We've just finished our first weekend of previews here at Kansas City Repertory Theatre, and audiences are loving our show!
I've been here at KC Rep for 2 weeks, designing their production of Evita in the Spencer Theater on the UMKC campus. My journey on this show, though, began back in April when I first heard that I was being considered for the design position.
From the very first phone conversations, I was asked how I was planning to conquer the big issue of the production: how to make seven people sound like 30. It's an Evita with a small cast and an even smaller band. With only 10 adults in the cast (three of whom are Eva, Che, and Peron) and only seven musicians in the band, the trick would be how to make the sound full and satisfying while having only a fraction of the people the show was originally conceived for.
But, we'll get to the details of that in Act II, so be sure to come back for even MORE exciting information! For now, let's get into...
My primary aim with most musicals is making sure that the vocals are clear, the words are distinct, and the orchestra is full and satisfying. The clarity of lyrics can sometimes be impinged on by a loud orchestra, so these goals are not always in line with each other, but there are a few techniques that can help out a lot in these endeavours.
I made two decisions almost immediately: this show would have multiple "vocal zones" and it would be a speaker system comprised of point-source boxes. At most big rock concerts today, you'll see these huge long line arrays of speakers hanging left and right. They are very good at covering huge areas with a very even sound and so are great for big audience areas. Line arrays can be used in smaller theaters, too, with great success, but I knew with this musical, I wanted to use more, smaller speakers to distribute the sound. With only 2 or three arrays of speakers, you're sending a lot of sound from a small number of sources, but with more smaller speakers, it's easier to source the sound to the stage (that last sentence was a huge simplification.) Anyways, it was my instinct that this show did not want to be a line-array show, an instinct I stuck with.
As I began to receive drawings from the production staff and the scenic designer, I immediately recognized the design of the theater - it looked almost identical to the seating layout of the mainstage theater at Centerstage in Baltimore. Both houses have seating configurations in a fan shape, spanning about 120º, so the apron is a little deeper than a standard proscenium but not quite a thrust stage. So, I decided to design a distributed LCR system for each of the three seating sections, fanning out from the stage. In this system, each seating bank gets a left speaker, a center speaker, and a right speaker. You end up with some strange speakers seemingly pointed sideways in the middle of the house, but I promise it works.
The 3 speakers in the "Cluster" and the two "Outer" speakers look fairly typical for a musical sound system, but the "Inner" pair look like they're strangely hanging in the middle of the house, pointed at the wall. But, for an audience member sitting in the left bank of seats, it gives them three speakers (Outer Left, Cluster Left, and Inner Right) from which to hear the show and gives me another speaker to help image the sound of the vocals to the performers onstage.
Image is REALLY important. Like, essential. When sound designers spend hours listening to what sounds like random static noise, we're actually aligning the sound system - we're making sure that all our speakers work together to create a unified sound, emanating from the stage or wherever the sound needs to "source" from. The main variables we use are volume and time. By adding very small amounts (milliseconds) of delay to the speakers, we are able to align the sound coming from a performer on stage with the sound coming from a speaker, even though that speaker is far closer to the listener than the performer. This tricks our brains into believing that the sound we are hearing is coming from the performer.
But, as performers move around the stage, their distance to the audience changes, and so the sound from the speakers suddenly is reaching the audience's ears before the performer's acoustic sound reaches their ears. Within a few milliseconds, this isn't a problem, but eventually, the effect will be lost and all our hard work will be for naught. One solution is to have not one delay time for each speaker, but multiple delay times. Whaaaaaat?? Yeah! So, we make a matrix (just like math class in high school) that lets us delay a person's voice differently depending on where they are onstage.
Let's say Taylor is standing down center, singing. I might have one setting for the speakers for Sam when they're singing downstage, but a completely different setting for the speakers for Jamie, who is far upstage and on a tall platform. With a matrix, I can make the sound of both these performers reach my audience's ears at the correct time.
Confused? Me too. It's a ton of numbers, a healthy dollop of trigonometry, and a LOT of listening to make sure the math is actually correct. But, if I'm successful, the sound reinforcement will be barely perceptible. All that work to make sure no one notices. Maybe that's why we don't have a Tony anymore.
The image above is from the planning phase of my design for Evita. The colors correspond to potential volume (darker colors corresponding to "quieter") and show that, with the outer, inner, and cluster systems all on, the front section of the theater has a nice even bright-green volume. There are a few hot spots in the middle of the left and right banks of seating, but those will be easily evened out by turning down the Inner Left/Right pair of speakers ever so slightly. Also, I like to draw the stage in ArrayCalc so that I can see not only what's hitting the audience, but also, hopefully, that it ISN'T hitting the stage! I like to avoid feedback if I can help it. :-)
If you *really* want to nerd out, you'll notice that these 4 colorful plots are specifically for the frequency 4000 Hz, which is, in my opinion, a pretty good frequency for musical theater prediction, because it lands squarely in the sibilance and vocal-clarity frequency range. If you were to predict at 8kHz (one octave higher than 4kHz) you could quickly drive yourself crazy by trying to get an even response across the house, due to how small the 8kHz wavelength is, and going lower than 4kHz could mean you miss some valuable information about speaker coverage. That said, speakers can do some crazy things in the 200-600 Hz range, so it's important not to ignore the low end, either! But 4kHz is where I start for aiming of speakers and speaker interaction. Whew. This paragraph could easily be expanded to an entire article/blog post all on its own...so I won't continue here.
So there it is! The beginnings of a system design for Evita! I hope you'll come back after intermission to read me wax philosophical about vocal effects...