Salsa Sound Sees Virtual Crowds as the Right Sounds at the Right Time

Audio is derived from authentic sources and can match nuances of the game

Despite concerns about “inauthentic” sounds of games and matches by A1s and others on the technical side of sports broadcasting, it appears that “faux” crowd sound is becoming part of the sports-broadcast experience as various leagues and teams come out of pandemic-induced hibernation. European soccer leagues, such as the Bundesliga and the English Premier League (EPL), have used preprogrammed crowd sounds for televised games, as has CBS Sports for the recent NWSL Championship.

Salsa Sound, a three-year-old start-up germinated in the well-regarded audio program at the UK’s Salford University, is at the forefront of that trend, initially positioned by its automixing software aimed at broadcast matches. That led to development of the company’s fan-experience software, called Front Row, which offered augmented audio, such as chants and effects, customized for specific teams.

Salsa Sound’s Rob Oldfield: “I think virtual crowd sound is quickly going to become part of the picture.”

Both the technology behind those services and the sizable audio library of crowd sounds built up for Front Row serendipitously positioned Salsa Sound for sports’ slow return from lockdown, with games played and races run without fans in the stands. vCrowd, the product of that, offers broadcasters dynamic crowd sound that can be nuanced to the action of a particular match and using actual audio elements sampled from prerecorded broadcasts of any pair of teams. Its first application was on a Manchester City–Real Madrid match in June, and its U.S. debut was later that month for the NWSL Challenge Cup broadcast on CBS.

Since then, Salsa Sound’s virtual-crowd-sound system has been vetted by a number of leagues and broadcasters. Although no firm announcements have been made yet, Salsa Sound founder/CEO Rob Oldfield indicates that both basketball and football are interested.

“We’ve had prototypes out to several leagues and broadcasters, and the reactions have been quite positive,” he says, noting that any resistance to the idea from the professional side has been overwhelmed by the way fans and viewers seem to need and welcome the sense of normalcy that virtual crowd sounds offer. “We listened into webinars to what engineers and producers were saying about ‘inauthentic’ crowd sounds. But, at the end of the day, [sports] is about storytelling, and the average Joe doesn’t want to be distracted by the lack of crowd sound. Instead, they want to be sucked into the game by it.”

Authentic Sources
Salsa Sound draws on a substantial library of crowd sounds sampled from a variety of UK EPL stadiums and teams. Oldfield says it’s important that the sounds for a particular match have originated at the same stadium from which the match is broadcast. These are recorded using Salsa Sound’s own sampling engine. For the NWSL broadcast, the sounds were derived from clean-feed broadcast NWSL archives transmitted by CBS Sports, from which Salsa Sound cut samples and loops.

Critical to the implementation of these samples, however, is the ability to mix and ride them according the ebb and flow of a game. This is done on a track pad that can not only swell the volume of a crowd as needed but also can change its fabric, using what Oldfield calls “an excitement algorithm,” which can increase the velocity and feel of a crowd reacting to a goal or a bad call.

“You can’t use the same crowd reaction to a goal as to a near miss,” he explains. “There needs to be ways to realistically portray emotions like outrage and joy. And it has to be made up of actual crowds for a particular match. You can’t use sounds from a Manchester United game for a Phoenix Suns game.”

Europe vs. U.S.
Oldfield notes that European broadcasters and leagues were much quicker to implement virtual crowd audio than their U.S. counterparts. But now that it has been used for several events here, he expects the uptake to gain momentum in coming weeks, as leagues like the NBA and NHL resume play and the NFL readies its 2020 season, almost all of which will be without crowds in the stands — at least initially.

“We’ve been watching how soccer [broadcasters] in Europe have been handling it, and I think some people, including me, might have been a little skeptical at first that it would sound hokey or cheesy or just be a distraction,” Steve Karasik, VP, remote production, CBS Sports, said ahead of the NWSL Challenge Cup. “But, after listening and watching those European soccer matches, we felt that it really did add something and made for a normal viewing experience for the viewer watching at home.”

Surprisingly, Oldfield has seen little credible competition for a service like Salsa Sounds. Besides Sky Deutschland’s Sky Sound Atmo system and alliances between broadcasters and some sports-videogame producers sharing audio elements, it’s a wide-open field. That may be because of the widespread hope that the root of demand for it — the pandemic — will be short-lived.

However, says Oldfield, the concept of virtual crowd sound and the technologies behind it may well find traction in broadcast and live sports coverage even after the pandemic has been tamed.

“On television,” he explains, “vCrowd gives broadcasters a tool with which they can weed out elements like racist chants or abuse aimed at a particular player, weaving it in and keeping the crowd ambience consistent. It can also be used to pump [partisan] crowd sounds into a stadium live,” he adds, for practice sessions or for away games to cheer the visiting team.

For now, though, Oldfield believes that broadcasters and teams alike are happy to have something that helps restore a sense of normalcy to games.

“As long as the sound matches the nuances of the game and of the teams and the venues,” he says, “I think virtual crowd sound is quickly going to become part of the picture. I say that not only as the CEO of Salsa Sound but as a fan, too.”

Password must contain the following:

A lowercase letter

A capital (uppercase) letter

A number

Minimum 8 characters