Part 1: The Challenge
by Simon Hayes AMPS
“The only way I feel we can make this movie is to record all of the singing live”
Tom Hooper, the director, dropped that bombshell on me in our very first meeting and those are the words that would intrude on every waking thought for the next 12 months.
I have always worked very hard to capture original performances and never rely on ADR; I believe that performances captured on a movie set are rarely bettered in the isolation of a vocal booth, without the presence of other cast members, and months after the movie has wrapped. By then, the actors may be shooting another movie and be immersed in completely different characters, accents and mindsets. Of course, if a director feels he can find a better performance in ADR, it is a very useful tool, but it is such a shame when the cast and director think they have all really nailed a performance on the set, to re-record because of poor sound quality.
I had previously recorded musicals where small elements of live production sound were incorporated into the musical numbers. The last and best known of these was Mamma Mia!, starring Meryl Streep. Many musical numbers were lip-synced in the conventional fashion to a pre-recorded track prepared months earlier in a music studio. But there was a particular number that Meryl specifically asked to sing live. The action required her to climb a steep wall and she felt that the tracks prepared in the studio would not permit her to match her expressions with the action.
We kept a pre-recorded music and vocal track on standby but instead of playing the vocal and music out of a high-power amplified playback rig, we fitted her with an “earwig” and played her the music minus the vocal. This is a tiny, wireless, in-the-ear speaker that, once fitted, is difficult to see unless you are looking directly into the ear. Originally designed as hearing aids, they have been adopted by security professionals and other industries that require sending audio discreetly to personnel.
The recording we made on Mamma Mia! was successful. This experience gave me confidence that, although risky in comparison to going the pre-recorded playback route, recording live singing on set is achievable. It gives a vocal performance that more closely matches the on-screen visual performance.
So, in that first meeting when Tom asked me if I thought it was possible to shoot the whole of Les Misérables live, without hesitation I answered, “Yes.” As Tom and I spoke, he told me of earlier experiences in his career that had resulted in him gaining a respect for production sound and a dislike of committing performances to ADR.
We began to formulate a plan to serve as a rudimentary workflow for the film. He said to me that he knew there had been major technical advances in recent years, not just in the equipment we use to capture sound in films, but in every sector of the movie industry and within the wider electronics industry. His next comment really shaped my own opinion and confidence in how seriously he was committed to recording the film live when he said, “I want you to use every single piece of modern technology available to us to record the performances on the set live with high-enough quality that will ensure we won’t need to return to an ADR studio to re-record performances because of poor sound quality.” This was the backing I needed to realize his vision and take on a unique project that had never been attempted before.
Let me take a moment to qualify that last comment: In the early ’30s before modern production and post-production techniques, singing was recorded live on movie sets. This was before the idea to pre-record vocals and lip sync to playback had been conceived, and before the ability to ADR performances in post-production existed. These very issues become the subject of the famous movie, Singin’ in the Rain.
However, to my knowledge, none of those movies had attempted a workflow that allowed the actors to set the tempo, with the timing of their acting taking precedence over the musical orchestrations. This was Tom’s revolutionary idea. Having agreed with him that it was possible to record vocals with high-enough quality on the set to use in the finished film, he then began to explain to me the next piece of his unique plan.
He asked me about other live recordings I had done using a prerecorded track for the talent to sing to. “What would have to happen if an actor wanted to take a moment to reflect on something within their vocal performance?” My reply was that it was possible to take tiny moments but only within the strict confines of the pre-recorded music track being played to them in their earwigs. Tom went on to ask, “What if they want to take a longer break, a larger pause?” I replied it would be impossible because the vocal would become out of sync with the music. Tom said to me that was exactly what he wanted to avoid. It was his vision that actors would be able to reflect on emotive moments or take time to complete actions without being worried about falling behind the music. They would set the tempo and the orchestration would happen in post-production. The orchestra would play to the actors’ performance and the acting would drive the music, not the other way round.
At this point, I knew he was asking me to take part in a venture with huge risk, something that had never, to my knowledge, been attempted before. As Tom waited for my reaction, I thought carefully. He asked me if I was prepared to try this and I asked if there would be any pre-recorded music at all. He said no. He wanted the actors to be followed by an electric keyboard played on-set, with the pianist reacting to the actors’ performance rather than setting tempo in the customary fashion.
I responded that I had the technology that would allow the actors to hear the piano in a hidden earwig so that microphones could pick up a clean vocal recording. If the music department and picture editor could support Tom’s vision to orchestrate afterward, I certainly had the ability to successfully bring this methodology to the movie set.
Tom then told me about his collaboration with Gerard McCann, his supervising music editor. Gerard had also worked with him on The King’s Speech and shared his conviction that allowing the acting to set the tempo of the music was a valid approach. He asked me to meet with Gerard as soon as possible so we might immediately begin collaborating on developing a methodology to achieve these ends. We had now been talking about two hours continuously and this was our initial meeting!
Discussing the technical challenges of recording the singing live, Tom told me he would need to cover the action with multiple cameras. If the cast made a perfect take, he would want to have it well covered with wides, mids and tights all at the same time. Being able to preserve a “perfect” take was only part of his vision for the film; he also wanted to allow his actors to freely overlap their lines. That technique can greatly enhance performance authenticity and vibrancy but it presents difficulties for the editor who must find a place to make cuts. It also necessitates that all speaking and singing roles must be recorded on-mike at all times, whether on-camera or not. If the sound mixer has both actors (on screen and off screen) covered, it is possible with a skilled picture editor and dialog editor, to preserve the on-set performances by carefully weaving in and out of the speeches, cutting on syllables and tiny pauses. Utilizing both multiple microphones and multiple cameras, it is possible to exercise control over the material. Adding to the complexity, Tom intended to extend this flexibility to interactions between the principals and the chorus. He and his DP, Danny Cohen, planned to use as many cameras as needed to ensure that every good performance would be captured sufficiently so that re-takes for coverage would not always be needed. This was his plan for filming Les Misérables.
I agreed wholeheartedly that this was the best way to bring the immediacy of a continuous theatrical performance to the film audience as I began to formulate the plan of how I might accomplish the task.
We discussed why boom mikes are traditionally the preferred method for capturing movie dialog and why radio mikes are generally treated as a secondary method. I explained to Tom that Production Sound Mixers usually favored boom mikes because radio mikes historically had three huge issues compromising their reliability. The first was range. It is only in recent years that radio mikes presented enough range to be able to be used on movies without indiscriminate splats, pops and hiss. The development of the Lectrosonics digital hybrid system has been a great step forward, yielding not only far greater range but also a significant improvement in sound quality generally. Results are practically indistinguishable from a cabled system. So, issue number one was covered and range was no longer a problem.
The second issue was that lavalier microphones have always been compromised by their tiny size. Their sound quality has always been a long way behind what was considered “studio quality” in the music industry. Performance on a set has been inferior to what can be accomplished with a well-placed boom microphone.
Calling again on my experiences on Mamma Mia!, I knew that there was now a better answer. I tested many mikes for that film and found that the difference between the various lavs on the market was just a matter of audio taste. There wasn’t one product that could really be called “better,” just many products each with their own audio voice. Some on my crew would prefer one and some another. During discussions about the small live element limitations of lavaliers, I met with Benny Andersson from pop group Abba and his longtime engineer, Bernard Lohr. They told me that when they first took Mamma Mia! onto the stage, they were presented with lavaliers for the first time, and coming from a music industry background, had no experience with them. They also tested products from many different companies and were consistently disappointed. Then they asked DPA (a Danish company whose studio products had already impressed them) if they could test the company’s range of lavaliers. Benny and Bernard both told me that DPA lavaliers were the closest tonal match to the mikes they used in the music recording studio and that the Mamma Mia! stage musical had benefited greatly from their sound quality.
To confirm the wisdom of this choice for the vocals in Les Misérables, I set up a test with the DPA 4071 against the three strongest alternate candidates and invited my crew to participate. This time, everyone was in immediate agreement that the DPA 4071 was the winner. It seemed to sound less closed, chesty and constricted than the other lavs and exhibited an openness previously heard only on good-quality condenser mikes used on a boom. It was a really transparent mike that, unlike the others, didn’t add its own “voice” to the recordings.
The other problem I had experienced with lavs was their inability to cope with high SPLs. When you most needed them on scenes that were challenging to boom, like fight sequences or exterior action sequences where the talent was likely to be shouting or screaming, they would often distort due to their capsule being tiny and ‘hitting its end stop.’
I was impressed that the DPA could handle amazingly high SPLs. Toward the end of the test, members of my crew were screaming into it while it was rigged on the chest, trying to get it to crack off and distort but it remained smooth. Yet when the vocal was brought down to the quietest whisper, an almost ‘breathed’ dialog, it was sensitive enough to capture the change in level instantly and without coloration.
Upon research I found out that DPA had designed the lavs not just with vocals in mind, but that they had created a tiny mike that could be used in music recordings when an instrument needed to be closely yet discreetly mic’d. This gave me enormous confidence in my decision to use them for live singing on Les Misérables.
So I knew that I had a good microphone candidate to address the issue of recording the singing with clarity and accuracy.
The third issue with radio mikes did not have such an easy answer. It is the primary reason that, even with huge advances in radio and microphone technology, most Production Sound Mixers still prefer to capture dialog with boom mikes and rely upon radios only when a boom cannot be effectively deployed. Lavaliers must be hidden underneath costumes. Whether the clothing will “sound good” or cause rustle over the dialog is often a gamble because the visual choice of costume generally takes priority over radio mike placement. If an actor is wearing a nice cotton T-shirt, it is possible to get really great sound quality but, if the costume designer needs to have the actor in a silk shirt, the dialog may be unusable. The “plug-ins” now available to dialog editors and re-recording mixers have become amazingly advanced and many background noises can now be effectively cleaned up but severe clothing rustle remains notoriously difficult and removal efforts will often remove some frequencies of the voice as well. On normal dialog a small amount of cleaning up of the voice may be acceptable but the artifacts of a ‘cleanup’ are much more noticeable with singing.
I decided to take up Tom on his challenge to use every advance in modern technology to enable us to record the best quality vocals possible.
I explained to him that the booms would have to be a secondary way of capturing the vocals due to the multi-camera cinematography. Sure, the booms would work fantastically on some shots but we could not completely rely on them. For radio mikes and lavaliers to be our primary method of recording voices, we needed a plan to effectively eliminate clothing rustle. My proposal to Tom was that we place the mikes on the outside! To his credit, he continued to listen to me although my suggestion completely broke with film industry tradition.
I intended to take advantage of advances in CGI to allow greater freedom in microphone placement. I proposed working closely with the costume department and obtaining swatches of matching fabric that might be used to cover the microphone mount. A small cut in the exterior of the costume would permit mounting the mike on the outside and a camouflaging piece of matching fabric would make it inconspicuous. Since the mikes would be attached using DPA’s recommended mount, and since no fabric would touch the grill, the application should be noise- free. Although the mike would be clearly visible to the human eye, on a wide shot and on a moving costume, it would be very difficult to see, and on a tight shot it would actually be beneath the bottom of the camera’s frame line. For the medium shots we would rely upon advances in VFX technology to paint out the mikes. Tom was immediately 100% receptive to this idea, commenting that the VFX department could spot the mikes that needed removal once picture editing was complete. At that moment I knew that not only had we formulated a unique plan, eschewing many ageold film industry traditions, but also that the plan placed an importance on production sound that I had previously only dreamed of.
At this point, Tom and I had been talking for three hours and he told me he had another appointment. I expected the meeting to end when he told me that Hugh Jackman was rehearsing with the Music Director, Stephen Brooker, Cameron MacIntosh’s longtime collaborator, and I was honored when he asked me to come and meet them both.
As I walked into the studio, Hugh was in mid-song and I was struck with the huge dynamic range of his voice and what an accomplished singer he was. At the end of the song, Tom introduced me to both Hugh and Stephen and explained I was the sound mixer who was going to record the musical live. Hugh looked at me warmly and told me how impressed he was that I had taken the challenge. I couldn’t have met two more welcoming individuals who would turn into friends and collaborators in the coming months.
Over the next few days, I began the task of technically planning what I would need to facilitate Tom’s vision.
I began to think deeply about the project and my previous experiences working on musicals. Since the singing was pre-recorded and the singing was generally lip-synced, the tempo had been set months in advance of the actors coming on set. Tom had explained to me that he felt live singing held an energy and truth that miming could never fully replace and he believed he could detect a falseness that disconnected the audience from the performance.
I considered this. His thinking about mimed singing closely paralleled my thinking about ADR in general. A few lines in an action scene might pass unnoticed but longer passages can be ruined by the need for dialog replacement. I began to think about normal musicals where a singing number might run for three minutes followed by dialog scenes and then by another musical number. The singing would be mimed to playback and the conjoining dialog scenes recorded live in the usual manner. I believe that the audience senses miming immediately but they are conditioned to accept it by other musicals and by MTV experiences. They can subconsciously accept the theatricality for a short while as part of the willing suspension of disbelief. Just as the acceptance is becoming ragged, the musical number ends and the audience again experiences live recording and forgets the distrust building during the mimed number and the cycle starts again. We accept this pattern of connect and disconnect because recording live singing is so difficult and time-consuming. To ask an audience to connect with a lip-synced mimed performance for more than 2½ hours would, I think, run counter to their instincts.
This led me to another subject I considered while planning Les Misérables: why do audiences not connect with ADR and miming?
In my opinion, we, as human beings, have incredibly sensitive instincts that stem from our beginnings and are deeply rooted in our genetics. When connecting with other human beings, we are predisposed to process and evaluate the subtlest facial expressions, changes in voice, mannerisms and body language. Why do we do this? To try and work out whether the person we are talking to is honest, whether he is worthy of our trust. This is an instinctive part of survival and taps into out most basic reactive quality: fight-or-flight. We do this subconsciously most of the time, only noticing a problem in our conscious mind if our subconscious has flagged an alarm from our continuous evaluation.
The more I thought about this, the more obvious it was that the cinema performances that engage us are the ones that our subconscious accepts as true. A truly great performance is one that keeps the audience in the moment and doesn’t allow any subconscious alarm bells to ring, bringing us back to reality and creating distrust in the performance.
Tom Hooper’s vision of a completely sung through Les Misérables would not, in my opinion, have kept an audience captivated for long if it been lip-synced from start to finish; there would have been an eventual disconnect for even the most ardent of fans. After all, who would go to see a stage play that was mimed?
Before commencing technical preparations, I met with the producers to explain my plans. I am fortunate to have an excellent working relationship with Working Title Films, having recorded several films for the company, and I’ve worked with producers Eric Fellner, Tim Bevan, Debra Haywood and production executive Sarah Jane Wright many times. I explained that to make the project possible I would need a much larger crew than usual. I always use two boom operators on my projects so I may record off-camera lines, giving the editors more options and reducing the need for ADR. So they were familiar with my preference for a larger crew but this project had even larger needs than usual. I intended to use two boom operators with mono booms to capture the individual singing lines and also a 3rd boom fitted with an MS stereo mike for the chorus and group ensembles. Members of the chorus would be individually radio mic’d but this track would add width and texture to the recording. I had also worked out that on most days we would be running as many as 20 radio mikes and 75 ear pieces so I would need two sound assistants just dedicated to rigging radio mikes and wrangling earwigs—a huge task in itself. To really be able to guarantee sufficiently quiet backgrounds for Tom to use the live on-set vocals, we needed a member of the sound team whose sole responsibility would be generally spotting background noise issues as they arose and dealing with them professionally and swiftly.
The need to have all the participants mic’d-up for rehearsals was a further complication. In a normal shoot, the cast can sometimes do early ‘block through’ rehearsals without being mic’d to save time and then be rigged just before final rehearsals. Our plan for Les Misérables called for the lead singers to take their own pace and a piano accompanist would follow them and provide a tempo for the chorus and other singers. Since the pianist was to be off-set to keep the piano clear of the recording, it was essential that he hear the performers through their mikes and that his piano track be fed to earwigs worn by all the actors. This meant it was imperative all the actors were fully rigged before the first rehearsal. This was nothing new to me and my team having just finished Ridley Scott’s Prometheus, where all the actors had to be mic’d inside their space helmets and fitted with earwigs so they could communicate with each other inside the soundproofed glass helmets. Of course, it took a while for the rest of the crew to realize it wasn’t possible to call a rehearsal unless the sound team had completed rigging the actors.
Debra and Sarah-Jane agreed that if Tom wanted to shoot the musical live, the sound team had to be employed in larger numbers than a regular movie. They understood that usual crewing levels in the sound department could cause schedule issues while shooting due to the far greater workload.
At this point, I could put together my dream team: a seven-man sound crew.
Arthur Turner and Robin Johnson, my boom operators, have been working with me since 1997. We’ve done more than 40 movies together and work as a tight team. We each know our roles and can work out complex issues quickly and efficiently with minimal discussion. For this project, Arthur and Robin held the titles of Key 1st Assistant Sound and 1st Assistant Sound, respectively. Since the project was unique, and their roles and responsibilities were so much more than pure boom operating, we thought it right to use titles more in-line with our colleagues in the camera department.
Joining Arthur and Robin would be Paul Schwartz, who has often worked with us as a 2nd Unit Boom Operator. He would be operating boom number three, the stereo boom and also help coordinate the production sound equipment and interface with the music department’s Pro Tools rig. His title was Sound and Music Maintenance.
James Gibb would be our 2nd Assistant Sound and would serve as chief radio mike and earpiece technician. He has worked with us on about 12 movies and is very capable. We brought in Andrew Rowe, another collaborator from 2nd unit work on other pictures, to assist James with radio mikes and earpieces. He would also be responsible for talkback systems and monitoring.
Duncan Craig was brought aboard as Sound Trainee. He has worked three films with us so we were privileged to have someone so experienced in that role. His responsibilities were carpeting the actors’ feet, soundproofing anything noisy and generally helping with anything sound-related.
So, we had a seven-man sound team in place—a bigger main unit crew than I had ever worked with before. They were all handpicked from the best technicians I know and all excellent choices not just for their individual skills but also for their experience working together as a team. As events would prove, this large crew was absolutely necessary for the job required.
Editor’s note: In Part 2, Simon Hayes will tell us about the implementation of these plans.