With 44 million TV viewers and a 10,000-strong live audience, the Grammys are one show that you really don't want to screw up...
The problem is that these tracks get mixed in someone's home on a pair of nearfield speakers for a totally different medium.” Front-of-house mixer Ron Reaves motions towards the four strings of 12 JBL Vertec 4889 line-array boxes, steeply raked to cover the bowl of the Staples Center arena, as well as the seats in the upper balconies. "The [backing] tracks are not exactly optimised for playback for 10,000 people over a huge PA system. When you translate them to 90 15-inch speakers, I have to do a fair amount of EQ to make them sound the way they were intended to. When you mix on a 10-inch speaker, you're going to get way too much low end on the tracks. They typically hype the low end, so it sounds good on a television speaker but not through an 18-inch sub. Add that to what's coming off the walls and I'm fighting it in the 160Hz to 300Hz range. I often high-pass those tracks because they have too much low end for what we're doing.
"But on the other hand,” Reaves adds, "that kind of extra bass is welcomed by the guys in the truck.” This is a reference to the pair of Music Mix Mobile (M3) outside broadcast vans parked at the top of the Staples load-in ramp, where mix engineers Eric Schilling and John Harris trade mixes of the rehearsals back and forth in a MADI-enabled workflow that lets them give each of the 22 artists time to get their acts together over three days. They will then have to do it one more time, flawlessly, inside of three hours, for nearly 40 million viewers around the world — the show's largest audience since 1984 and larger than last year's by nearly 50 percent.
And every year, the challenge is to balance the needs of the TV and live audiences. The show needs to sound amazing in 5.1 surround on television for millions of viewers around the world, whose level of excitement and engagement translates into record sales. At the same time, they also have to make it a knockout for the crowd inside the Staples Center, where the first score or so of rows is made up of the most powerful and, in some cases, sharp-eared people in the music industry. "What we have to remember, and what sets this show apart from any other you mix, is that the front rows of the audience are made up of record producers and engineers, and they don't want to be assaulted, but you also still want the sound to have some impact,” Reaves says.
After nine years using the Yamaha PM1D console for mixing both FOH and monitors, ATK Audiotek, the sound system providers for 22 of the last 27 Grammy Awards shows, have brought in Digico desks. Reaves, working his 10th Grammy show, employed a 256-input Digico SD7 console at the FOH position. This sits next to a 96-channel SD10, operated by ATK Audiotek Vice President for live events Mikael Stewart, who is doing the production mix. There are two 96-input Digico SD10s for the monitors — one for each side of the split main stage — worked by Tom Pesa on stage right and Michael Parker stage left. All the consoles process audio at 96kHz.
Reaves has nothing but good things to say about the now-discontinued PM1D, but agrees that it was time for a change. "The show is getting more complex every year, and the ability to have more inputs and a faster control surface really helps. I had gotten to the point [on the PM1D] where I could have operated it in a coma, so I have to think a bit about what I'm doing as I move through this new console. There was a learning curve, but the effort is definitely worthwhile.” He says he's also impressed by its ability to run Waves plug-ins, though he opted not to do so for the Grammy show. "I just didn't want to add that extra layer of complexity to it. Maybe next year.”
Louis Adamo, of Hi-Tech Audio (the company that supplied the consoles), says that the main concerns surrounding the new console arrangement had to do with snapshots and mic-pre settings, since all FOH and monitor consoles were using the same mic pres. "We needed to make sure that everyone was clear on what actions would affect what the other guys had going on, and establish a communication and workflow protocol between them so no one saved bad information or caused changes for someone else at the wrong time,” he explains. "Otherwise, the questions were simple things you'd expect when working on a new platform — none of them had used Digico consoles for more than a show or two before this one.”
Pesa adds, "I can recall when the PM1D was unveiled here for the first time and that was a big deal, it was exciting. That's how this is, too.” The Digico is complex, but flexible, says Pesa, and so has been the crew's approach to learning it. "Everyone is looking for their own way to learn it,” he says. "I think the challenge really comes from within.”
With the show's live audio running completely on an Optocore network for the first time (a smaller-scale network was used last year, just for the rear delay speakers), the four consoles and the mirror versions of them used for fail-safe redundancy had their inputs and outputs linked via fibre-optic cable, for a total of 448 I/O between them: 288 shared inputs available to all the consoles, and 200 shared outputs. The monitor consoles had control of preamp inputs for instruments, while inputs for vocal mics and the Pro Tools playback system were under the control of the FOH console. "We sat down and found a level that we all felt comfortable with as a starting basis, and we can then all use digital trim from there to adjust individual inputs,” says Reaves.
Jeff Peterson, the system designer for ATK, adds: "When you go to a digital infrastructure it's a whole new proposition of full scale versus what we used to use as 0dBu with analogue amplifiers and how much gain you're putting out. It's a whole new digital world and you have to have parameters like your limiters all lined up the same [across the boards].”
Peterson, who sat monitoring the show on a MacBook Pro laptop throughout the rehearsals and the broadcast, points out that the Optocore is being used to drive the entire system this year. "There's just one A-D conversion at the mic preamps, and back to analogue in the amplifier, all via a 24-channel MADI feed,” he says. Ten XTA speaker processors are sending an AES signal to the amplifiers. Another first is driving the Powersoft K10 amplifiers digitally, though ATK opted not to use onboard DSP for the amplifiers yet. "ATK will probably move to internal DSP processing next year; this year we're still using the XTA external processing,” he says.
Redundancy was critical with this many firsts. Peterson and Stewart set up the system, including all of the consoles, at ATK's shop for an entire week, running every possible failure scenario they could conceive of, including pulling out cables and performing emergency reboots. If, for instance, the FOH console were to fail, the production console could immediately be manually switched over to cover that task as well, drawing a music mix from the broadcast truck over the Optocore network. Or if an amp rack were to fail, the Powersoft amps would be switched to their analogue outputs to the Vertecs.
As part of the mandate that the broadcast sound must take priority, Peterson says that the PA system is adjusted for that. "We go to great lengths to avoid reflections from the rear of the arena going back to the stage that could pollute the broadcast,” he says, and this is done using speaker zones and delays. Again, it's part of the pursuit of a balance that will keep the room and the live audience energised but won't overpower the broadcast. "By reducing the size of our main clusters from what a touring system might use, aiming them away from the reflective rear walls, and hanging additional speaker clusters to cover the upper and rear seating areas, we keep the energy returning to the stage under control,” he explains. "Every delay speaker or cluster is delayed back to the main speaker clusters. The same thing goes for fill speakers, like front fills and pit fills. Tuning and time-alignment was done with Smaart 7 software from Rational Acoustics, using Earthworks measurement mics and a Lectrosonic wireless mic transmitter. The Lectro transmitter makes tuning a lot easier when running a test mic around to all of the different coverage areas. It saves us running 500-foot microphone cables for the tests.”
Peterson says the house volume will consistently range between 100 and 105 dBA, though it's crept up as high as 110 at times. "It can't be too loud but it also can't be too quiet — it's a real fine line.”
Michael Abbott, audio director and coordinator for the Grammy show for 26 years, says the new digital infrastructure came at a time when the show itself was becoming more complex, with several multiple-artist set pieces, such as Maroon 5 and Foster The People joining the Beach Boys for their 50th anniversary appearance on the show. That performance alone used 110 inputs. "We also have string sections performing with Paul McCartney and Bruce Springsteen, and neither of them use click tracks, so we can't do a pre-record to add to the sound to embellish it during the performance,” he says. "There are 22 performances this year, versus 18 last year, plus we have a show across the street.”
The latter takes place on a stage built on the LA Live plaza, using four JBL 4889 enclosures per side for the Foo Fighters, three ATK C6 boxes per side for Chris Brown and eight ATK CSW 218 subs per side. The consoles used were Yamaha PM5Ds for Chris Brown's FOH, production audio and monitor mixes, a Digico D5 for the Foo Fighters' FOH and a Midas Pro 6 for their monitors. "That show we're treating as integrated into the rest of the production via fibre-optic from the stage to the Music Mix Mobile truck, rather than as a stand-alone show like we did a few years ago with the Foo Fighters.” There were 134 SMPTE fibre strands and 26 pairs of single-mode fibre from the outdoor stage, making the most fully digital Grammy show also the most complex in terms of inputs.
Up in the truck, broadcast music mixers Eric Schilling and John Harris, the latter of whom is also a principal in Music Mix Mobile, were further refining the ping-pong approach they've been taking to the Grammy rehearsals for the last several years. It well reflects the stage design, which has A and a B performance stages side-by-side on the main platform, so while an act is rehearsing or performing on one stage, the other stage is being prepared for the next show. All audio and other signal connections are organised on wheeled risers or carts, rolled on and connected in a choreographed ballet of stagehands and electricians.
Until 2008, a single truck was used to mix all of the rehearsals and later the actual broadcast. That year, as a way to expand the number of musical performances during each show, they used a matching pair of trucks. Truck one would mix the day's first rehearsal, which would be recorded through the van's Avid Profile desk to a 160-channel Pro Tools system. The mix comes together during the approximately one hour allotted to most acts, with Schilling taking input from the artists' own FOH mixers and/or record producers. Once the rehearsal is over, the tracks are ported via MADI over fibre to the second truck, where Schilling will go to fine-tune the mixes. His mixes are saved in the console's memory, while Harris starts mixing the soundcheck of the second artist.
The two will repeat this circular trek over and over as each artist comes in for rehearsal. In 2008 this arrangement enabled the show to go up to 16 performances in three hours, up from the dozen or so that had been typical a decade earlier. This year, they reached 22. This approach has also created what Schilling calls "a real comfort zone” for artists: "The confidence level is tremendous, knowing that we can take the time they need to feel good about the mix we're doing for them,” he says.
Similarly, both mix engineers work in very familiar environments no matter which truck they're in: the M3 Horizon truck is a carbon copy of its sister M3 Eclipse, from their Avid Icon digital desks and Genelec DSP 5.1 monitoring systems right down to the acoustical treatments in the control rooms.
The connectivity between the two trucks carries both audio and video. Audio is transferred between the trucks via MADI, while video is recorded as a SDI-HD Pro Res file, and shared for playback via an Avid Video Satellite software option. M3 engineer-in-charge Joel Singer designed the system. For the actual broadcast, the M3 Eclipse truck does all the live surround-sound music mixes, which are then fed to the Denali main broadcast trailer located inside the Staples Center, where engineer Tom Holmes mixes them with audience ambience and production dialogue audio supplied by Stewart's and Reaves' consoles. From there, they head to satellite distribution.
With its emphasis on a traditional red-carpet approach and an apparent terror of going off script, the Grammy Awards show may seem a bit old-fashioned, particularly compared to the sleeker Brit Music Awards that take place a week later, but its importance at a time of distressed sales in the record business cannot be underestimated. Dave Bakula, senior vice president of analytics for entertainment at media metrics company SoundScan told USA Today that he wouldn't be surprised to see a triple-digit bump in sales of Adele's 21 as a result of her six awards that night (turns out she sold over 700,000 units that week). The day after the show, Americana duo the Civil Wars' Barton Hollow shot to the number 5 spot on the iTunes album chart, and they had two top 10 titles on Amazon's Movers & Shakers list, which ranks percentage growth in sales over a 24-hour period. The cause and effect is undeniable. You have to credit great music for that, but some of the credit also has to go to how that music is presented, to both the live and broadcast audiences. .
The Grammys have 230 radio frequencies allocated to the show. Dave Bellamy, a partner in Soundtronics Wireless, the RF vendor for the show, says they used 218 of them, and many of those simultaneously for the multi-artist segments. Bellamy is constantly busy allocating frequencies and tracking down noise, which emanates mostly from the LEDs along the bottom of the stage. At times, he says, those RF emanations were inducing a noise floor as high as -68 to -78 dB. Optimum RF microphone performance starts around -97dB, so Bellamy had to get the noise floor down from -78dB to at least -96dB, which is quite a jump on a logarithmic scale. Part of the solution was for the stage crew to purchase and apply a microwave-absorbent material around the bottom of the offending part of the stage set, at a cost of £17 per linear foot, which translated into nearly £1,600 per roll. These are the kinds of costs that will be incurred as the US moves further into the spectrum reallocation process that began with the so-called White Spaces issue, in which the Federal Communications Commission eliminated most of the spectrum range between 700 and 800 MHz for broadcast and reallocated it for consumer wireless device use, significantly compressing the amount of spectrum available for live events. "That's really the big issue going forward for wireless — less spectrum to work with,” he says.
Bellamy's assiduousness and some serendipitous good fortune helps the show go off with only a single wireless hitch, and that one is putatively attributable to a completely different industry trend. Country singer Jason Aldean's vocal microphone shuts down during the ride-out on his duet with Kelly Clarkson, and Bellamy says the likely culprit is the failure to replace a Teflon retaining ring that held the mic's battery firmly in place after the microphone had been customised. Bellamy cites the proliferation of sparkled and bejeweled personal wireless microphones that have led to a minor epidemic of such failures, as collateral damage to the form-versus-function debate that live-sound folks find themselves in as artists look for more highly personalised microphones.
More and more of the music audio used on the Grammy Awards show is pre-recorded, delivered from a Pro Tools HD system run by operating engineer Pablo Munguia. The use of backing tracks has increased in recent years, in part because it makes more sense for certain types of music, such as for Chris Brown's athletic rendition of 'Turn Up the Music' on the show, and also because timecode-striped recordings can be more easily synched to stage lighting and video effects. But even self-contained bands will sometimes sing to music bed tracks, which underscores the desire to keep as much of the show under control as possible, especially as a run of unanticipated events, such as Kanye West's notorious interruption of Taylor Swift's first-ever Grammy acceptance speech in 2009, have made broadcasters, producers and advertisers nervous.
Pre-recorded vocals are another and more sensitive matter. Ever since the Milli Vanilli controversy (in 1990 the German duo's Best New Artist award was rescinded after it became known that the group's lead vocals were sung by other singers), the Grammy shows have emphasised the integrity of live vocal performances. But crowded stages and even more crowded frequency bands for the 40-plus RF vocal microphones used in the show, and the need to shoot for as much perfection as possible, have increased the number of still-rare occasions when lead vocals are pre-recorded. Vigourous dance numbers, such as Rihanna's twisty terpsichorean manoeuvres on 'We Found Love' at this year's show, are another compelling reason for the use of more pre-recorded lead vocals.
While Rihanna's vocal moves on stage were clearly not matching what was coming out of the PA system or on television during a number of moments of the performance, and during her subsequent duet with Coldplay's Chris Martin his guitar strumming wasn't quite synching up with what we saw and heard, Michael Abbott, the audio director and coordinator for the show for the last 26 years, regards it as a non-issue, aware that authenticity and expectations can sometimes be mutually exclusive. In fact, says Abbott, he does keep an eye on the various blogs and Twitter feeds that shadow the show, forums populated by both fans and production professionals, and notes that neither cohort ever gets it completely correct as to whether someone is lip-synching. And there are times when tracked lead vocals are a necessary part of the theatrics, such as this year's performance by Katy Perry, which opened with a misleadingly lit body double belting out 'ET' until a transition (in the form of a very real-sounding but totally intentional 60Hz audio glitch) revealed the real Perry picking up the live vocal of 'Part Of Me' from a box being lowered to the stage from above. The same goes for the use of various Auto-Tune types of processing on vocals, which has evolved from subtle corrective use to purposeful creative special effect. Abbott concludes, "I leave it to the viewer to decide, not only when it might or might not be a [pre-recorded] track, but also if and when it really matters.”