Penn Jillette, the taller and more voluble half of illusionist team Penn & Teller, knows how to beat a horse to death and then some. In the case of The Aristocrats (www.thearistocrats.com), that horse is one of vaudeville’s most famous jokes. The film, which drew critical raves, has 105 comedians, including Robin Williams, George Carlin, Andy Dick, Richard Lewis and Steven Wright, telling and retelling the thing, each trying to outdo the other for extreme levels of filth and vileness. The film’s grainy, backstage feel contributes to the rawness of the experience. However, that grubbiness extended to the production audio, making The Aristocrats a candidate for major sound surgery. (Keep reading – Film & Video has the audio files to prove it.)

Eric “Travis” Wilson settled in at Jillette’s home studio, Vintage Nudes, at the magician’s architecturally notorious house in the desert near Las Vegas last year to post The Aristocrats in an environment not unlike that in which the film was shot. It’s an intimate studio featuring touches like classic Bettie Page-era pin-up playing card decks covering the horizontal and vertical surfaces that aren’t filled with gear – workstations running Pro Tools and LCD displays – a space that’s representative of the way movies are increasingly being post-produced in the DIY digital age.

“I’m a big fan of getting great performances out of cheap, crummy gear,” says Wilson, referring to the hand-held consumer-grade cameras and mish-mashed Best Buy microphone systems the movie was shot with.

What Wilson faced was track after track laden with background noise and hiss, level inconsistencies and other spurious artifacts. The problems derived from issues like mismatches between line-level and mic-level input sources, signal-to-noise problems, drop-outs, clicks, pops – you name it. Then there’s the scene in which Billy the Mime acts out the central joke – completely silent, except for Penn Jillette’s cackling laugh as he unsteadily holds the camera. (In a bit of technical irony, viewers will notice that Billy wears a wireless microphone pack anyway.)

Wilson got the sound from the audio tracks of the Avid production cut. There are plenty of plug-in solutions for all of these audio gremlins, and he wanted to apply them in real time to create a cohesive, consistent final soundtrack. But, as he points out, each one-minute-or-less segment had its own particular issues. “And programs like No Noise are real memory hogs because of how many calculations they’re doing per second,” he explains. “If you have more than a dozen or so [events] running simultaneously, it crashes the system.”

The solution was to analyze and categorize the problem areas, determine which combinations of them applied to each comedian’s audio track, then create “fix” stems: tracks with various combinations of processors, such as No Noise, Waves UltraMaximizer and Focusrite EQ and comp/limiter plug-ins that each audio track could be dropped into for real-time corrections.

For instance, one such fix stem was named “Phyllis,” for Phyllis Diller. Diller’s track combined the problems of background noise, level inconsistencies and EQ issues. “I listen to the other tracks and determine which ones have the same sets of problems, then put those tracks into this [stem],” he says. The Waves Maximizer plug-in was used to set certain threshold limits on distortion areas. “I used it to set an upper limit of 0.3-dB below zero full scale,” he explains. “That 0.3-dB threshold is so that the consumer-level [disc] players won’t see full zero and distort.”

Given the huge number of segments, the corrective techniques required certain compromises. Within the fix stems, for instance, EQ problem resolution generally had low-end roll-off applied. “It’s nice to try to keep the fullness of the voice by rolling off around 90 Hz,” says Wilson. “But the reality is you have to go closer to 200 Hz to get at the full scale of some EQ problems.”

Automation on the Digidesign Control 24 console was often used to deal with dynamic inconsistencies, with Wilson riding faders to even out the level. But other problems required more complex solutions. At one point in Jay Kogen’s delivery of the joke, the hard consonant “C” was clipped from the word “calling,” resulting in “…alling.” Wilson trolled the other tracks until he found a similar sound ‘ in this case, it was Drew Carey speaking a similar word ‘ and he cut and grafted it from there to the front of the truncated word on Hogan’s track.

At one point, a drop-out compelled Wilson to record himself saying an entire line and then cross-fade it onto the errant spot. “So it’s actually my voice in the movie for about 110 milliseconds,” he says.

In general, Wilson approached the processing with a relatively light touch, seeking balance between eradicating the problems and the new artifacts that are collateral to the restoration process. “I was thinking about the playback environment for the theatrical version. I had to think about air loss and dispersion from speaker systems that are going to get rid of the upper octave on you anyway,” he says. “A little sparkle of hiss on the top wouldn’t have been heard by anyone, but the artifacts resulting from the aggressive noise reduction might have. I actually did a quieter version but I backed off on it.”

Good. You’re not going to want to miss a word of this movie.