Black Box Theater Spatial System Tuning

This week I am involved in an event featuring onstage puppetry with some unique technical challenges. The production was originally staged in a small black box theater with carefully controlled sightlines. Adapting the touring version of the production to the 700-seat theater required an out of the box approach – or more literally, an in the box approach.

The audience was seated on the stage on risers, and the stage right wings became the performance area, with a grid rigged overhead for sound and lighting. The grid was surrounded on all four sides by theatrical soft goods to mask off a box, leaving only a small proscenium-like opening in the front where the puppetry performance would take place. The rest of the “box” was filled with lighting and audio equipment.

The sound system was an 8-channel spatial rig, designed to provide the perception of sound cues coming from various locations throughout the performance area.

In addition to a standard flown L-R point source pair for mains, there was another three boxes flown across the upstage edge of the area in an L-C-R configuration (I will refer to these as the “surrounds” for lack of a better term). There were also a pair of sidefills on the deck, firing across the performance area, and a pair of 18″ subwoofers placed under the audience seating area. 

The show’s technical documentation organized the system into three stereo pairs: Main L/R, Surrounds L/R, and Deck L/R, plus the Surround center and the subwoofers, which makes 8 channels in all. The 8 channels would be driven from an 8-output QLab file that travels with the show. My job was to arrange and optimize the system ahead of time.

Optimization Strategy

Since I wasn’t familiar with the specifics of the production or the design, I had to make some educated guesses in terms of the tuning. My approach was to get the entire rig functioning and tuned to a neutral baseline, with the levels approximately balanced when measured from the middle of the audience area. From there, the production’s sound designer could move forward and modify that as desired.

Main L/R

I tuned the Main L/R pair much like I would tune any L/R pair in a theatrical context. I started by measuring the left side on-axis at three positions from the front to back of the performance area.

We can see an important principle at work here: we aim at the furthest seat, and as we move forward in the seating area, we are gaining level due to proximity but losing it axially. Ideally we offset the two and end up with a minimum variance situation. Note the consistency above 1 kHz in the above traces. 

The pink trace is taken in the last row and has the flattest response, and the lowest level (furthest away). The dark red trace is taken in the first row of seats, which is closer (louder) but also far off-axis vertically, so we see a significant HF rolloff. The middle trace (mint green) splits the difference. Since the loudspeaker loses directivity below 1 kHz (when the output transitions to the woofer), we see simple gains as we get closer due to proximity – axial losses in this lower frequency range are negligible.

It’s a point source so we need to pick a single EQ solution that’s a best fit. Here’s an average created from the three traces in black, the inverted EQ (1/EQ) trace in green, and the result in red.

The seating area is a series of decking in stepped heights so we must be cautious of strange ground reflections here, and I ended up reducing the 300 Hz cut a bit after taking this measurement.

The process is repeated for the right side, and I’ll show it via an “optical average” this time rather than have the analyzer do it:

Here we can see that the EQ cut at 280 or so is placed a bit too low for the bump in the response around 315, so I moved that up after taking these measurements. Let’s look at the match between the final L and R average responses:

I am not concerned with the response below 63 Hz or so – that’s mostly environmental noise in the measurement, as indicated by the lower coherence. (Since it’s an average, the coherence is a bit higher than it is for the individual traces.)

Surround L-R

Now we move on to the flown L-R surround pair. These sources do not have a direct line of sight to the audience seating due to the masking drapery. Matching the HF response of the mains via immense HF boost is obviously inadvisable, so my plan is to make sure that the pair is symmetrical (match each other in tone and level). I measured from the center seat of the front row, but any position equidistant from the two should suffice.

Here we can see the stage left box (maroon trace) has a significant level advantage. This could be because the other side has a cabling issue (lost a leg on the XLR) or because the level pots on the back of the boxes got bumped after I set them. Since the rig is up in the air, I opted to simply adjust the level of that output at the console until they matched. If further issues crop up, time to get out the Genie lift.

How much level offset? In Smaart, use the Ctrl+ Down Arrow keys to lower the trace until it matches up with the other side. (You can also click and drag the trace up or down directly.) This ended up as 4 dB, which I adjusted at the console output. Now we are matched:

There’s a small HF difference but I’m not going to go after that just yet given the environment (I’m measuring amidst Genie lifts and lighting booms).

Surround Center

This box comes reasonably close to the flown L/R (note the clear effect of reflections on the phase trace between 1k and 2k. I’ll revisit this in a bit.

Deck L-R

The lighting crew was building booms while I had to verify this system, so I moved both boxes into the same location side by side, set the measurement mic on the floor about 15 feet away, and compared their responses:

The slight phase deviation up top is because I didn’t bother resetting the analyzer’s delay when I switched boxes, and of course they are not exactly in the same spot. Just for fun, we can note that the two traces reach 180° offset at about 6.3 kHz, or a time offset of 0.08 ms, or a distance of about an inch.

They’re not in their final position so I’m going to hold off on EQ for the moment.

Subs

These are a pair of the new QSC KS118 and they’re awesome. They’re stuffed up under the middle of the rearmost seating riser (the only one tall enough to fit them underneath).

I’m running them in Deep mode for extended LF response, with the XO set to 100 for the time being. 

Leveling

Now that everything is behaving, I placed the measurement mic dead center in the audience area and set the levels to something approaching reasonable.

For this I use relatively high levels of smoothing because the general trend of the vertical offset is the main thing I’m paying attention to at this point.

Here is main L/R along with the rear L/R surrounds. The HF rolloff from those is evident as they’re firing through thick drape, but we are matched well from MF on down. Any further EQ on this will be dictated by the wishes of the designer upon arrival.

Now we add in the center surround box (purple):

And here we have an interesting choice to make. The box has a tonal response matching the flown surrounds, but it is currently level-matched to the mains in the HF. Our options are:

1. Shelf down the LF so it matches the mains in the HF, and the entire rig in the LF. I would choose this option if it is to be operated as an independent center channel.
2.  Attenuate it by approximately 8 dB so it matches the flown L/R surround on either side in both level and tone. This is the option I would choose if the three are to be operated as a L-C-R soundscape.

I ended up choosing the first option to start, and make a note to ask the designer if a different approach is desired. Here we see that box’s response in purple with the 1/EQ in bright green – simply a -4 dB low shelf up to 1k.

In the event that the designer chooses to have Flown C match Flown L/R, I can simply bypass the EQ and lower the gain by 8 dB to match the response of the side boxes. 

I don’t yet know the final placement of the deck sidefills but they are a dual 12 box aimed off axis to the audience and placed against a stage floor, so the LF buildup is much more than all the other boxes in the system:

Again, this is a decision for the designer but to start I used a slight LF shelf to tame it just a bit and bring them closer to uniformity. Below we see the deck fill responses before and after the shelving EQ, with the inverted EQ response shown in green.

The end result is that when a signal is panned between the loudspeakers, the level should stay relatively consistent. The major choice to be made by the designer at this stage is whether the upstage center box should match the mains – as it does currently – or be rolled off at HF to match the upstage L/R.

Thanks to Fletcher McDermott for his help with this post.