Dimensional Focus: Vertical Plane
Designing, deploying, and aligning a modern sound system is a very data-driven process. This raises the question of how much data to gather. How detailed do our predictions need to be? How many measurements do we need to take? My view on this topic is that I want enough data to make a fully informed decision, but stop at the point where additional information doesn’t change the decision I’ll make. The recipe for efficiency is “all the necessary data, but no more.” In a large venue, it might be a three or four minute walk to drop a measurement mic in certain spots in the venue, so we want to be very deliberate about where our mics are going and what we’re looking to learn when we get there, to avoid unnecessary trips and minimize tuning time.
In this two-part series, we will apply the same logic to the creation of prediction models, which inherently simplify venue geometry to a series of raked planes and curved surfaces: if we only wish to spend time measuring and drawing venue architecture that will factor into our design decision, what must we include? What can we safely leave out? How can we best allocate our design time resources to the most important decisions? The answer is situation- and venue-dependent. Here we will look at two examples: one for the horizontal plane, one for the vertical.
Let’s start by looking at a design for the Gorge Amphitheater in George WA. It’s easy enough to shoot the vertical profile / landing strip elevations, but modeling the horizontal curvature can be much more time consuming, and we don’t want to hold up the rigging markout. Since we only need about 160° of horizontal coverage and our Mains + Sides can comfortably get us to the 180 line, our side hang will have a gentle and generous overlap with the mains while still making it to the necessary outer edge. Experience tells us that between 50° and 55° toe-out will do nicely, and the tour rigger knows how to hit that without elaboration, which allows us to completely avoid devoting any time to modeling horizontal curvature and focus our limited design time on the vertical front-to-back profiles, which will require some careful consideration.
We know where each hang is going and where it’s aiming so we can build profiles quickly (FIGURE 1A&B). The On-Axis Profiles mode of the On-Axis app is useful for this.
FIGURE 1A - Onaxis profiles for mains (+/- 33ft) and sides (+/- 66ft).
FIGURE 1B - Isometric view of the on-axis profile planes
The mains require some extra attention here. The red and green portions of the profile are handled solely by the main hangs. The blue portion of the listening plane starting at about 280 ft is the rear lawn area. The venue’s lawn delay system doesn’t start coverage until about 70 feet into the lawn depth, and has gaps between the towers until further back, so we need to throw some energy from the mains onto the lawn. The usual deployment on this run is 16 boxes for mains and 12 for sides, but since the mains have a much more significant throw than the sides in this particular venue, we’ll steal a pair of boxes from the side hangs, and deploy the rig as 18 in the front and 10 on the side.
Although a high trim height is available to us, flying lower with a larger up-tilt means the low-mid energy lobe will naturally travel in the correct direction, meaning the ArrayProcessing algorithm won’t need to work too hard to steer it, giving a more natural-sounding end result. Although the throw distance is quite significant, we are not asking the algorithm to redistribute energy in a dramatic way. (Figure 2)
FIGURE 2 - Side profile of the main hangs throwing some "overflow" energy onto the lawn (blue).
Center line impacts are more spaced on the blue plane, meaning a gentle HF pattern that feathers into the lawn system. This also avoids the “ice pick to the forehead” effect that we hear when we have a heap of 0° angles in the far field - all the low mid has dropped considerably, and the HF is lost to air absorption, leaving a speakerphone-like mid-bandpass response that’s both immediately recognizable and fundamentally unpleasant. Comparing the 4 kHz and 250 Hz contours confirms that we will maintain overall tonality as the level drops with distance. (Figure 3)
FIGURE 3 - The 4 kHz contour (solid line) and the 250 Hz contour (dashed line) remain parallel from front to back, indicating a stable overall tonality with distance.
Finally, subwoofers: we’re carrying flown arrays of seven per side, which is a reasonable line length for a typical arena or shed depth of around 175 - 200 feet, but not sufficient line length required to tighten the vertical pattern enough to launch the two lowest octaves all the way up the hill (see Figure 1 in Taming Red Rocks for the importance of line length in such a situation). The “default” trim for this design is the same as the main hangs (35 feet in this case), but raising the subs to 45 feet and tipping them back 5° helps reduce the LF building between FOH and the barricade by 3 dB as well as slows the drop rate up the hill (Figure 4). Plus it looks cool (Figure 5). If we need to go further, we can apply a gentle log delay taper to steer upwards (see #3 in Quick Solvers), although it didn’t prove necessary in this case. And again - the LF coverage is starting to lighten up in the firsts few rows of the pit - right on schedule, as one of the design goals on this show is to keep the stage clean of LF. The sub energy is then filled back in for those listeners with four single-18” front fill subs placed in the pit, run at a much lower SPL.
FIGURE 4 - 50 Hz landing strips from front to back, 3 dB / color.
FIGURE 5 - Two photos of the final deployment.
In this particular situation, adding the full 3D curvature and horizontal plane to the model doesn’t change anything about how I would have chosen to deploy the system, so it’s not something I wanted to devote any of my limited design time to. The OG landing-strip-style predictions can still suffice in certain situations where we don’t have any surprises or complex geometry in the horizontal plane. With experience and judgement, some or all of this information can be omitted from the modeling process if time is short. (Some of my arena models infamously lack curved corners, instead built only from inclined plane for rear and side rakes. In some situations this is enough information, in others, it isn’t. The skill comes in knowing which is which). If, for example, you weren’t sure about how far out the sides should aim to accomplish the necessary coverage width, then you should model it instead of guessing, or use a low-risk mechanical deployment (single point hang can be re-angled with a rope or tie-line, or a three-point hang using a delta plate can be adjusted easily if we don’t nail it the first time).
This design is an example of putting more design time and bandwidth into the vertical plane when the horizontal can be safely deprioritized. In the next article we’ll look at an example of the opposite situation.
