Actually, they were. Three weeks of singing, choreography, and acting rehearsals preceeding their recording session on Jan 31, 2000.
No. The actors auditioned seperately, and had never met before the final callback auditions for the film. Interestingly, we did try to get a mildly famous singing group to participate, but they weren't interested.
We had experience with casting for musical theater, and it was basically the same thing. Singing, dancing, and acting auditions all came into play.
I wrote it back in December of 1998, including all the lyrics for all the major songs. I looked for a composer for many months, finally lucking into Scott Van Essen in the summer of 1999. He took many months to compose the music, and a lot of the lyrics changed through collaboration.
The refigerator song was all Scott's idea, though I made him reduce the length from 45 seconds to it's current 18 seconds for screen time considerations. Scott also wrote the music and lyrics for all the finale songs. It's a heap of music.
Hard cash, not including tuition (see 1.5), this film probably cost $24,000. More than half of that is film expense, including 13,000 feet of raw stock, processing, telecine, negative cut, film recording, optical soundtrack, answer print, etc. The rest is production design, food, costumes, engineers, mixers, etc.
Yes, this was my graduate thesis film at USC. For those in the know at USC, this was a '581.' A 581 is an advanced graduate film not to exceed twelve minutes in length, funded by the filmmaker. The school provides no funds or materials, but does provide some equipment and facilities.
The film is over fourteen minutes, and required special faculty approval for the length.
Being a USC student made some things easier, and several key things cheaper. Of particular benefit was group insurance coverage and a SAG agreement already in place. USC also has a very good relationship with Clairmont camera which made for a simple camera package donation. I had access to AVID (which I used sparingly, see 3.1), ProTools, and a rewind bench. Of key importance was access to the Domino system for film scanning and recording (see 2.6), which is now gone from the facility. For this I paid tuition (on this class alone) of around $7,000. Was it worth it? Probably, but not definately.
Oh, and I also had to give up the copryright to the film -- see 4.3.
Just being a student at all provides many benefits, such as free LA permits and discounts at most vendors.
The school provided no actual funds or expenses.
A good question. The choice of format was primarily an artistic one:
1. We wanted a widescreen presentation, since the frame frequently featured five or more people standing in a horizontal line. Why not anamorphic? See 2.2.
2. Presentation soundtrack was very important. I definately wanted 5.1 sound with very good dynamic range to feature the music and better isolate the individual singers in the big crane shot. Only 35mm digital sound would feature this (though theoretically correct HD projection could do it as well). Certainly 16mm (with its built in low-dynamic mono sound) would be unacceptable.
3. Color saturation was important for the film's color scheme. The yellow from the Chorus' sweaters loses much of its saturation in the transfer to video.
Finally, we had a good deal on the camera package and processing, possibly a better deal than Iwould ever get again, so I took advantage of it.
Why not Super16mm? I was reccomended against it for technical reasons, though in hindsight it might have been an acceptable way to go and save money. It would have seriously complicated the CG work.
We shot 35mm on Kodak Vision 500T (5279) film using an Arri BL4 35mm camera. The camera package was courtesy of Clairemont Camera, who commonly provides older packages to USC students. They almost never rent a BL4 to paying customers, so they provide them for student use.
We ended up shooting about 13,000 feet, which is about 9:1 ratio.
1:1.85. We wanted to go as widescreen as reasonable without resorting to those darn expensive anamorphic lenses. Some other reasons:
a) Most of the shots in the film feature five or more people in the frame. Wider aspect ratios provide for better composition.
b) Anamorphic ratios don't transfer to video very well without some sort of cropping. Sadly, since the majority of viewings of this film would be on video, I didn't want to make a hyper-letterbox video [when you show the full 2.20 frame on video] Nor did I want to pan-and-scan: it's expensive and besides, if most viewings are on video, why bother shooting an image most people will never see?
c) 1.185 is a standard accepted ratio for 35mm. Other wacky things, like 1.33 and 1.66 weren't even considered.
We primarily shot short ends purchased from Studio Film & Tape in Hollywood, though we had to use some 1000 foot recans for longer shots and the crane shot.
We used a Chapman 17 foot crane arm with a Power Pod remote camera mount. The Chapman crane was provided at a modest student discount, while the Power Pod was part of the Clairemont camera package. Like most crane shots, the shot took six people to execute: three to push and pull the dolly (it weighs about 2000 pounds), one to control the pan and tilt of the crane arm, one to do pan and tilt on the remote head, and one to focus.
We used video playback to check the shot, and we did nine takes. The storyboards called for one continuous shot. Sadly, the shot was never perfect, and it the final film there is an insert shot of the back of Ed's head where two different crane shots are spliced together.
We actually had 9 track playback on the set, so each singing group could properly lip-sync to their recording. It was loud, as you might imagine.
Lighting consisted of replacing all the flourescents in the ceiling with daylight Kino-flo's and a few tungstens in the corner for highlights.
The shot took about a half a day to setup and execute. When it was done, everybody had sweaty palms, since the shot was critical to the film and so very expensive (crane + many costumes + lotsa production design + burning film).
The car was photographed using green screen to save the time, insurance, and difficulty of an actual tow shot. In hindsight, this was probably a mistake. The processing, work, and expense to do the compositing probably ultimately outweighed the problems of a tow shot.
Many people suggested cutting that scene, but I wanted to keep it in because a) I thought the song was one of the better ones b) the second shot of the empty car for me is the strongest one showing just how much he misses his chorus.
There are four sequences that went through the computer: both car shots (for compositing), the head-flying title, and the credits.
The footage was scanned in using the school's Domino system (which had a film scanner) with the help of Eric Furie. The frames were scanned in at 3k, downrezzed to 2k, and transferred to my computer. I did the compositiing and titling in After Effects using Ultimatte. Most out put was done on the school's Domino (which had a film recorder), and some was done at Foto-Kem.
The schools Domino is now gone, and these were the last frames ever recorded on the machine. In fact, the film recorder was supposedly out of alignment, which may explain some of the softness of the shots. It took several months of delay while Eric with the help of Quantel technical support tried to calibrate the recorder.
The car shots were mistakenly shot on the same film stock as the rest of the film (Vision 500T), which added tons of grain. Since we were outside, it was really a poor choice. Even a normal low asa color stock would have been better than that high-speed stuff. Also, the day was cloudy, so the lighting on the green screen was uneven. In fact, in the second car shot, the green screen was pretty washed out, and the shot had to be done with tons of rotoscoping and multiple matte pulls. The grain doubling on the background plate was reduced by adding extra blur outside the car.
Also, another mistake was shaking the car on set to simulate driving. Not only does the shaking not look much like driving, it all had to be stabilized away before the composite anyway so the plate matches. Then we had to add shaking to the final image. I also added the freeway overpass shadow in the second shot, which seemed to help the composite.
There are still problems in the final film, but the shots are quick enough that most people don't notice.
The heads are scanned directly from the preceding shot in the office, but obviously the alignment has troubles and the dissolve is not a smooth one. We never did figure out what went wrong, and couldn't afford to do more work to fix it.
In fact all the problems could have been fixed given more time and a lot more money. But even with student discounts I would pay one to two dollars a frame for scanning and about one dollar per frame for output now that the school's domino is gone. I couldn't afford that.
We scheduled for 7 shooting days, and finished in 6. I've personally never heard of another short that finished under days.
We also shot an additional two hours of pickups several months later.
We prerecorded nearly all the music before shooting, and played it back on set. Playback was primarily done using a Nagra IV, with a DAT tape ready for backup. The playback on the big crane shot was done with a DA-88 playing back eight tracks simultaneously, with a large number of personal stereo amps and speakers in select locations for the amplification. All sound was recorded using a Nagra IV-TC on one channel, with timecode slating.
When the film ended up in telecine, the telecine operator called us up and reported that our first two sound rolls were completely blank. We still have no idea what happened, the production mixer was a fairly seasoned operator, and claims he did confidence checks. All subsequent sound rolls were fine. This caused us to completely reshoot Ed's close-up in the doctor's office [See if you can figure out which close-up shots are the reshoots and which are the original shots -- they were photographed months apart!], and completely ADR the doctor. In fact, that whole scene is a sound nightmare, mixing original production sound with pickup production sound with extensive ADR.
Worked out okay though.
Chris did all of his editing on Adobe Premiere. The dailies were transferred to DV with timecode intact, and he just edited that. I also transferred the dallies into my computer so he could e-mail me his premiere project files and I could view the edit. I had to write a special script to convert from QT (he was using a MotoDV card) to my AVI files (I was using a Fast Avmaster). We were able to convert the telecine files to premiere to use for batch capture.
When we were about done, we tried (and tried) to transfer an EDL from Premiere into AVID Film Composer. This never worked, and AVID technical support told us it was impossible. It would have worked into Media Composer (for a video finish) but not Film Composer (for a film finish). So, we had to transfer the cut into AVID by hand. It wasn't so bad, since the timecode numbers on the dallies matched, it was just a match the number cut-and-paste routine.
Once into Film Composer, I conformed a workprint so we could view a projected version for further edits. I, with the help of Ann Raziel, did all the further editing tweaks on the AVID. I kept the conformed workprint up to date, but never reprojected it (though I did watch it on the flatbed). Mostly I wanted to check the telecine and give the negative cutter an actual conformed print for matching.
Speaking of telecine, it's a good thing I checked the film on the flatbed! All the sound had been synced up in telecine, and several shots were TWO FRAMES OFF. We found this after the mix, so we couldn't just slip the sound. We actually had to slip frames of film. This was only found while watching the print, and couldn't be detected from the video. Very scary.
Boy, I sure wish I had something like FilmMatch at the time.
I used Pro Tools v5.0 to do all the sound editing. We did our final mix at Disney from the protools session.
We did our own foley in the USC foley room. There are some foley glitches, but it's mostly pretty good.
We did ADR in the USC ADR room, and that came out okay. ADR was mostly Dr. Fred in his office, with some additional Betty off screen lines and general walla lines.
Chris scored us a good student deal at Disney. It was still horribly expensive by my standards, but remarkably discounted from normal rates. It was a 5.1 mix with an additional LtRt (stereo surround) downmix. Dolby donated the encoding to Dolby Digital, and NT audio provided a small break on the optical negative.
Il Mio Coro. I learned that when it was translated for the film festival in Rome.
The film has had 26 festival showings so far, resulting in 10 awards. Pretty good record. There are several more festival showings lined up (see the screenings page for details). It doesn't get into every festival we apply to, though, so it's not like this is the greatest short film ever made, according to those that judge such things.
We have finished the soundtrack, which sounds great. It contains 26 minutes of music, including all the finale songs (all 9 of them). A complete lyrics sheet is included in the CD. 26 minutes of music is pretty remarkable for a 14 minute film. The soundtrack is available for purchase on the merchandise page.
The copyright on the music is NOT owned by USC, but by Scott and I. See Q4.3
All films made using any University equipment connected with a class become property of USC. This is a very controversial issue, since a lot of students (me included) contribute huge amounts of personal funds into their film productions, as well as paying a pretty steep tuition (something like $7800 for this production class alone).
In the past, this wasn't any big deal. It didn't affect the student's ability to send the film to festivals or make copies for his personal use. And there really wasn't much of a market for short films. Now things have changed a bit, making things less clear. There is some market for short films, including DVD compilations, airline in-flight movies, foriegn television, some domestic TV (sc-fi channel, IFC), and the internet. Remarkably, the university splits the revenues with the student (which it doesn't have to since it owns the copyright). They also provide complete legal assistance in signing distribution contracts and numerous connections to sales channels, so actually the chance of recovering some revenue is quite a bit more likely with the University's help.
Currently the USC festival office forbids the distribution of the film on the internet without permission of the festival office. This is the first time in history where USC has really exercised its copyright in the exhibition of the films. The festival office is actually very well tuned to the needs of students, and having the full length version on the internet can really destroy a film's festival chances and distribution chances, so I feel they are making the right decision here. Still, it's weird. In general, I feel the University's backing in supporting the film was worth losing the copyright for the short film.
For a strongly negative opinion about this, check this.
|© Richard Doherty, My Chorus, 2003||
For info: Richard At TheDohertys.com