Hello again. Over the years I’ve worked on several films and TV shows that required editing crowds for different sorts of sports, both indoors and outdoors, concerts, and, my personal favorite, panic, battles, and riots. Umm, just kidding there. Mostly. Crowds are actually a bit more difficult to cut than they might seem on the surface. There are two obvious problems to be aware of in cutting crowds for just about any event and these are: getting the right flavor of material and keeping them from summing into an insurmountable wall of noise.
There are basically three possible places that crowd recordings usually come from. The first, and really the best, is if the production mixer can get good, clean, stereo production or wild track recordings on the set. Unfortunately, these are often either short takes, recorded in mono, or have some sort of set-based intrusive sound in them such as wind machines, production vehicles, or Assistant Director vocals. In the event they do exist and don’t fall into one of the traps I just mentioned, you are starting off from a great position. Tracks recorded on the set have the obvious advantages of correct size of crowd ( two people to two million ), correct makeup ( men or women, domestic or foreign speaking, etc. ), and correct character ( excited, concern, anger, panic ).
The second source of crowd material is the Group ADR stage. These can work well if your group Supervisor is really good and paying attention to the specific requirements of the project. The concern here is that most ADR groups consist of eight to twelve people so getting material that represents larger gatherings can be a little problematic. It’s best not to rely on Group ADR for the weight of masses but it is particularly useful in creating a close-up layer of specific voices that will play in the mix over and above the cheering or screaming millions that will come from crowds cut in sound effects. Library sound effects is, of course, the third source for preparing crowd material.
For effects editors, the gates are pretty much wide open to come up with any type crowd, in any size, and in any static or changing character that the scene requires. It’s not all drudgery, though. We also get to do fun stuff like crowds heard through the ears of an outer space visitor or maybe by a Earthling who is drunk. Also, scenes are often presented in slow motion for one reason or another and that’s a whole area unto itself in designing crowds.
While all crowds are basically a see-it, hear-it type of sound, there are some differences in cutting them that would be useful to discuss. I’ll start with concert or pageant crowds. In “Miss Congeniality” (2000), there are various sequences which require layers of laughter, applause, and reactions. Pageants, concerts, and other similar audience type events are mostly reactionary to whatever happens on stage. In addition to the basic elements of steadies and reactions, I’ll add cheers and whoops to bring up the energy level in reaction to a specific event or song.
In “Hannah Montana: The Movie” (2009), directed by Peter Chelsom, there are several concerts which have endless armies of screaming young girls. I had to cut elements for the crowds that were paying attention to what Hannah was singing on stage but I also had to cut tracks of shrill, squeaking, screaming girls! They didn’t seem to react to anything in particular, they just screamed and screamed and screamed . . . . Easy enough to cut once I thought to turn my monitors down by about 15db! I’d like to mention that, while in no professional regard whatsoever, working back-to-back on “Hannah Montana: The Movie” with Miley Cyrus and “17 Again” with Zac Efron, did get me some significant credibility with my teenage nieces and nephews!
I usually find myself creating a structure of three general levels of crowds for almost any type of scene actually and they are: base crowds or steadies, also known as idles, specific but still large-group-based reactions, cheers, etc., and, finally, a close-up layer of single voice specifics such as shouts or laughs. The base crowds play throughout the entire sequence as needed and form a support level for all the other elements to sit on. I like to cut these as five-channel ( L-C-R-Ls-Rs ), tracks in order to fill out the theater or room the event takes place in. Cutting a specific center mono track can also help the dialog mixer cover angle or mic changes in the production dialog. These tracks don’t have any specific movement or recognizable vocals to them which might be problematic in terms or preparing the M&E version for delivery. They are intended to deliver size and weight of the crowd but nothing necessarily close-up or sync relevant.
The second layer that I edit is a crowd-level reaction pass. These do have pan and sync characteristics and by their nature, often have vocal specifics to be aware of, again, for the M&E. Finally, the third layer is played for on-camera, close-up shots of the audience and these are the single-person shouts or laughs that play right up against whatever the speaking actors are doing in that particular shot. I find this type of layering works for just about any type of crowd scene from small recitals to huge planetary doom scenes.
Sports films are a great opportunity for crowd scenes and one that I worked on was “Rollerball” (2002). This film had a good combination of raging sports crowds for the Arena matches and also a few interesting street riot scenes. The games had the usual mix of base crowds and reactions to players getting pulverized on the track, but it included another great element: chanting and footstomps. We start one of the early contests below the arena in an access corridor. From above, we can hear the crowd is already getting into the event and I cut both chanting crowds and bleacher foot stomps in sync with each other which built in both pace and loudness as we approached the end of the tunnel and crescendoed as we move out on the track with the players to start their warmup laps. Here, I did get to cheat in some inappropriate chants as we didn’t have wild tracks from the set but it worked ok because they were non-identifiable vocally and pretty much got covered up by the other elements in the scene. They ended up being felt more than heard but along with the footstomps, made an intriguing pulse-like element in the soundtrack.
As I mentioned above, I also worked on “17 again” (2009), and the film had several high school basketball games. There was one game in particular in which Zac’s character starts to go into a bit of a trance as he finds out a critical bit of personal information and the whole room slows down and begins to time-warp. I took the crowds of yelling spectators and pitched them in sync with the slowing camera pov while simultaneously changing their character from happy high school kids to slowed-down panic screams and wails to play up the surreal tone. In the regular backgrounds, I also cut in some underwater tones designed to supplement the pitch-slowed crowds and play up the claustrophobic, near-drowning, feel of the scene.
For some reason, war movies and riot scenes always seem to skip the nice gentle ramping up of the crowd’s energy and just go straight to the screaming and panicking. For “Windtalkers” (2002), John Woo’s film about Navajo code talkers in World War 2, I edited some of the Japanese attack sequences and was lucky enough to have a few of those rare wild tracks I mentioned in the beginning of this article. We had great attack yells of large groups of soldiers and, crucially for the sequence, big group “Banzai’s” and other attack screams. This one is a good example of crowds having very specific natures that have to be addressed in order to be believable by the audience. There was only a small layer of ongoing attack vocals that I could use because most of the Japanese soldiers were seen running and shooting, not yelling as they ran forward. Further, I couldn’t use any other language crowd because they wouldn’t be yelling anything in any other language but Japanese. Unlike the “Rollerball” chants, I couldn’t cheat in attack yells of African Tribesman or Roman Legionaires! “Forward, Claudius! Put them to the sword!” simply would not have worked here at all . . . . For the American soldiers, I had a little more leeway though. I actually had access to some really terrific death yells and run-by screams in the library that had been recorded for “Glory” (1989). Mid-19th century yells from the Civil War somehow seemed right at home in the mid-Twentieth. The nature of these battle scenes relied much less on beds of yelling soldiers and more on single person grunts and screams but it was still a layering of sounds much like the concert material.
I seem to work on lots of films with big crowd panic scenes. “Snakes on a Plane” (2006), “Poseidon” (2006), and “The Final Destination” (2009), all feature big crowd death and destruction scenes. They are all pretty much just what you’d expect: big in-your-face screaming crowds. I wanted to be a little different in each case though so I started coming up with other types of crowds that I could mix in with the up-front screams. In order to ramp up the energy, I started to add in things like angry men from simple arguments on up to boxing crowd shouts and yells. I found a recording of rodeo cowboys yelling back and forth. Not what you might expect at a Nascar race but there they were. One track was of a group of girls laughing that I pitched down for “Snakes on a Plane” to give the moment a more demonic feel. They weren’t distinguishable in the mix, but it added just a bit of a creepy note. In the end, I scoured the library at Soundelux in Hollywood, where I worked on these three films, to come up with any sort of angry, shouting, spikey vocals that I could layer over the top of the more constant scream-based tracks. Thankfully, the library there is extensive to say the least and I found all sorts of odd little bits that livened up the crowds as they met their horrible, inevitable ends!
One crowd that I specifically want to mention occurs in “The Great Buck Howard” (2008). This is actually sort of an “anti-crowd” event. As the grand finale’ at each performance, Buck Howard, a personification of the real-life mentalist, Kreskin, played here by John Malkovich, is required to locate his paycheck for that show by reading the audience’s collective mind to determine its hidden location. We see the trick performed successfully a couple of times in the film and are left uncertain of whether or not this is a well organized charade or if he is truly locating his check by thought waves.
The moment I want to point out comes one evening at a performance in front of a huge Las Vegas showroom audience. Buck, who has become progressively more anxious and unstable, is not able to determine the check’s location and the crowd, which had been laughing and murmuring along with him up to that point, finally falls silent as they slowly tumble to the fact that Buck has lost any ability he may actually have had in terms of his mental gifts. As Buck lurches from table to table, desperately trying to “read” anyone in the audience, the crowd in this huge room just sits there, silently. The absence of any sound from the audience at this point just becomes heartbreaking as Buck continues to flounder hopelessly and helplessly, alone in a room full of people.
Written by Bruce Tanis for Designing Sound.