Monday, June 23, 2014

From Beijing Opera to Fart Noises

This past few weeks has been a trip from, well, not from exactly sublime, and not to exactly ridiculous, but certainly of extremes in sound design.

To me, every show is different. Some times that difference is best expressed as a choice of palette. Other times it seems to reach deeper into a philosophy of design.

I did a children's production of "You're a Good Man, Charlie Brown" a while back, and in keeping with the high concept of the production -- that all the action on stage was an elaborate make-believe being played out by children on a playground -- many of the key sound effects were sourced from children's voices. I had a very amusing VO session where I asked our young cast to improvise airplanes and machine guns and AA for the Snoopy vs. the Red Baron number that opens Act II.

For a recent production of "The Drowsy Chaperone" my philosophical take was that the sounds of the play-within-the-play were obviously pre-recorded effects, but everything that belonged in the world of The Man in the Chair was ultra-realistic -- as if we were actually inside his New York apartment. This was why I rang a real telephone bell, and also attempted to use the actual speakers of the vacuum-tube phonograph prop.

Some of these choices are dictated by the material. It is traditional, for instance, to use a recording of Mel Brooks himself at one place in "The Producers." Many effects -- particularly for more "stagey" or meta-theatrical works -- may be implicitly written into the musical score, with the expectation that they are to be performed by the pit orchestra.

So, Disney's "Mulan Jr." This is another of the youth-theater friendly Disney musicals. They are a mixed blessing; on the positive, you get the name recognition of the originating Disney property, and you get a script which is designed to make maximum use of the large casts typical of children's theater, providing enough interesting roles that no-one need feel left out. The musicals are very professionally put together, with detailed staging notes, lots of neat little side notes, and other material of interest to the young actor (or the inexperienced theater company).

On the flip side, they are also in at least some part labors of love by people who really, really know musical theater. So there are theater in-jokes, but are also filled with musically challenging material. They inevitably suffer from the curse of Meredith Wilson (that is, at least one song in which several different people are simultaneously singing different musical lines).

They are also expansions of source movies that generally had many fewer songs. A number that drops again with licensing issues (or perhaps just performance practicalities) for some of the pop music that accompanied the original. And, as in all musicals, the added songs are equally hit, and miss.

Mulan of course already walks the difficult line of cultural appropriation. The team that put the movie together went to China and collaborated with musicians there. The resulting score is a subtle blend -- avoiding the ching-chong of "Flower Drum Song" and hinting of traditional Chinese melodies and song structure even as it is a thoroughly Western score. As originally scored, there are parts for a small number of traditional Chinese instruments to pad out the rather large orchestral needs.

Well, sort of. There actually isn't a score per se. When you license the script (from MTI) you get a rehearsal piano score and pre-recorded backing tracks. It was not expected that you attempt to perform live accompaniment for your production of the musical.

Well, in any case, our original production concept was to consciously use some of the staging and vocabulary of the Chinese Opera. Minimalist scenery and props, color cues from the Chinese Opera, masks, a dragon dance costume for the character MUSHU, etc.

And it seemed to me that this concept translated into seeing the sound effects as being created as they would be for the Beijing Opera stage; instrumentally. The script itself claims that effects are produced by the visible actors.  In practice, however, this was also as much miss as it was hit.


My experience was a sometimes-painful collaboration with the music director. He had the chore of getting through the wealth of material with limited pit resources. The best choice was piano, bass and drums, and he was busy enough just writing out parts for the other instruments out of the vocal rehearsal score.

We messed around with adding synthesizer keyboard, or additional (mostly improvised) percussion elements. We also flirted with the idea of pre-recording some elements of the scene (not necessarily anything that was already in the written score). The above excerpt is the only survivor of those experiments.

In the final production, the score was realized by the pit, but the Opera elements were brought in with selected sound effects, some character accents, and the introductory percussion (I stripped out all the pitched elements).

The "wind" and "avalanche" cues heard above were also used during the big action scene. When the actual "armies" appear in the play, the cast was pounding on the floor with bamboo sticks. I went and recorded those same sticks and they are the prominent percussive element in that cue.

To break down who did what took several days of crawling through script, vocal score, Disney backing tracks, director's notes, rehearsal notes and recordings, and a full orchestration we managed to find online (in Finale format, but Wine wins again; I was able to run a free Finale reader written for Windows).

In the end, it was a lot of work, the blend between the worlds was not as smooth as it could be, and we fell far short of the kind of intense concentration on that staging conceit that is used so effectively (although from a different cultural tradition) in Sondheim's "Pacific Overtures." But I think in the end the use of a few gongs here and there, and the concentration on percussive elements for sound effects instead of realistic samples, gave the show a distinct and interesting flavor.

A half-dozen shows later and I'm doing "Shrek." The high concept for this production was pop-up book. In my mind, a core idea was that of escaping the dominant narrative; that SHREK, FIONA, and the other fairyland creatures struggle against the expectations of others. "You mean the stories that say I'm a 'wicked' witch?" asks one. And against their own internalization of these narratives as well; FIONA's "This isn't the way it is supposed to be!" and SHREK's own "I guess I'd be a hero..." Which is beautifully realized in SHREK's song at a key moment; "An ogre and a princess, it's complicated. You've never read a tale like this. But fairy tales...should be updated."

But this didn't help me much for sound. I was left with more low concept than high concept. Because the set realized the "pop-up book" idea with lots of things that unfolded and turned, and changes were made in partial light, I chose to accent the scene movements. My original desire was to put paper elements in it, as if it was book pages or pop-up illustrations in motion, but the final versions were various sorts of wood and stone sounds.

The other idea I went in with is that SHREK is an ogre and as the play says, "Ogres like nasty." So I wanted plopping and squishing and other disgusting sounds in there.

And, after all, there is the fart song. It is basically a retread of "Anything You Can Do" from "Annie Get Your Gun," only with farting and belching as the contest. I had the delightful experience of putting every fart and belch from my sample libraries into one long file, and adjusting EQ and pitch and normalizing until they all sounded like they existed in the same sonic space. Then I assigned them across the keys of of MIDI keyboard using the low-profile shareware sampler VSamp Pro.

To reduce the clutter in the pit, I'm actually running the MIDI from a 24-key mini-keyboard through a stage snake back to my own computer, where the sampler player is resident. This also allows me to assign the bodily noises to a dedicated channel on the FOH mixer, so they can be dialed in to the appropriate level with whatever vocal energy the cast has that day.


The most difficult cues were the dragon cues. Partly because the realization of the dragon in scenery and props and costuming happened so very late in the technical process, meaning I didn't have anything to guide me by. But largely because it just took so much work to come up with the vocabulary for the dragon.

The above is my third attempt to create the wings effect. There would be more than three attempts, but we already opened.

The basic wing sound comes from a foley session I did way back on "Peter Pan." I set up a nice mic in a quiet room and whipped a sheet around in front of it. Unfortunately the room wasn't as quiet as it could be, and I had to apply excessive levels of noise reduction to the result.

(As a sideline, the abandon-ware SoundSoup has resurfaced with a new company, and I spent the money to upgrade it to compatibility with my new computer. And I haven't used it yet; the noise reduction available in the free software Audacity has suited me just fine so far.)

I combined clips of the sheet noise with some generic "whooshes" from the library. I tried adding a snip from a fluttering wing to give a little more character, but in my mind dragon wings shouldn't have feathers. I also tried resonance filters and even ring modulator to get a little more lizard skin feel in there, but none of that quite worked and I took it out again after playing with it for several hours.

Of course there's a whole bunch of pitch-shifting and time-stretching going on, as well as a general down-sample, and in the final cue, subharmonic synthesis, false stereo, and a slathering of reverb.

By the time I was ready for this final take we were rehearsing with orchestra and in time, so I could take a reference track. The onboard microphone of the laptop was good enough for that. I assigned up-stroke and down-stroke to a sampler, and "performed" the wings into a new track in Reaper.

In previous attempts I'd added a little pitch bending for Doppler Shift -- which was a lot easier once I discovered Reaper's Pitch Envelopes. Unfortunately the pitch algo is not as flexible as the one I'd been using for this kind of effect previously, and the artifacts were simply too objectionable.

Instead of doppler (and volume envelope) I combined the wings track with a "by" of a hanglider from the BBC sound effects series. That and a little bit of pitch-shifted rattlesnake -- purchased for the show at SoundDogs.

These elements gave me most of a vocabulary to build the other dragon actions; tail lash, head appearance, and a big chomp on FARQUAD...but there wasn't enough time left before opening night to properly employ them.

To add to the complexity of the wings sequence, the director needed them to pass into the rear surround speakers we hung for this show. I had originally baked this into the cue, but vagaries in the performance timing made it necessary to split it up into manual fade cues in QLab. So the first "Go" in QLab starts the wings, the second one pans them into the rear speakers, the third returns them to the stage. And with my free hand, I am also attempting to throw some of the DRAGON's vocals into the rear speakers to follow the wings there. Which has not so far worked as well as I'd hoped but I have seven more weeks of performance to get it right...

1 comment:

  1. Nice retrospective on these sound designs! Regarding Doppler effects, I too have issues getting artifact-free pitch shifts in Reaper. It's not a panacea but have you heard of the (free!) Proximity VST by Tokyo Dawn? It's given some nice results for subtle psychoacoustic shifts and some good wilder ones with lots of automation applied. By nature I'm not much of a commenter but I enjoy reading your posts and thought this might be useful to you in a future production.