture Engineers and this article was published in the July, 1942 issue of their Journal.


Music Department, Walt Disney Studio, Burbank, California

Summary.ľA non-technical discussion of Fantasound from the musician's point of view. The use of Fantasound is reviewed as a basis for discussing ways in which it can be used in the future.

Fantasound has been demonstrated to the public only in Walt Disney's Fantasia, but to accept or reject Fantasound on the basis of its use in that picture would be unjust. Fantasia is a remarkable showcase for an experiment in sound engineering because it uses music as a vital function of the picture. However, the dramatic effectiveness of Fantasound was limited by three conditions peculiar to this production.

(1) During its actual picture footage Fantasia uses only music on the sound-track. This eliminates the possibility of placing and moving dialog or sound-effects in the multiple speaker system that Fantasound includes. Dialog and sound-effects are the "real" sounds of the movies with which the audience is thoroughly familiar. Because of this familiarity it is quite possible that the location of these sounds in the theater could be more easily registered than the placement of musical sounds.

(2) The music that Fantasia interprets was conceived long before sound-film was available for use. The compositions were designed for concert performance and were so well designed for that medium that any orchestral changes made to improve reproduction greatly affected their basic character:

(3) The original recording of the entire orchestral performance of Fantasia had been completed before it was known what dimensional effects would be available in the theater. It was thus impossible to guess what method of recording would be most efficient for reproduction in Fantasound.

This is in no sense to be interpreted as an apology for Fantasia or the methods used in it. It is merely a description of certain obstacles that would not be confronted in the usual feature.

The future of Fantasound depends upon the efficiency with which the original sound material can be transferred to film and upon the dramatic effectiveness of the total result. These related factors dictate the future of Fantasound because they represent, respectively, the expenditure necessary and the expenditure warranted by box office returns.

Before suggesting a method of recording an orchestra that might be practicable for future productions in Fantasound it seems advisable to describe briefly the method employed in Fantasia. During the original performance, each of six sound cameras recorded the close pick-up of a particular section of the orchestra. A seventh camera recorded a blend of these six close pick-ups, and an eighth recorded a distant pick-up of the entire orchestra.

In preparing the final re-recorded track from this original material several weaknesses became apparent. Because of acoustical pick-up the separation between the six sections of the orchestra was merely relative. In the material on the woodwind channel, for instance, the woodwinds usually predominated, but material from other sections of the orchestra was definitely present. Many times, because of differences in performance level, the material from adjacent sections would be as loud as, or louder than, the woodwinds directly picked up. This lack of complete separation was not an insurmountable obstacle in creating an artistic balance for ordinary reproduction, but it greatly limited the dramatic use of orchestral colors in Fantasound. If we wished, for dramatic reasons, to have a horn call emanate from a point to the right of the screen, our purpose would be confused by hearing the same call, at a lower volume, on every other speaker in the theater. Greater separation in the original recording could have been achieved only by greater segregation of the sections or by moving the microphones closer to the individual instruments. To go any further than we had gone toward segregation of sections or close pick-up would have impaired quality of performance in one case and recorded tone quality in the other. On the point of efficiency of the Fantasia recordings we must observe that only one-third of the material recorded on chosen performances was used in the final dubbing. The unused film contained sound that was too repetitious of, other channels, too poor in quality, or, during long sections, too unimportant in the design of the composition to help the total result.

Since the completion of Fantasia we have recorded orchestral performances of five compositions for possible use in Fantasound. It is not likely that these can appear as productions for a long while, but the method that was used may provide a possible approach to future Fantasound projects. The recordings were much less expensive and, there is every reason to believe, can be much more effective dramatically than the Fantasia recordings. We concentrated upon the achievement of two qualities of Fantasound that seem to us to be important-the illusion of "size," possible to attain by proper use of a multiple-speaker system, and recognizable placement of orchestral colors important to the dramatic presentation of the picture.

For the illusion of "size" or "spread," we used a three-channel recording set-up. Channel A was fed by a directional microphone far enough from the instrumentalists to cover the entire left half of the orchestra. Channel B recorded the right half of the orchestra. Channel C recorded a distant pick-up of the entire orchestra. This three-channel system recorded the "basic" tracks of the composition. It is important to note that in planning the material for these "basic" tracks any orchestral color or passage for which we might have special dramatic use was omitted from the performance. The recording of this special material will be described later.

In reproduction over the Fantasound system this method of recording the basic tracks has great flexibility. To regain the natural spread of the orchestra, the A channel (left half of the orchestra) appears on the left stage speaker, the B channel (right half of the orchestra) appears on the right stage speaker, and the C channel (distant pick-up) appears-on the center speaker. The distant pickup appearing in the center adds an illusion of depth which is beneficial and also provides a more practical "cushion" for the solo instruments or other special material that would normally appear in the center. The "panpot" (described by Garity and Hawkins in the August, 1941, JOURNAL) can execute practically any variation of this reproduction plan that could be demanded. Each track can appear on any one stage speaker, any two stage speakers in whatever balance desired, or on all three stage speakers in any balance. The house speakers can be added to the left and right stage speakers in whatever set balances desired, or they can replace the left and right stage speakers so that-sound comes only from left and right house and center stage (as in "Ave Maria" in Fantasia).

In the recording of what I have termed special material-material whose location it is important to register-we employed the only method that assures absolute separation. The section of the basic track with which the special material is to synchronize is used as a playback on earphones available to conductor and instrumentalists. The physical difficulties of this method can be minimized by careful planning of the orchestration. It is usually possible to avoid the occurrence of the same melodic passage or rhythmic pattern in both the special and basic material. This makes synchronization less critical and also allows more freedom in performance of the special material. As advantages, the playback method offers complete control of the volume relationship between special and basic material; complete freedom in locating or moving the special material; and freedom to choose the pick-up, in recording the special material, that produces the finest quality in reproduction.

As an example of the use of the playback method, in The Swan of Tuanela, by Sibelius, there is an English horn solo that is vitally important in the design of the composition. We knew that this English horn should be a principal actor in dramatizing the score. We had recorded the composition played by the complete string orchestra omitting, among other instruments, the English horn. We then recorded the English horn alone, using the performance by the strings for the playback. A relatively distant pick-up was used, which gave the tone of the English horn brilliance, but also lent a feeling of mystery in character with the subject. Because of the complete separation achieved it is possible to submerge the solo in the rest of the orchestra or to make the solo stand out in a clear relief physically impossible to attain in concert performance. The solo can locate as its source one of the three stage speakers or, by balancing its volume between two speakers, can seem to locate a definite point between them. The solo can come from the left or right unit of house speakers without the stage speakers or, if power or diffusion are desired, can come from every speaker in the theater. The solo can move in such a way that it seems to follow the pattern of a pictorial effect; it can change from offstage to onstage; or it can change its source, by a smooth, irregular movement of the panpot dial, so that it seems to float through the theater. I have mentioned a single composition and only a few of the effects possible. However, it is clear that the restrictions offered by this tentative method are infinitely less than those offered by the method used for Fantasia. (The Fantasia score contained only one example of complete separation-the solo voice and chorus of "Ave Maria" were recorded by the playback method to an orchestral accompaniment recorded a year and a half before. The vocal performance of "Ave Maria" was the last material to be recorded for Fantasia, and we were able to use everything Fantasound had to offer. It is interesting to note that for many of those in the audiences-at least in New York and Los Angeles-Fantasound was "turned on" only for "Ave Maria.")

The advantages of volume range are probably more obvious than the advantages of other features of Fantasound. To be able to use the upper volume range without distortion and the lower range without submerging the tone in ground-noise has been the dream of every dramatically minded sound-director since the advent of sound reproduction. Experience shows us, however, that this greatly extended volume range still has important natural limits. If sound is reproduced so low that it is unintelligible or so high that it causes physical discomfort, there must be adequate dramatic reason. Either extreme is likely to irritate.

Dialog and sound-effects, as material for use in Fantasound, have one decided advantage over music. They do not have to be recorded differently from the customary recording of ordinary sound. Their placement, movement, and extended volume range are all accomplished after they are normally put on the film.

Dialog is the only sound medium in whose reception the audience has been well rehearsed. The average member of the audience has heard the sounds that the screen sound-effects imitate, but he does not ordinarily analyze their character or location with any great care. He has listened to music but, perhaps wisely, he does not bother himself with the details of its complex pattern. In the reception of speech, however, he has trained himself to register, in great detail, character, pitch, volume, and location. Location of sound source is an unconscious function of his daily group conversation, group work, and group play. It is reasonable to expect, then, that when dialog placement has dramatic meaning it will be efficiently received by the audience-at least, more efficiently received than the placement of sound-effects or music. Because of the visual limitations of the screen, dialog, in Fantasound as in ordinary reproduction, comes normally from the center of the stage. For this purpose the center stage speaker is adequate. Because the ear is critical of voice placement, however, it is not far-fetched to attempt the location of characters by changing the speaker source. If an actor appears in the area at the extreme left of the projected frame, or if the implied location is slightly to the left of the projected frame, placement of the voice on the left stage speaker supports the illusion. Such use of the three stage speakers creates the possibility of dialog between extreme left and extreme right or between center and either side without greater sacrifice of intelligibility than would exist in dramatic productions on the stage. Obviously the device could be over-used to the point of annoyance, and should be limited to dramatic situations that are definitely improved by the illusion. In the treatment of off-stage voices the house speakers could be used to advantage. When a voice, or a group of voices, comes from the left or right unit of house speakers, an effect of reverberation is added to the original recording. The loss in intelligibility and in point source definition could have dramatic value because they imitate these same losses in the reception of real sounds from a distance.

Fantasound is able to make its greatest contribution in combining dialog, music, and sound-effects. In ordinary reproduction one of these three mediums must, with rare exceptions, be dominant while the other two are sacrificed. In Fantasound it is possible to follow the continuity of the dialog clearly and still receive the full emotional impact of the music, or the dramatic realism of atmospheric sound-effects. As a possible use in the theater, consider that the center stage speaker would be saved exclusively for on-stage sound-dialog, music performed on the screen, or realistic sound-effects. The house speakers and, at a lower level, the side stage speakers would project music or general sound-effects at a level natural for them. As long as the music or effects are pertinent to the story being portrayed they will not distract and would not cause the dialog to become unintelligible. This physical separation of sound-tracks also reduces to a minimum the unpleasant phenomenon produced when a well-modulated track is "pinched."

If these comments seem to wander it may be because Fantasound is at the wandering stage of its development. We have the tools and we have not decided what we intend to build with them. These tools may not be available in the theater "for the duration," but this might be an excellent period during which to develop a practicable, effective plan for using them. It is within the power of Fantasound, as an idea, to revitalize the industry. This power, however, can not be fully developed until script, direction, music, and recording are planned with Fantasound as an organic function.

©1942 Journal of the Society of Motion Picture Engineers.