Magazine article American Cinematographer

Disney's Fantasia: Yesterday and Today

Magazine article American Cinematographer

Disney's Fantasia: Yesterday and Today

Article excerpt

"The beauty and inspiration of music must not be restricted to a privileged few but made available to every man, woman and child. That is why great music associated with motion pictures is so important, because motion pictures reach millions all over our country and all over the world. Their influence is immensely powerful and deep. We cannot measure how greatly music and motion pictures contribute toward a higher standard and enjoyment of living, increasing the well-being of each one of us, as well as our nation, by giving us not only recreation and pleasure, but stimulation and nourishment of the mind and spirit."

These words were written in 1940 by the most famous symphony conductor of that time, Dr. Leopold Stokowski, as an introduction to the program book for the Walt Disney production, Fantasia. Although this was not the first movie to present serious music visually, it is by far the biggest and became (eventually) the most successful.

Fantasia was also the first feature to be presented with a multiple track-multiple speaker sound system, thanks to Disney, Stokowski, and a recording crew headed by William E. Garity. It was called Fantasound.

That Fantasound was highly effective is illustrated by an excerpt from Time magazine's cover story of November 18,1940, which followed the Broadway premiere:

"The music comes not simply from the screen, but from everywhere; it is as if a hearer were in the midst of the music. As the music sweeps to a climax, it froths over the proscenium arch, boils into the rear of the theatre, all but prances up and down the aisles... Musicians and sound engineers who came to hear soundman Garity's gadgets perform found that such recording had never before been even approached."

Now, after a half-century, an almost exact replica of the original version is back, Fantasound and all. Thanks to modern technology plus a lot of ingenuity and hard work, Fantasia looks and sounds magnificent.

The idea of Stokowski and Walt Disney teaming up as two geniuses with but a single thought seemed to some like putting ice cream on chateaubriand. Disney's greatness lay in visuals, but his musical naivete was often evident. For years he had been combining art and music (including some light classics) in his Silly Symphony series, but he admitted to having a "tin ear" and a strong dislike of symphonic music. Although a legend persists that the 1930 ancestor of the Silly Symphonies, the Skeleton Dance, is Saint-Saens' "Dance Macabre" set to pictures, it actually has a score by cartoon specialist Carl W. Stalling and its only bow to the concert hall is a brief xylophone quotation from Grieg's "Dance of the Dwarfs." After Fantasia wrapped, Disney said "... I can listen to it now. It seems to mean a little more to me." However, his later efforts at putting music to pictures, such as Saludos Amigos and Make Mine Music, utilized mostly popular music.

In the late 1930s the great orchestras of the world were in the hands of the most colorful and lionized conductors in musical history. Their names echo even now: Toscanini, Koussevitsky, Monteux, Ormandy, Reiner. But highbrow music was epitomized for the public by Stokowski, the tall man with the elegant profile, expressive hands and mane of fluffy white hair who conducted without baton or score. He had been musical director of the great Philadelphia Orchestra since 1912 and a top recording artist since 1917. He had co-starred with Jack Benny in The Big Broadcast of 1937 and with Deanna Durbin in 200 Men and a Girl. He was Greta Garbo's frequent companion (top that, Toscanini!). Whenever a conductor was depicted in a cartoon, it was most often a caricature of Stokowski, such as Bugs Bunny's unforgettable 1949 impersonation of the maestro in Chuck Jones' Long Haired Hare, with an awed Hollywood Bowl crowd gasping, "Leopold! Leopold!"

According to Stokowski, he and Disney met by chance in a Hollywood restaurant. "I was alone having dinner at a table near him, and he called across to me, 'Why don't we sit together?' Then he began to tell me how he was interested in Dukas' The Sorcerer's Apprentice as a possible short, and did I like the music. I said I liked it very much and would be happy to cooperate with him."

Stokowski had long before worked at improving the live orchestral sound in concerts by regrouping the players from the traditional seating arrangement-a practice condemned by some conductors and adopted by others - and by having an "acoustical shell" positioned back of the orchestra. During the 20 years he had been recording for RCA Victor he had experimented at separating the various choirs of the orchestra so that the more delicate instruments wouldn't be drowned out by the stronger ones. Now he convinced Disney that an ordinary sound track would be a disservice to the music. The sound should have directional integrity and it should make the orchestra sound "live" rather than giving it a one-dimensional simulacrum of life.

The limitations of SOF recording in 1940 made it necessary to compress the dynamic range to utilize a usable range of about 35 decibels, about half the range of a symphony orchestra. Also, the sound of an orchestra emanates from a broad area, the width of a stage, with added resonance from the walls of the auditorium, rather than from a single source such as the big speaker in a theater of the period. Disney felt that directional effects with the sound seemingly coming from different parts of the screen or even from off screen, would add to the dramatic impact of the animation.

The problems of creating a recording and reproducing system that would double the dynamic range, spread the source of the sound over a wider area, provide additional source points for certain sounds, and create directional effects, were given Garity, who worked out the intricacies of an effective multi-channel, wide range recording system. Important contributions were made by C. O. Slyfield, J. N. A. Hawkins, W. C. Lamb Jr., C. A. Hisserich, H. M. Tremaine, P. J. Holmes, Melville Poche, H. J. Steck and E. A. Freitas. RCA Victor took an active interest in developing the new technique.

The Sorcerer's Apprentice, starring Mickey Mouse, was in the planning stage in November 1937 and began actual production at the old Disney studio on Hyperion Street on January 21, 1938 under the direction of Jim Algar. The music was recorded in the first version of Garity's multi-channel system at RKO-Pathe Studio in a three-hour session starting at midnight on January 9. Stokowski conducted a specially organized orchestra of 85 studio musicians. The nocturnal session was Stokowski's idea: the musicians would be alert and edgy from drinking coffee to stay awake. When completed the pictures would tell the story (by Goethe) that inspired Paul Dukas' descriptive music without resort to sound effects or dialogue.

The completed Sorcerer's Apprentice in two reels was terrific, but it cost more than $125,000. At the same time the company was in the process of building its new $3 million studio in Burbank. No short subject could recoup such an investment, so Walt Disney and his more down-to-earth brother, Roy, decided to expand it into a feature production by adding other segments. Stokowski and composer-critic Deems Taylor came to the new Disney Studio in Burbank and a long list of compositions was compiled and considered. The final program chosen was as follows:

Johann Sebastian Bach's "Toccata and Fugue in D Minor," orchestrated by Stokowski from the original organ composition; Piotr Ilich Tchaikowsky's suite from "The Nutcracker" ballet; "The Sorcerer's Apprentice;" Igor Stravinsky's ballet score, "The Rite of Spring;" Ludwig von Beethoven's Symphony No. 6 ("Pastoral"); Amilcare Ponchielli's "Dance of the Hours" from "La Gioconda;" Modest Moussorgsky's "Night on Bald Mountain," arranged and orchestrated from the composer's unfinished piano sketches; and Franz Schubert's "Ave Maria," orchestrated by Stokowski with new lyrics by Rachel Field.

The music was recorded over a two month period by the great Philadelphia Orchestra at the old, acoustically perfect Philadelphia Academy of Music. More than 420,000 feet of sound recording film was used to get the 18,000 feet remaining in the final cut.

Eight push-pull RCA ultra-violet variable area optical recording channels were utilized with as many as 33 microphones. Close pickups of violins, cellos and basses, violas, brass, woodwinds and tympani were recorded on six separate channels. A mixture of these channels was recorded on the seventh channel, and a distant pickup of the entire orchestra was registered on the eighth channel. These channels were combined to produce three push-pull tracks. A special optical printer was developed to print the three tracks side by side along with a fourth track on sound track film. The fourth track, which was made after the final rerecordings of the three program tracks had been made, was an automatic control device consisting of a composite recording of three different oscillator frequency-tones. Each control tone was rectified and provided a gain-controlling element for the amplifiers.

Standard-width tracks were used for recording and rerecording but these were blown up to double-width when the final prints were made. This four-track sound positive was separate from the Technicolor picture positive although, as a safety measure, the picture positive carried a conventional single sound track closemixed from the original eight tracks which were used in producing the three program tracks.

Such a recording demands a special reproducing system. The designs utilized standard equipment wherever possible. "A great many equipment combinations were explored on paper, probably several hundred," Garity noted in his report to the Journal of the Society of Motion Picture Engineers for August 1941. "Of these, ten different systems have been built up and tried out . . . Even though Fantasia has been released, development has not stopped."

The seventh system was the first to be built by RCA and cost about $100,000. Subsequent units cost about $30,000 each. The eighth was a rearrangement of the seventh and was installed in the Broadway Theatre in New York for the world premiere on November 13,1940. The ninth was sent out with eight of the roadshow presentations and the tenth was installed at the Carthay Circle Theatre in Los Angeles for the West Coast premiere and first run.

A multi-track sound-head was used on an otherwise standard RCA re-recording unit and synchronized with the picture projector with Selsyn motors. The scanner beam put a slit of light across all four tracks simultaneously and the slit-images were optically projected to their proper photocells. The signal output of each track was fed through a relay fader system and then amplified by separate preamps. Each of the three program pre-amps fed into a variable gain amp which fed a 20-watt driver amp. Each driver amp fed two 60watt power amps connected in parallel, yielding an output rated at 120 watts per channel but actually capable of delivering 200 watts.

Three separate multispeaker systems embodying a total of 36 loudspeaker units were placed on the stage at left, center and right. Each consisted of four large high frequency baffles fed by eight low frequency speakers and one large cellular high frequency horn fed by four high frequency speakers. The additional units used at the Broadway Theatre consisted of two additional 50-watt power amps, each of which drives 22 small speakers mounted on the sides, back and ceiling. The installation was varied according to needs of the various auditoriums.

The visuals consist of Technicolor Multiplane animation of the highest level interspersed with imaginative live action of the orchestra, the conductor and the narrator. Altogether, Fantasia took about three years to make and utilized 11 directors, 60-plus animators, 20 background artists, and a small army of writers, designers, researchers, ink and paint personnel, special effects men, sound recordists, sculptors, models, actors, and others-close to 1,000 persons in all. More than a million drawings were used. The cost approached $2,250,000. At 133 minutes (including intermission), it was the longest cartoon ever produced.

Disney planned to release Fantasia as a concert tour at advanced prices. Because of Fantasound, only a dozen theaters would be able to show it at first, but more Fantasound units were planned and the showings would extend to 76 other theaters. It seemed a good plan, but it was doomed by unforseen circumstances.

First a jurisdictional dispute between the International Alliance of Theatrical Stage Employees and the International Brotherhood of Electrical workers threatened to doom the Broadway premiere. A last-gasp compromise and several 24-hour work shifts by both unions made it possible to open the premiere as scheduled. Then, with America's entry into World War II, national defense priorities halted the making of any more Fantasound systems.

Disney's distributor, RKO Radio, put Fantasia into regular release on April 10,1942, cut to 88 minutes and with a standard sound track. The picture was too heavy for the children who made up most of Disney's regular audiences - and for most of their parents as well. Word of mouth spread quickly and the picture became a boxoffice disaster. Reissues in 1944,1946 and 1953 met with invariably dismal results.

Disney's Buena Vista Distributing Co. brought it out in SuperScopeand "Full Stereophonic Sound" in 1956. The wide screen process made Tchaikowsky's sugar plum fairies look fat and the dinosaurs of "The Rite of Spring" resembled balloons. There were further reissues in 1963 and 1969, after which it stayed in release because a new generation saw in it a "psychedelic experience." So it was that, long after Walt Disney's death in 1966, his "folly" became a financial success at last.

A new edition with rechanneled stereophonic sound was released in 1977, but even this suffered in comparison to the current product. It was decided in 1984 that the sound should be completely redone. Irwin Kostal conducted the new performances while Taylor and Stokowski disappeared from the scene. The Fantasia addicts were livid: it just wasn't the same show without them, even though Taylor's words were spoken by Tim Matheson and the picture bore a dedication to Stokowski.

But now Walt, Stoki and Deems have been avenged.

Fantasia Becomes Tantastic' Again

When Disney executives first considered re-releasing Fantasia with a restored original score, more than two years ago, they took a measured approach. A particularly rough two minute segment of the "Nutcracker Suite" from a copy of the original optical track was sent out to various firms for cleanup to gauge the feasibility of the idea. Could the relatively primitive recording be resurrected to theatrical quality?

The answer had been 'no' in 1980, one reason for the subsequent version featuring new performance recordings conducted by Irwin Kostal. Sound restoration techniques came a long way in the 80s, however, and Disney eventually found the answer at its own Buena Vista Sound. As it turned out, restoring theatrical quality sound was only the first step. The result, after months of painstaking work, was a complete and faithful reproduction of the original Fantasound roadshow experience.

Like the original Fantasia, the sound restoration started small and then snowballed. Oscar winning re-recording mixer Terry Porter used many conventional techniques as well as a few unusual "bells and whistles" to work his magic on the faded tracks, and soon found himself falling in love with the project. Porter recalls the genesis:

"At first it was really an experiment to see if we could get it into some kind of theatrically acceptable shape using the new technologies. So I spent some time, just basically using conventional, up-todate techniques - most of them things I usually do with production tracks in features-and did a little cleanup on the test segment."

The decision was made to restore the sound on the basis of that experiment. Initially, Porter's assignment was to clean up the tracks and prepare them for release in normal Dolby stereo. But as he began to sift through the details and steep himself in Fantasia history, Porter found himself drawn to the project. Before long, he was hooked.

"When I first got involved," he says, "I had no idea how innovative Fantasound was, or that Fantasia was first a stereo show. I spent time in the archives, learning exactly what Fantasound was, and I was amazed. I then went back to the executives with a proposal: to recreate the original roadshow sound experience that they had in 1940 - Fantasound."

Before Fantasound could be recreated, there were months of careful restoration work to be done. Source material was woefully deteriorated. The performances, conducted and overseen by Stokowski, were finished to nitrate optical playback beginning in 1938. Because of nitrate's inherent instability, a transfer was made in 1955 to magnetic tape. This 35 year-old transfer was Porter's source - so far, so good.

There was a problem, however. After the Fantasound roadshow toured the country, much of the 'equipment had been dismantled and contributed to the war effort. By 1955, the only available Fantasound playback system was being stored at the RCA building in Burbank, which had yet to be equipped with the new 35mm magnetic recording equipment. Disney, however, was so equipped.

"It was someone's decision at the time to make the transfer from the RCA building to Disney over phone lines," says Porter, incredulous. "So the source I had to deal with came from a disintegrating nitrate optical that was in poor condition, was made with magnetic recorders - a technology in its infancy - and was transferred over phone lines! So we had a lot of problems with the three-track master. If s easy to see why in 1981 they said it was impossible to restore."

Porter sized up the problems and assembled an array of eight or nine primary tools with which to solve them. "Mainly there was a tremendous hiss level. We removed probably 3000 pops out of it. There were some very bad distorted areas, and occasionally there were areas where the mag stripe had flaked away or was missing. We also found some very nasty hums - four different sets of frequencies, in multiples of 60 - my guess is they were caused by the phone line transmission."

Rather than making a number of passes through the material - one for pops, one for hiss, etc. - Porter applied all his tools and expertise simultaneously, during one long, careful, frame-byframe run-through. This allowed him to hear how each process affected the others.

"The work was done in a theater setting," he states. "It worked well that way because I could evaluate the whole package at once. I was also able to listen to it as if I were the audience in a theater environment, instead of working in a small control room looking at graphs and meters. If I ran across a pop or some sort of system problem within the music, but the music still gave me the emotional impact it was supposed to, and removing this pop somehow hurt that aspect - I left the pop in. It's not worth hurting the music. Being in the theater, I was able to make a good judgment.

"My partner [MeI Metcalfe] and I were tempted at times to do a little 1990s work to some crude forms of panning that they did," he admits, "but we stayed away from that. This is a piece of history, something they did 50 years ago. Our goal was to present it in 1990 like an audience would have heard it in 1940. And we stuck by our guns on that."

One of Porter's strongest weapons was the SAE 5000. "If s actually called a depopper. You locate the pop to the exact frame and then automate it in and out for just that one frame, almost a sprocket's worth-very tedious work. These pops are very, very small - we're talking about a fiftieth of a second - and generally it couldn't be detected that something was being removed. But sometimes, especially if there was a series of pops, they would leave a hole or a glitch that you could actually hear. In that case, I would program in a very short piece of reverb, the same length of the pop, to fill it in. That's a little of the smoke that you use. You're not hurting or changing the music at all. It just samples that fraction of sound right before the pop and fills it in."

Although the material was never actually in the digital domain, certain pieces of digital equipment had a place in Porter's process. "In order to deal with some of the hums and hiss, I used Quantec, which is run by a Mac computer. They have what's called a high-low pass, which cuts off the extreme low and high frequencies, and notch filters, which you can use to notch out some of the hums."

Porter's experienced ear estimates the frequency of a pure hum. With the frequency pinpointed, a few adjustments are made to the Macintosh, which then removes that sliver of the bandwidth. Again, the absence is unnoticeable.

For noise that is spread over more frequencies, Porter employed several pieces of equipment. "As a result of the primitive recording techniques, there was the general hash, this broad banded surface noise-like pink noise, like a waterfall. The Cat 43, as if s called in the business, is a device that's been around a long time, made by Dolby. If s a four banded noise reduction device. We use it in dialog cleanup quite a bit. It works great with camera and generator noise, and broadbanded air conditioner noise, the kind of problem you encounter with production tracks.

"In conjunction with the Cat 43, we used an MCF Band Splitter, which splits up the sound into four discreet sets of frequencies -lows, mid-lows, mid-highs, and highs. I had four noise gates hooked up tojjhis band splitter. A noise gate reduces the amplification of the signal, and then releases itself as soon as the frequency hits it. If, for example, there was a section with some french horns working in some lower registers, I could go in and clean out some of the high frequencies with these noise gates. So I was constantly adjusting to reduce the amplification of various frequencies where there wasn't any music. In conjunction with the Cat 43, it worked really well for general surface noise."

These noise reduction techniques concentrate on cleaning up or limiting the frequencies that aren't occupied by music. But the opposite can also work-a kind of reverse noise reduction. Porter, using his ear, would locate the bandwidth where an instrument was playing, use the GML Parametric Equalizer to sweep the frequencies, and boost those frequencies. Expanding the music, in effect, reduces the noise.

Other footholds were gained because of the way the stereo signal was distributed in the Fantasound system. In most conventional stereo recordings of today, the strings have a left-right configuration, the percussion is center, etc. The object is a general, wide feel, rather than an articulated, moving aural picture. On the other hand, Stokowski, Disney and their cohorts thought nothing of panning an entire string section from one track to another and back. In the three track configuration, if the music shifted completely to the right, Porter was able to eliminate 2/3 of the noise simply by closing his left and center pots.

"All these pieces had to be used in moderation," he cautions. "Anytime you start to use equipment excessively, you immediately start hearing it in the music. In fact, any time you do any noise reduction, you do take out certain elements of the sound that you wish you hadn't. After all our extensive noise reduction work, the music sounded slightly dull and flat, so I used another Quantec with another Mac computer and a digital graphic equalizer to do my salt and pepper - to equalize it back to as close to the original timbre of the music as possible - and was able to bring back some brightness and some life to it."

One other piece of equipment was a Lexicon 480 reverb chamber, which, at first, Porter hesitated to use, because it would be the only break in his f aithfulness to the original recording.

"In the archives I came across a transcript of a meeting between Walt Disney, Stokowski, Garity, Hawkins and Tremaine. They were talking about the presentation, and Stokowski was concerned that the music panning back and forth and popping in and out of speakers so suddenly might be very disconcerting to the audience. Of course, they didn't have anything like an electronic reverb chamber then. So one of the criteria they set up for the road show presentation of Fantasound was that the auditorium must have hard walls and a long reverb time. This would result in a natural reverb decay in the room that would soften all the pans. While the left speaker was coming in, you still could hear the right one trailing off.

"Of course, in 1990, theaters are constructed to be as dead as possible. So I added back what I thought was a suitable amount of natural reverb, using a concert hall program on my 480. It works well - the sharp pans are kind of rounded out a little bit. That was the only thing that I felt justified in adding to the original."

With a clean recording on his hands, Porter turned his attention to recreating the Fantasound presentation. Digging through the Disney archives, he learned exactly how the system was developed, and how it worked in the original roadshow. There were obstacles, however. The three track magnetic transfer told Porter how the music panned across the front of the screen, and the dynamic range - where it was loud and where it was soft. But there was no recorded example for the three-way speaker system positioned in the rear of the roadshow auditorium.

Originally, the sound in the off-screen speakers was controlled by a complicated mechanical system triggered by notches on the side of the sound reel. The notches ran over a roller relay, which turned camshafts, which in turn boosted or backed off amplifiers driving particular speakers. Note by note, the sound was controlled and directed to various parts of the auditorium. For Porter, this was a major missing piece in the Fantasound puzzle, until one day his research efforts paid off.

"I found a real coup in the music archives," he recalls. "I found Stokowski's original score sheets, with handwritten notes on them explaining note by note, phrase by phrase, where the music was to go. It would say, 'Left wall, rear wall, kill the fronts' for a certain passage. It was a complete map to exactly what they did in the auditorium."

The final step was perhaps the easiest - translating the Fantasound into today's theaters.

"The original Fantasound was basically six dimensional (screen left, center and right; auditorium left, rear and right), so we put it on 70mm, which also has 6 tracks. In normal 70mm, you have left, center, right, one surround channel, and two boom channels for the very low frequency information. I just used all six channels full bandwidth for the music. With a few patch cords and a little rewiring in the theater, we were able to achieve those Fantasia dimensions without too much pain."

Throughout his work on the project, Porter was awed by the innovation and creative energy that culminated in the Fantasound system. He wonders what might have come next, had the war and other factors not drawn that energy in other directions.

"I ran across blueprints and diagrams of some fascinating things they were designing at the time, but it all just stopped. I don't know how many people have ever bothered to look at these archives, but I was flabbergasted," he says. "It baffled me to think that these people were able to come up with all these ideas off the top of their heads and actually develop them. We use so many of those same ideas today, only the technology is more advanced. Walt was certainly excited about doing things like this, and Stokowski was a genius.

"We got so excited as we started getting into this thing to find out what they had done - it's all been locked up in such a mystery to so many people. Who knows where Disney would have gone with sound?"

While we ponder that one, we'll have to be content with knowing how far he did go. Porter's work on the restored version of Fantasia contributes to an entertaining and faithful record to study, and more importantly, to enjoy.

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.