Fiddler on the Roof Sr. Practicum – Jacksonville State University – March 2008 Sabbath Prayer 0:00 / 0:00

Fiddler 13 Photos

The audio recording of Fiddler was part of my final project at Jacksonville State University. It was actually two classes; Special Problems in Drama + Sr. Practicum = six credit hours of independent study. The result of the classes is a minimum 40 page paper detailing the plan, the execution, and the experience.. The first part of the project was to handle all of the live audio needs of the production, and second to make a multi-track recording of it. One of the first challenges is to find ways to accomplish both goals with one engineer – myself.

Jason Bozman - Gadsden AL

Even though the recording of the show is forever and the experience repeatable, the live audio production is paramount to the recording. One reason is that we only get one shot at any given one of the fifteen or so performances. For the recording, only one of those performances has to be a keeper. But perhaps the most important reason to give priority to the live audio production is the audience, who paid good money to be there.

The sound reinforcement needs of the show are rather substantial and requires quite a bit of attention during each run. Our input list consisted of about 24 channels of wireless “face mics” using Countryman E6 and B6 microphone elements and a 14 member pit orchestra. Also, we used a couple of microphones on each side of the stage in the wings. Performers who were not onstage, or about to make an entrance, sang around these mics to provide reinforcement to the chorus component of the ensemble. As a side note, we used a closed circuit video system to provide a view of the conductor backstage. In the pit were two percussionist, double bass, 2 keyboards, 3 trumpets, 2 trombones, 3 woodwind players with various instruments, and a sad violin.

The theatre seats nearly 400 and is very intimate making the sound reinforcement needs unique.  The composition of the orchestra is designed to meet the minimum requirements of the score, still the orchestra will most always be too loud for such a small theatre. The face mics are used to add just a little kick to the performers voices to balance the pit. Usually very little coverage of the orchestra is necessary using only a DI for the keyboards for the onstage and backstage monitors, and a bass DI or mic, to help support the overall ensemble sound. In this instance however I added a series of spot mics, one for each section and an overall stereo pair. The intention was to use these mics for the recording only, but I did find them useful at times in the house.

One of the big goals, that is always present when doing sound reinforcement for a musical, is transparency of the sound systems influence. We don’t want the audience to know we are doing anything at all, and go to great pains to conceal our physical presence, and the presence of our work. Often this requires careful time alignment of the various components of the system, to ensure the voices appear to be originating from the mouth of the performer, even though it would be physically impossible, for these singers, to balance the orchestra without a little help. However, that is a long discussion for another day. Still, the topic is covered in detail in the paper I wrote documenting the project, as is the entire sound system alignment procedure as well.

MY16AT ADAT I/O (bottom)

Now that we have touched on the requirements of the sound reinforcement part of the project we can consider what we need to do to record each performance. Since the house mix will require nearly all of my in run attention, the recording system needs to be practically self sufficient. Because this is a university show, my equipment choices where limited to what was available at the school, and whatever I could acquire on my own. That is not to say the inventory at J.S.U. is lacking in any way. In fact, it is quite nice and includes a Yamaha LS9-32, 20 channels of Sennheiser wireless, and an adequate F.O.H. speaker system.

The LS9-32 has the ability to directly output each input channel to any output on the console, from anywhere in the input channel’s signal path. Because the LS9-32 is limited to 16 outputs, it was necessary to add two MY16AT expansion cards giving us the necessary number of outputs to cover our speaker system, and the “1 to 1″ recording of every input channel. The MY16AT transmits its I/O in the ADAT format which is convenient because many computer audio interface’s support it.

At the time the best option we had for adding 32 channels of simultaneous I/O to a computer  was the M-Audio Lightbridge.

M-Audio Lightbridge

The computer for the gig was a monster Intel quad core PC running Sonar which ran flawlessly for fifteen performances. The computer recorded continuously for nearly 2.5 hours each night and never thought twice about it. One reason the DAW was so stable is that the PC was not worrying with AD conversions just shuffling the information from the console to the hard drive in a manor humans can make sence of. The primary recording drive, by the way, was USB 2.0. – seriously – and it worked just fine.

Direct outputs where taken from each input channels post AD conversion and before any of the channels processing resulting in 32 tracks of raw input. Each night, I have to hit record and make sure that no inputs are clipping. Since all of our meters are post AD, clipping will occur at the same time on the channels input as it does on the recorders input. If for some reason the two clip asynchronously Yamaha has provided a “direct out level” to compensate.  This makes the rig very low maintenance and easy to monitor.

F.O.H. Position

In the Mix

The largest body of work in the post production process for this particular situation was selecting which performances to use, and editing them. Fifteen performances times 2.5hrs each equals  37.5 hours of audio that has to be sorted and evaluated. What I was looking for was basically a performance of a particular song that contained no “deal breakers,” which is any snafu that cannot be easily edited or covered. For example, a snafu could be a performance issue like intonation, major timing errors in which the ensemble falls apart for a period of time, or perhaps someone missed and entrance. Maybe the snafu was a technical issue like RF interference in the wireless system, or a faulty microphone element, which happens very frequently despite our best efforts. So the goal is to find the best performance of a given song and the best recording of it. Given all the human elements involved with a production of this scale, out of fifteen performances I ended up with one or two choices per song

The DAW was recording all inputs continuously, regardless of whether or not the performer was onstage, in the wings, or in the bathroom. ( They were well warned and besides – that’s show biz). I find that 90% of mixing music is preparing to do the mix. Organizing the tracks, creating and routing buses, setting up effects, and cleaning out all the garbage is a lot of busywork, but if you slack off here, the results will not be anything like they could be. The creative stuff simply cant happen, for me anyway,  until the tracks are all cleaned up and a work flow is established.

I began mixing by working on getting the orchestra straightened out first. Once I had a basic mix of the pit I then started working on getting the vocals together. I found blending as many as 20 individual voices a very time consuming thing to do. A good ensemble blends itself naturally with each part adjusting itself in relation to all the other parts continuously. What this means for me is a lot of automation. I don’t have to recreate the natural balance and blend entirely by artificial means. The adjustments made by the performers are still there. What is missing is the point of reference to which the performer was adjusting. In short, we don’t know exactly what the performer was hearing, which is exactly the information used by the performer to blend naturally.  To make matters a little more difficult, because the performers do not stand still on stage, that point of reference is constantly changing.

The first step is to organize the vocals by leads and chorus, part, and into male and female groups. Then I started to build the vocal ensemble from the bottom up against the orchestra. Once a general balance is achieved I used automation to adjust for any balance issues that occur over time, probably caused by the performer trying to balance within the ensemble he or she was hearing. This is not necessarily difficult work, but it does take patience.

It is important to note that the intention is to not to create the best possible performance in post, but to make the best performance sound a good as it possibly can. A little dirt never hurt anyone and after all, its live.

To Life!!!