March Madness 2021: Broadcast Audio Mixes in Crowd Sound — Real and Augmented

Short turnaround, venues’ varied capabilities complicate the effort

March Madness is off to a noisy start, thanks to a combination of fans finally allowed in the stands and augmented crowd sounds.

Aboard NEP Supershooter 9 mobile production unit, A1 Dave Grundtvig is mixing games from Lucas Oil Stadium, whose football field has been divided at the 50-yard line into two courts. He takes in a feed from the Firehouse Productions crowd-sound system, which is also fed into the venue PA system and mixed in with the actual crowd noise, from the 6,000 or so spectators allowed in, through the six ambience microphones Grundtvig has deployed.

At Lucas Oil Stadium’s March Madness venue, a feed from the Firehouse Productions crowd-sound system is fed into the venue PA system and mixed with actual crowd noise.

“The PA hasn’t been overpowering, and we’re getting good reaction from the fans,” he notes. “Overall, it sounds very natural.”

The crowd-sound content used for the NCAA games is significantly different from that heard during NBA play.

“Early in the NCAA project, which we started in December, we realized the crowd sounds that we had created for the NBA were of the wrong size and emotional intensity,” says Sonic Designs Audio Designer Dan Gerhard, who with his partner, Associate Sound Designer Ellen Fitton, created the content for both the NBA bubble and March Madness. “So we started over and created crowd sounds specifically for the NCAA.”

Custom crowd content was necessary for the NCAA series for several reasons, he explains, most notably because college-game audience members are younger and tend to have family members involved, which he says affects how the crowd reacts and sounds.

“The nature of the tournament is also much more intense than a typical NBA game, and there is no ‘home team’ fan element within the NCAA bubble,” he says, adding, “And with the exception of Lucas Oil Stadium, the arenas are generally smaller than an NBA arena. This necessitated the use of younger, more enthusiastic voices [from] a smaller, more intimate crowd in overall size, which would be better identified with the NCAA and therefore more realistic. There is no overlap of material between NCAA and NBA.”

What’s missing from the mix, notes Grundtvig, are the cheerleaders and school bands, which no artificial sound system has managed to simulate yet. “It sounds good, but it could never sound like it does when the stands are full and the bands are playing. But we’re still able to get the sound on-air well-balanced.”

A Variation on Augmented Crowds

The augmented crowd sound will vary from venue to venue, mainly because not every venue has a dedicated sound system for the artificial audio. The variability is largely due to the short notice on March Madness this year.

“Our entire rehearsal period for three days before the first game sounded like there was an arena full of people, and that is a strange feeling,” observes Mark Dittmar, VP, Firehouse Productions, which is providing the games’ crowd-sound PA systems and its mixers — they prefer the term sweeteners — who fire preset crowd-effects audio cues instead of riding faders. “The front-of-house mixer loved it because he could find the pocket he needed to be in right off the bat, instead of adjusting things. All those adjustments you have to make as the game goes on, he was already there; he knew exactly what the entire room was going to sound like.”

As a result, Dittmar adds, the venue sound was very consistent from rehearsal to air, thanks to what he calls a virtual aircheck.

Audio Is a Team Effort

Joe Carpenter worked aboard Game Creek Columbia at Butler University’s Hinkle Fieldhouse — site of both the 1986 film Hoosiers and the story it was based on — for first-round games last weekend. One issue he faced was the ambience of the venue, whose ceiling acted like a parabolic reflector, exacerbating the room’s innate boominess. That was largely mitigated with the addition of acoustical treatments. The interaction between the room, the house PA system, and the augmented-crowd-sound speaker arrays presented the same challenges that Carpenter’s colleagues on the broadcast encountered.

At Butler University’s Hinkle Fieldhouse, the ceiling acted like a parabolic reflector, intensifying the venue’s boominess and requiring acoustical mitigation.

“The NCAA wants it to be loud and exciting, and we’re trying to do a television show,” Carpenter points out. “It’s the age-old battle. It’s not necessarily a conflict; I understand both sides of it, and we always find a way to make it work.

“It’s a team effort,” he continues. “It’s just that there’s a lot of elements, and a lot of them are in different places than usual, like where they seat the fans, which is in the end zones by the baskets here. I had to move my crowd mics three times. Nothing’s ideal.”

Another challenge created by COVID configurations is to match A1s with different production and announcer teams for each game, rather than creating a unit to serve an entire bracket. One way that is manifested is the often significant disparity in natural voice levels between announcers.

“For instance,” Carpenter explains, “the other day I went from Carter Blackburn, who projects really loudly and very directly, and Deb Antonelli, who surprised me by how directly she also projects, to Ian Eagle and Grant Hill, who naturally speak at a lot lower level. I found I was literally putting 12 dB more input gain in my play-by-play headsets and about 8 dB more in the color headsets. All of that high-mid from the room and real and artificial fans increased everything by 20 dB in my center channel. I don’t think I’ve ever bounced back and forth so much between 5.1 and stereo in any event. I was struggling to create lanes to put my effects in there.”

The Building ‘Sings’

Phil Adler served as A1 at Purdue University’s Mackey Arena, where the First Four games were played. Mixing aboard Game Creek Celtic, he had 24 channels of cabling in the venue, but some had to be devoted to other audio — intercoms, for instance — limiting to 10 the number of ambience microphones he was able to deploy, which in turn limited the sound-effects field for the surround-sound broadcast.

Combining real and artificial crowd sound in the broadcast mix overcame the inherent resonance of Purdue University’s Mackey Arena.

“It’s all about what the whole stereo spread looks like,” he explains. “What I was doing was to make the crowd the widest thing in the mix and then place the court in between that. [Although] the court fits inside the crowd and the crowd is a widest thing, the crowd wasn’t terribly wide. I did try to make the crowd sound wider using the stereo spread [within the 5.1 mix], with more of the crowd in the rear channels.”

Adler notes that audio crews also have been working under some adverse conditions, including leaking roofs and cabling shortages between trucks and venues. The sonic variety of the venues can be heard in the diversity of their ambiences: the paucity of fans means that some of the older buildings’ character can show through.

“My building had serious ‘singing’ at about 230 Hz,” he notes, referring to the venue’s inherent resonance, caused by a combination of its acoustical nature and its mechanical systems. “That’s the resonant frequency of the building’s fans, air conditioning, air handling — its natural noise. It’s just a general hum. I had to do some serious [EQ] notching just to get rid of it and then work around that. We had about 1,500 people in there, but that’s just not enough, spread out like they are, to overcome the building noise in a venue that holds more than 14,000.”

The combination of the real and artificial crowd sound in the broadcast mix ended up being fine, Adler says, but only after he “battled” with the volume level of the venue PA system, which was also playing the augmented-crowd sounds.

“They were also feeding that into the house, which is fine,” he explains, “because my crowd mix would pick up the same crowd bouncing off the walls, which can make it sound more natural. We just had a problem with [the augmented crowd sound] being too loud at first. Once we got it dialed in, I frankly couldn’t tell you at one point who was real and who wasn’t.”


Password must contain the following:

A lowercase letter

A capital (uppercase) letter

A number

Minimum 8 characters