The Front Row Experience — Precision Audio Services

I recently overheard a concertgoer, waiting in line outside a venue, tell his friends that he loved to be in the front row, right up against the barricade, even though the sound is typically not as good, but that it was worth it to be closer to the artist. While priorities may be different for each concertgoer, I couldn’t help but think that perhaps this isn’t a tradeoff that any concert attendee should have to live with, especially given what a front-row ticket costs these days. Should the biggest fans of the artist, who waited outside for hours to get as close as possible, be expected to settle for a lousy sonic experience as a given? Is there anything we as systems designers can be doing to improve the situation for the first few rows?

After experimenting with various approaches to this issue, here are a few design concepts that I have found to consistently improve the front row listening experience.

Turtles All The Way Down

Whenever possible, I try to achieve enough curvature on the flown mains to cover all the way down to the barricade or first row - that is, the center line of the bottom box hitting at listener height for the audience member closest to the stage. Although this is not always practical - those first few linear feet of coverage disproportionately consume splay resources, an idea that we will be examining more closely in a future article - it helps a lot when it is possible. I find that having the mains cover all the way down to the front gives a much more cohesive feel to the coverage, as there is no coverage “hand off” between sources when walking from the last row to the first.

This usually means more curvature of the array, and the accompanying increase in vertical splay may not be possible if box count is limited, and also shifts the center of gravity of the array towards the rear, which might present mechanical complications as well - see Center Yourself for more.

This can be hard to see in isometric view (Figure 1, left) so I use a section view of the array to inspect the center line impact of the bottom cabinet with relationship to the first row (Figure 1, right).

FIGURE 1 - Getting an accurate view of the coverage at the bottom of the array can be difficult in isometric / 3D view (left). - I use a Section view of the array to inspect the bottom cabinet’s impact point in relation to a listener standing at the barricade (right).

Asking A Lot

Front fills aren’t magical - they can’t penetrate through human bodies, so we have to be realistic about what our front fills can reasonably be expected to accomplish. With a stage height between 4.5 and 6 feet or so, they may be good for about 2 rows of a standing-height audience. They can’t reasonably be expected to cover the first ten rows under normal circumstances, and if the stage is lower - say, below 4 feet - their effectiveness is almost certainly limited to the single row of bodies along the front.

The devil is in the details here - when walking the coverage of the system in an empty venue, with nothing to block the front fills, they may seem to perfectly cover the first 6 rows before your mains system picks up the coverage baton on row 7. With human listeners packed into the space, this is no longer the case, and crowded pit sections are among the most difficult and inconvenient for a systems designer to walk through and evaluate during the event, so the problem often goes unnoticed during show even though it was fine during sound check.

This is not to say that front fills are “overrated” or unnecessary. Even in situations where the geometry of the mains system is sufficient to cover all the way across the front row of the listening area with no gap, I still find the listening experience to be somewhat “weird” (that’s the technical term) when a front-row listener is watching an artist perform directly in front of them at eye level, while localizing the audio coverage to the array twenty feet up and to the left. Tolerance for this type of “image distortion” seems to vary by listener and, to some extent, context and environment, but in all cases the resulting listening experience is far more natural and intelligible when a front fill is provided to drop the localization to a front-and-center position that more closely aligns with what’s happening on the stage.
Years ago, I had broached this topic with an audio vendor who didn’t want to deploy front fills because the prediction showed gapless coverage from the flown mains all the way across a theater’s front row. After the show, a man who had been sitting in the front row remarked to the vendor that he had a strange listening experience. He could perfectly hear the show, and understood all the lyrics, but still found it to be an uncomfortable listening experience for a reason he could not describe. Deploying front fills rectified the solution for the show the following night.

Swishy Swishy

Last year, a FOH engineer I work with remarked to me that he had found a YouTube video filmed by a fan in the front row of one of our shows. The fan had replaced their camera phone’s audio with a board recording of the show that had been published on the internet by the artist. The FOH engineer said “you can tell they’re using the board mix because the organ is in stereo. If this were phone audio, the organ would be mono.” To which I replied, “well… It doesn’t have to be!”

I grabbed an extra helping of jumpers from a cable trunk and quickly required the front fills in alternating pairs - 1/3/5/7 receiving the Left side of the mix, and 2/4/6/8 receiving the Right side of the mix, rather than the traditional mono sum feed to all the fills. This produced two immediately noticeable benefits: First, the listeners down at the barricade were now treated to the same semblance of width perception as many listeners further back in the venue, as the show in question used stereo mic arrangements for the guitars and keyboard elements. Second, and of even more interested to me, was that the “swishy swishy” combing interaction (that’s the technical term) that typically occurs on the seam between two front fills was reduced by about 90%, due to the fact that adjacent fills were reproducing only semi-correlated rather than fully correlated signals. This approach quickly became integrated into our daily system design.

Note that this is the same benefit afforded by supplying swapped drive signals to side hangs (Left side hang receives Mix R, and Right side hang receives Mix L). The common explanation is that this creates some amount of width perception for listeners sitting in the overlapping coverage area of both a Main hang and a Side hang. Although this is true, the effect is small because the angular displacement between the two hangs is minimal for those listening positions, so there’s not much “width’ to be had. However, the benefit compounds in the form of the reduced combing interaction between the two sources. Both effects combine to create a more natural listening experience through that coverage seam.

Our PA tech pointed out that we should swap the drive signals to Right on 1/3/5/7 and Left on 2/4/6/8, so that listeners in the center position (between fills 4 and 5) would experience the same orientation of the stereo image as the mix engineer further back in the venue. Of course, “stereo” is a very slippery word in the context of live sound reinforcement, and I find myself preferring the description “perception of width” to “stereo image,” but whatever you choose to call it, the result sounded more natural. It might be more appropriate to think of this as “two mono mixes” rather than “stereo” in the truest sense.

FIGURE 2 below shows the alternating front fill strategy in prediction, represented by alternating colors. The inset shows the input routing for the front fill amplifiers, with D1/D2 carrying Digital Mix L/R, and A3/A4 carrying Analog Mix L/R. (Digital and analog inputs are open simultaneously due to the redundancy strategy, see Thoughts on Redundancy and Failover for more.)

This example shows a typical 60’ by 40’ arena stage with a 7’ barricade depth. I used my front fill lateral spacing solver, available on the Resources page, to arrive at the proper fill spacing.

*FIGURE 2 - Alternating front fill drive strategy in MAPP 3D indicated by red fills for Right and blue fills for Left. The inset shows input signal routing in R1.*

Note: I tend to eschew the common practice of timing each individual front fill back to the mains. For the reasons just discussed, the front fills nearest to the center of the stage don’t share a seam with the flown mains when the venue is full and if they don’t share a seam, it makes even talking about their timing relationship rather nebulous in my opinion. What they do share a seam with is their neighboring front fills. If all the front fills are individually timed back to the mains, it means higher delay times towards the center of the stage, which, for any listeners listening to more than one front fill, pulls their localization away from the artist on stage, which is distinctly unhelpful and contributes to the “weird.” With a flat fronted stage, I typically let all the front fills ride together on timing, and set the timing relationship between the bottom box of the flow mains, and the outermost front fill (aka - the one that has the most substantial overlap with the flown mains). Substantial overlap in coverage means interaction, and that interaction is managed with timing. I typically make the fills about 2ms early at that point, which leans to a smooth imaging towards the stage with no weird gaps or seams as one walks through that area, but your mileage may vary based on the program material and the design geometry.

Stuck In the Middle With You

Even with the most fabulous-sounding front fills in the world, we would still be left with an undercoverage problem four rows back once the human meatbags paying customers fill the pit. Front fills might cover the first few rows, but beyond the reach of the front fills we have a hole created by the coverage gap between the flown Mains. It typically forms the shape of an inverted triangle, a.k.a the Triangle of Sadness (that’s the technical term) (Figure 3).

*Figure 3 - The coverage gap between the flown mains and beyond the reach of the front fills.*

Again, this gap might not be easy to spot when walking an empty venue with un-occluded front fills, but during the show, the absence of coverage here is definitely painfully obvious. Using wide-dispersion elements at the bottom of your flown arrays can reduce the size of the gap, as can toeing in the mains horizontally when practical to do so, but they don’t usually eliminate the issue completely, and it is exacerbated by both lower trims and higher horizontal displacement (wider stages). Luckily this is easy to spot in prediction if the frontfills are muted (the computer doesn’t know they’ll be blocked by humans in actual use) - Figure 4 below shows some real-world examples of the Triangle of Sadness in an arena (left) and a theater (right). Unluckily there is sometimes not much we can do about it.

*FIGURE 4 - The triangle of sadness is visible in HF prediction of Mains with the front fills muted. Arena (left) and theater (right)*

The solution is a fill speaker. One approach is to add small point source elements, flown just inside the main arrays, aimed into the center gap. Although this can work, it is not without drawbacks (it is difficult to match the dispersion of the loudspeaker to the shape of the coverage gap, and turning the loudspeakers in towards the center enough to adequately fills the gap unavoidable bleeds some HF energy onto the downstage center position. Drastically reducing the gain before feedback of our entire system for the benefit of coverage of a few dozen people is probably not a good tradeoff. It’s also less desirable in terms of sound quality in my opinion because it results in two loudspeakers aimed directly through each other at a shallow angle, which is never a fruitful approach if our goal is uniformity and smooth-sounding seams.

A better solution is a flown downfill, suspended above the coverage gap and pointing steeply downwards. This is a much easier coverage shape to make with a point source element, or even better, a few cabinets of a small-format array product with a tight vertical pattern. The coverage should have a harder cutoff edge down near the front of the floor at the barricade, and a softer fall off edge further back on the floor as it feathers gently into the coverage of the mains.

If only there were some sort of truss that was already hung in that position, to support lighting fixtures or something of that nature, that we could attach a small fill to. Logistically, this is rarely going to be a solution that is feasible and agreeable by all departments, but when the opportunity is available, we should take it. (Lighting designers who have background in the theater world are pretty used to seeing center clusters, although their role in a typical musical theater sound design is a bit different. Rock and Roll LDs are more likely to consider a center cluster an elusive beast.)

Does it work? Yup. (Compare Figure 5, below, to Figure 4, above.)

*FIGURE 5 - The triangle of sadness is eliminated by addition of a small flown centerfill. Compare to FIGURE 4.*

The examples shown here use a three-box array designed “upside down:” with the smaller splay angle at the bottom, to give a harder HF edge near the barricade, and a larger splay angle at the top, to give a softer HF edge to feather back into the mains coverage. This produces a cleaner outcome than the flown infill approach, leads to less bleed back onto the stage, and is easier to time - timing is set on the floor, halfway between the mains and the center fill, just a tiny bit late, and then timing the front fills as described above, which makes them early in relation to both sources.

The properly timed centerfill is perhaps the best example of “you don’t hear it when it’s working properly” as walking through the pit shouldn’t shift localization in any substantial way, but simply never leave you standing in a coverage gap.

Next time you’re in a position to do so, put on some program material you’re familiar with and listen for the Triangle of Sadness. Is there one? How big is it? Where are the edges? This is a good exercise to get in the habit of recognizing one of the most common coverage deficiencies in typical designs, which is the first step towards fixing them. (Figure 6) 😊