Taming Red Rocks

In the summer of 2023, I was hired to design, deploy and tune a sound system for a concert at the Red Rocks Amphitheater in Morrison, Colorado, a venue notoriously unique both for its acoustical properties and its challenging coverage requirements. Having never worked there before, I called a number of my colleagues who had mixed shows there, and asked what they considered to be the biggest challenges from an audio perspective.

I was treated to a large variety of anecdotes about various attempts throughout the years to cover the iconic 309 foot, 22° sloped seating area – custom rigging brackets, new roof trusses, flying dV-DOSC on top of V-DOSC. If this wide variety of system design attempts and iterations were to be catalogued in a coffee table book for sound engineers, it’d be a best seller. Note to future self.

Despite the variation in products and approaches used, two challenges were recurring themes: 1) getting sufficient LF and lo-mid energy up to the back of the slope, and 2) mix position is so close to the stage that it’s “off the edge” of the PA coverage, located in the center gap of light HF coverage between the main hangs. Fair enough – addressing these two issues would be my twin design goals.

The key to successfully projecting LF over long distances is simple enough, at least in theory: line length. Lengthening the line will increase the directivity of our array at low frequencies, a principle well established by Harry Olson almost a century ago, and dating back even further if we look to literature applying the same principle to optical physics.

FIGURE 1 shows a “landing strip” cut from the seating area prediction, showing the effects of increasing subwoofer line length on the front to back level drop at 63 Hz. We begin with a hang of three 18” subwoofers in a cardioid (FBF) configuration. Our prediction is showing us 3 dB / color division, which translates into a 15 dB drop in LF level from front to back.

FIGURE 1 - Front-to-back (left to right) landing strip from seating area prediction showing the effects of increasing subwoofer line length. 63 Hz, 3 dB / color

This illustrates a common limitation in large-format sound system design, and it’s not exclusive to Red Rocks: some vendors and designers specify subwoofer quantity based on the SPL that must be produced at FOH mix position, and today’s modern double-18” subwoofers are incredibly powerful beasts, such that three per side might be totally sufficient to create the desired SPL at mix position. But with a vertical line length of only about 5.5 feet (1.7 meters), it’s virtually omnidirectional in the vertical plane, and when contrasted with our much longer line array (which is in turn responsible for projecting generally shorter wavelengths), our sub array just can’t match the directivity of the full-range mains, leaving us with a system that loses sub frequencies over distance much more rapidly  than the upper 6 octaves.

Modern line arrays have no trouble projecting high frequency energy hundreds of feet (atmospheric considerations aside), but if the bottom of the spectrum isn’t keeping up, our last-row listeners are subjected to an unpleasantly bright, thin presentation, lacking impact.

Thus, our subwoofer line length needs to be increased to improve directivity and allow it to fight off the level drop over distance, remaining well-paired with our full-range sources. Increasing to six (11 ft / 3.3 m) is still not quite satisfactory, but a bump to nine elements (line length of just under 16 ft / 5 m) and tipping the entire array upwards by 5° bumps the green isobar out of business entirely.

We can also gain an additional 3 dB improvement with the introduction of some gentle delay taper (1.5 ms on the middle cardioid cell, which is elements 4,5 and 6, and 3 ms on the top three elements) and we have something quite respectable – a last row that is only 9 dB below mix position at 63 Hz, despite being over 250 feet further up the hill.

There’s nothing novel here – the line length mechanic is doing exactly what theory tells us it will. So we will specify a higher subwoofer count, and when it is pointed out that we don’t “need that many, they are plenty powerful,” we will agree, and respond that we’re trying to make a shape, not just an SPL, and we will probably end up turning down the gain a bit.

But our limitations here are practical in nature. A flown hang of nine subwoofers is heavy (we will need two-ton motors to fly it safely), we can only achieve the desired 5° of mechanical tip-back, and no more (this much tilt lands all the weight of the hang squarely on the downstage rigging point, preventing us from reclining the array further), and of course, someone far above our pay grade needs to approve the rental cost on the increased subwoofer quantity, the vendor must have sufficient quantity available, the truck space to bring it in and out, the time and labor to deploy it, and so forth. In that sense, the real-world limitations may mean the party is cancelled before it’s even started. In this particular case, the bean counters understood what I was asking to do, and agreed with the intent, so the request was granted. 

Okay, we’ve tackled the “LF not making it up to the rear” issue fairly simply (all the line length money can buy!), but the remaining challenge – the light HF coverage at mix position – requires a bit more nuanced approach.

How do we close up a center gap in HF coverage? We have three main tools: aim the loudspeakers inwards instead of pointing them straight ahead, move the Left and Right hangs closer together, or use loudspeakers with wider HF dispersion. Let’s do all three!

Luckily, the show in question has a relatively narrow stage footprint, and no upstage video element, so we have a lot more leeway than usual with our sightline considerations, allowing us to move our main points one beam further on stage than where the PA is usually hung, diminishing our center gap by 16 feet.

We can use a bit of trig (if we paid attention in high school math class) or a right triangle solver (if we didn’t) to calculate how far onstage we need to shift our downstage rigging point to add a 5° toe-in on our main hangs. Nothing in life is free – and this toe-in creates some under-covered seats on the sides in the first eight rows or so, so we will use a ground-stacked outfill array to restore coverage to the front corners of our seating area. This is, in my estimation, far better than leaving a hole at mix position. If the FOH engineer is mixing off the edge of the horn, they are hearing less HF than listeners on-axis to the PA, making the show sound bright for everyone else. Compound this with the LF loss due to a short sub line, and you have a perfect recipe for an unpleasant last-row listening experience. Restoring proper coverage to mix position should prevent this.

Finally, the vendor has four wide-dispersion elements available, so we’ll place those at the proper positions in the array to cover mix position, two per side. The wrinkle is that they want to pin the arrays to the proper angles at the shop the day before, so they’re loaded onto the truck ready to fly, so we need to figure out exactly where those elements need to go in advance.

I used a Google Earth satellite photo to measure the lateral distance from the downstage edge to mix position, and then superimposed a vector of the same length onto the prediction, indicating that elements 12 and 13 in the main hang should receive the wide elements. Obviously, this approach is somewhat lacking in precision, so the first thing we will do once we arrive on site is to take measurements and confirm that the model matches reality, and that our predicted array geometry is correct.  

It can be a bit of a tightrope walk to go into a new venue and do things the house crew hasn’t seen done before. When the head rigger, who has hung hundreds of PA’s off the same roof, asks you if you’re really sure you want to toe the PA inwards, because a lot of people choose to do the opposite…you can’t help but wonder if you’re about to drive straight into a pothole that everyone else saw coming and avoided. Do you really trust your design, prediction, and math skills enough to override the concerns of the technical minds who have an intimate knowledge of how to stage successful events in the space? Am I about to do something that everyone who works here knows not to do? This is probably the most challenging moment in the entire process, when we need to make a decision and be saddled with the consequences of it.

In this instance, our trust in the design process was not misplaced, and front to back measurements of one side of the PA (Main & Sub) reveal that we’ve met our level variance goals, showing a drop of about 5 dB from mix position to about 200 feet up the hill (FIGURE 2, top). Normalizing the traces shows that our full-range tonality is intact all the way down into the sub range (FIGURE 2, bottom).

FIGURE 2 - Front to back magnitude traces @ approximately 50’ (FOH / green), 120’ (orange) and 200’ (pink).

Bottom: normalized

(Interestingly, I observed about 2 – 3 dB less LF loss in the last rows of the venue than the prediction indicated, and ended up slightly sub-heavy in the rear, leading me to reduce the electronic beamsteering a bit on the sub array. My working theory is that this is because the PA is not radiating into free space, but rather has the boundary of the rigid seating surface as well as the rock walls on both sides, which form three walls of a rudimentary “horn” to funnel the energy up the hill and slow the inverse-square loss rate. Although acoustically simulating this is a challenging task, it’s something I would be interested to study in the future.)

Our onstage hang points, toe-in and wide elements have properly combined to ensure that mix position is in fact properly covered, and once we deploy our deck-stacked outfills, we have restored our horizontal uniformity as well. It’s a bit of exercise to climb all the way to the last row, but it’s worth it to sit down, catch our breath, and feel the impact of the low frequencies coming from 300 feet down the hill.

The 2023 deployment of mains and subs.

Outfills, 2023 edition.

Anyone familiar with my design approach will know that I never pass up a chance to spin up another design iteration if I feel there’s still potential for improvement, so when the same artist asked me to return to Red Rocks and handle their show this year, I gladly accepted.

We focused our improvements on four main points: first, the vendor offered a 17th box per side for our main hangs – effectively “the same geometry but with one more box at the top” to better cover the ADA seating plateau that is elevated above the last row at the rear of the venue. This proved a bit of a challenge mechanically, as this requires more uptilt on the rigging frame, and virtually all the weight of the array was already on the front point. By massaging splay angles a bit and adding a bit more curvature to the bottom of the array, the bottom boxes act almost as a counterweight for the top ones to pull the center of gravity back under the rigging frame. This gave me about 7° of over-coverage past the front row, so I had planned on simply muting the bottom box. However, the “pit” area ended up being filled with an additional row of removeable and ADA seating, a happy accident, meaning the bottom box landed right on target.

2024 mains deployment, stage right. Note the toe-in on the mains, and the yellow beam clamps where the PA is usually hung further offstage.

Second, the vendor had an increased inventory of wide-dispersion elements available, allowing us to make the bottom 6 elements in each array all wide, which extended coverage further into the under-covered corners, reducing the amount of territory that needed to be covered by the outfills.

Third, we used a larger-format element for our ground-stacked outfills. They were being driven heartily the previous year, and this year I wanted to increase the splay angles at the top of the stack to create a softer HF edge and help “feather” the outfill coverage into its seam where the mains take over. A more powerful loudspeaker with a larger horn gave us the necessary output to do this with ease, and created a better tonal match with the large-format mains.

Finally, the vendor had upgraded their subwoofer inventory to a newer, single-21” model. I still had nine per side to work with, but the larger driver meant a taller cabinet height, which yielded about a 15% increase in line length, and a corresponding improvement in vertical directivity and therefore front to back consistency (see the bottom-most bar in FIGURE 1). Eagle-eyed readers will note that the LF is now starting to decrease in the front few rows, making front fill subwoofers important. Accordingly, our deck-stacked outfill arrays were supported by two-per-side double-18” subwoofers in gradient configuration, restoring some impact to the front rows while maintaining a quiet stage – a priority for this artist.

Mains and Subs, 2024. Note the redistribution of curvature compared to the previous year.

I will readily admit that the toed-in mains can look a bit “wrong”, particularly when viewed from off-center in the curved seating area, but I am firmly in the “It ain’t stupid if it works” camp on this one, and already kicking around some ideas for my next incremental revision next year if I am fortunate enough to be invited back once again.

Showtime.

The author wishes to thank Production Services International and FOH mixer Peter Costello for enabling his mad scientist inclinations.