Sooner or later you'll encounter a dialogue problem you can't repair. The background noise is the same on all of the takes, so alternate files or noise reduction won't save you; or the actor is so annoyingly accurate that she always says the needed line as she steps on the creaky spot in the floorboard. Maybe there's no close-up coverage for the off-mic wide shot you're trying to fix, or the take chosen by the picture editor really does have the best acting despite the F-15 flying overhead. There are a million reasons for rerecording dialogue.
Everybody loves original production sound, so there's no point discussing the magic of the sound from the shoot, when the actors were hyped and in character. There's always the concern that replacing the originals with studio recordings will kill the charm. Nevertheless, if you're looping a scene then someone has accepted that it must be redone.
What, though, are the reasons for rerecording a line, a shot, or a scene? I like to divide them into three groups based on how the lines will be used: replacements, adds, and group loop.
Replacements are rerecorded lines of existing dialogue. When people think of postsync, these are what usually come to mind.
• Noise problems. Easy to spot: excess general ambient noise (a scene shot in a convertible or next to a waterfall), temporary loud noises (car horns, voices, sets falling), wind. The list is endless.
• Technical problems. These include a wide array of screwups: radio mic breakup, hidden mic clothing rustle, rain striking the zeppelin, or any other microphone-related problems; distortion; wildly underlevel recordings, and so on.
• Perspective and voice quality. Sometimes it's simply impossible to record a shot: very wide shots with lots of headroom and no place to put the boom mic, scenes shot in overly “boomy” spaces, or weird sound reflections.
• Acting. Some of the worst ADR nightmares occur when the director or editor doesn't like the read on a line and wants to “improve” it. This can turn into postsync perdition, since the director's new idea of the “right” read may not match the gestures or acting energy of the image on the screen. But try we must.
• Line changes. Sometimes lines have to be changed to fix story problems. This is where you become an expert at squeezing new text into old sync shots without it looking like a Godzilla movie. It rarely works.
• Focus control. To isolate the characters from their surroundings you may have to rerecord their lines. (Perspective and focus are discussed in Chapter 12.)
There are many other reasons to replace a sync line. What's important about replacement lines is that you have to prepare the track so that the new ADR line can be mixed with the rest of the dialogue. More on this later.
Not all ADR lines are intended to replace mangled, damaged, or drownedout production lines. Some of them are added on top of the dialogue.
• Story details (a.k.a. narrative emergency surgery). If the story is foundering because a few critical facts have gone missing, well-placed clarifications might save the day, like the antagonist muttering a bit of vital information while passing behind a post. Decades of exposition have taken place during long driving scenes in buddy films. You think those beauty shots out the window are there just to celebrate nature? Not entirely; that's a great time to plug story holes, reveal details about characters, and up the tension ante. Similarly, a character who's just left the screen can yell from the exit sign that she'll be home at 8:30. This sets up the next scene: It's 12:30 and still no Sally. Her small off-screen line adds to the drama because everyone else in the film had warned her that it was folly to go surfing during the shark warning.
• Breaths and other nonverbal sounds. Sometimes it's not the words that enhance the moment, but rather the breaths between them. A sigh can say more than words. A deep exhale of resignation tells us a lot about the situation, as does grandpa's heavy snort as the settles into his chair. And shallow, quivering breathing betrays erotic excitement. Breaths are a powerful tool, so use them only when truly needed, lest the scene come across as obvious and vulgar.
• Narration. We're supposed to hate narration in dramatic films, viewing it as admission of narrative meltdown. But some films have it, so you've got to record it.
Films are full of human sounds that don't come from the principal actors: the crowd in the bar; commuters on the train or in the station; the vendor in the background selling ice cream to kids; a riotous mob. What they have in common is that they come from actors brought in during audio postproduction to add life, depth, and the occasional story detail to the soundtrack. This is group loop.
Like many people, I tend to use “ADR” and “looping” interchangeably, all the while knowing that they aren't exactly the same thing. Call it what you like as long as everyone understands what you mean. Still, it's worth knowing the differences between them.
Long before DAWs—before projectors and mag film recorders and players could move in reverse, much less “rock and roll” to repeatedly play a line— sound editors rerecorded dialogue lines by creating physical loops of film, which contained sound and picture of a line to be replaced. A film loop was prepared with beeps and visual clues to cue the actor, and because it was a loop, it could repeat continuously.
The actor would repeatedly hear and see a line, and then, when he was ready, recording commenced. Each time the actor heard his line, he would immediately repeat the text, in a process that continued over and over until everyone was happy. Then up went the next loop, and so on. This system was great for rhythm, since most people can manage to accurately repeat the music of a phrase while holding onto its spirit. You couldn't, however, conveniently check the sync of a loop during the looping session.
As technology made it possible to better control mag dubbers and projectors and to preprogram the complex array of electronic commands involved in rerecording, it was all but inevitable that machine-assisted looping would come along. Meet ADR, automated dialogue replacement, which introduced a new way of working:
• The actor listens to her lines leading up to the line to be rerecorded while watching her sync picture on a screen.
• She hears a series of beeps as her cue approaches. There's likely a line, or streamer, wiping across the screen from left to right.
• On what would have been the last beep, the streamer reaches the right side of the screen and the actress sees her sync picture but no longer hears the guide track.
• A cue light glows and the actor speaks her lines—hopefully in sync.
• The monitor controller switches to playback and the actress hears the continuation of the guide track.
The process can repeat endlessly.
With gifted actors, skilled in the process, ADR is a real timesaver. All cue information can be programmed offline, and the “live” nature of ADR recording gives the director, ADR supervisor, and supervising sound editor an instant indication of whether or not the recorded line is acceptable. The ADR engineer can immediately combine selected pieces of chosen takes, giving everyone immediate feedback as to what works and what doesn't.
Slick as it is, ADR isn't for everyone, nor is it for all occasions. Many actors don't like the pressure of having to perform live, in sync, all the while focusing on a good, matching performance. For them, it's best to use a modified looping technique, which I'll describe later.
Another term you'll hear bandied about on recording stages is postsync. Some like this term as a general description of the whole after-the-shoot voicerecording process. If “ADR” seems too sterile but you're too much a stickler for accuracy to use “looping” incorrectly, then postsync may be just the description you've been looking for.
One of the major differences in between working on American films and Indian films is that most Indian films don't record sync sound. Around 90% of films that release in India are completely ADR. There are 28 states in India, each of which speaks a different language. So there are always actors from different parts of India acting in films where they don't even speak the language. Another reason is the loud traffic all throughout India, which makes it nearly impossible to record clean production tracks. These are some of the reasons why most filmmakers prefer ADR over production sound. But it's slowly starting to change; I am seeing more filmmakers taking the effort to record live sync sound in India.
Kunal Rajan, supervising sound editor
Achchamundu! Achchamundu!; Vishwaroopam
Postsync (or looping or ADR) demands organization. Someone has to compile the postsync requests from the director, picture editor, and dialogue editor and find the problems no one else noticed. Each line of dialogue is checked against the guide track and precisely spotted for placement. Copies of the spotting go to the recording studio for machine programming, to the supervising sound editor for approval, and to the production office for talent scheduling.
During the recording session, there has to be a “voice of reason” who soothes impatient actors while keeping in mind the needs of the film. And the hundreds, easily thousands, of soundfiles generated during the postsync sessions have to be managed in a way that tells the ADR editor what to do with the fruits of all this work. It's no wonder that there's a unique job title for the person responsible for managing the ADR on a film: the ADR supervisor.
A decent-sized film will have an ADR supervisor. It's too big a job to land on another busy crewmember. If the film you're working on has one, you can skip the rest of this section. However, many small-budget films don't have an ADR supervisor, nor do they have an ADR editor. Responsibility for ADR is divided between the supervising sound editor and the dialogue editor. If that's your situation, read on.1
One thing that all low-budget films have in common is, well, low budgets. There's no money to waste, and everyone wants the precious funds to go into the film, or at least into the right pockets. If you're responsible for supervising the ADR, remember—before, during, or after the recording sessions—that poorly organized ADR is a good way to hemorrhage money and fry nerves.
Today, most people manage ADR with software or software/hardware combinations that greatly streamline the entire undertaking, from spotting to editing. Although we still must determine what to loop, and then produce various forms of paperwork, record the stuff, and edit it, a bit of automation goes a long way to ease the tedium and reduce the likelihood of mistakes. Still, the process is the process, no matter how it's done. With that in mind, this chapter presents ADR workflow as a manual process. You may be using a slick ADR management program, but it can't hurt to grasp the underlying mechanism.
Momentarily forgetting the film's timeline realities (actors' schedules, recording studio availability, temp dub requirements, impatient producers), the ideal time to record ADR is just past the midpoint of the dialogue editing process. Decisions about postsync start when you spot the film with the director and picture editor, and they continue as you carefully go through the film looking for technical ADR calls. However, unless you're psychic, you won't be sure about the ADR call list until you've had a go at the dialogue. Lines you were sure you had to loop will be easily fixed with alternate takes, while many an unforeseen problem will rear its head and require looping. The later in the process the recordings are scheduled, the more accurate your list will be. Too late, however, and you squander your margin of safety.
If I have a seven-week dialogue editing schedule, I try to plan ADR recording to begin at the end of week four or the beginning of week five. This way I'll have almost three weeks of real editing to learn the tracks and then a few days to prepare the paperwork. Recording ADR for, say, three or four days leaves ample time to edit the loops and prepare tracks for the new lines and still have time for a couple more dialogue passes before the premix.
This isn't an ideal schedule because it steals more than a week from dialogue editing. But remember, this is about saving money, and somebody has to pay. In this case it's the dialogue editor. Of course, the more help that's available, the less you'll be distracted from editing. A good assistant can prepare ADR call sheets and recording logs, so you'll be in less of a panic. If the supervising sound editor can cover the recording sessions, you'll have more time for editing.
But ADR scheduling is rarely about your needs alone. Actors are often not available when you want them, so special recording sessions are necessary. The film is being submitted to an important festival and certain loops have to be recorded very early in the editing process. Or maybe studio or focus group screenings mandate decent loops early on. You just have to be flexible and not resent the lost editing time.
You'll get an idea of the ADR load at the initial spotting session with the supervising sound editor, effects editors, director, picture editor, and perhaps others depending on the structure of the film team and the nature of the production company. This first meeting isn't a detailed ADR spotting session but rather a first coming-together during which the filmmaker hands off responsibility to the sound department. There's much on the table besides ADR and dialogue, so the most that you can hope for as far as ADR is concerned is to get an idea of the scope of the postsync needs and to communicate this (good or bad) news to the director and producer. You'll also gain a grasp of the director's enthusiasm or reluctance regarding ADR recording. This will provide a clue of what has to be saved “at all costs” from production sounds and how aggressively you should spot the ADR.
As you edit the production sound, you'll encounter problems that no one noticed during the spotting session. Most of them you'll be able to fix with bits of alternate takes or clever use of room tone, but there'll be some you can't sort out. Add these new problems to the ADR call list and move on. When you come to a production line that the director called for ADR, study it, keeping in mind why it was flagged. Go back to alternate takes and see if there's anything you can do to save it. If you can rescue the line—solving the problems that brought attention to it in the first place—while respecting the spirit of the performance and delivering good sound, you may be able to remove it from the ADR call list. But not yet.
By now you have a handle on your major dialogue problems. You know what will and probably won't work. On your ADR notes you've marked what you've been able to fix as well as the problematic lines you've discovered since the initial screening—lines you want to loop. Several days before the scheduled ADR recording, you'll need to properly spot the film and begin creating the paperwork. “Several days before” is purposefully vague. It depends on how many lines you plan to loop, the availability of the studio, and the needs of the production company.
Screen the film with the supervising sound editor, going over each line and confirming what's in and what's out. Show her the called lines you were able to resurrect and then decide together which ones still have to be looped. Add to the list any “add” lines you feel would be instrumental in bridging a dialogue transition or clarifying or focusing a scene. When this ADR spotting session is finished, you'll have all of the information you need to create the ADR paperwork.
Whether you're working on Excel or Filemaker, or on any number of automated ADR spotting utilities, you must enter timecode, character code, and of course, the text (this is explained in detail in the next section).2 If you use ADR software, things like timecode and character codes are taken care of; without such software, you're on your own. As soon as you (more-orless) finish, sort the ADR cue list by character, and send it to the production office so that they know how many lines—and hence how many recording hours—are needed for each character. Remember, you can always modify the ADR call lists you sent to the production, but they need a place to start.
Find out if there'll be a special mix for tamed-down language. A TV version may require special postsync lines, not needed in the full mix, to replace potentially offensive exclamations, blasphemies, and steamy pillow talk. You, the supervising sound editor, and perhaps the director will become skilled at creating nonsense nonprofanities that match the mood of the scene and the sync of the shot. These special lines are recorded at the same time as the full-mix ADR, and their code names carry the suffix “TV.”
There's a lot to ADR: microphone selection and placement, room acoustics, dealing with actors, and knowing when you have the best take. Fortunately, the technical issues are rarely your concern—that's what the ADR recording engineer is for. But you're still left with the huge responsibility of managing the vast amount of information generated by the ADR process. Getting through an ADR session without blowing it or “blowing a fuse” requires organization and a decent sense of humor. By carefully spotting the ADR cues and organizing them in a way that's comfortable for the production company, the engineers, and the talent, you'll streamline the session. When you clearly have your act together, the talent will be more relaxed and you'll have greater authority. Controlling a session is vital, because you, more than anyone, know what is needed to complete the dialogue. If you've earned the respect of the actors and the engineer, you'll stand a better chance getting some extra takes or those few minutes beyond quitting time.
By the end of ADR recording you'll have accumulated countless takes, and you'll have very little time to edit them into your dialogue. Except in extraordinary cases you have to decide in-session which is the best take (or the takes you'll combine to create a good performance). That's why good paperwork is critical at each step of the process:
• Before the recording session. Compile the ADR calls, breaking down each long line into manageable lengths and precisely noting its start and stop times. From this you create a master ADR cue sheet (sometimes ADR call list)—a list of all ADR lines, in film order. Each line carries a unique ID number, which includes the character name. Count the lines each actor needs to record and inform the production company. Based on an average pace of about 8–10 lines per session hour, the production office will use this information to schedule the talent. The call list, the Bible of ADR calls, is what you'll use to create printouts of each actor's lines—simplified, easy-to-read scripts the actors will use during the recording session. Also from the master ADR cue sheet you'll make the ADR recording logs, which you'll use during the recording session to keep track of what's going on.
• During the session. The code names and numbers you assigned to each line will become the names of the corresponding soundfiles created in the recording session. During the session, use the ADR recording logs (or “session reports”) to take notes about each line and each take. Note which takes are acceptable and which take is the preferred “buy” (or “best,” “print,” “go,” or even “hero”).
• Back in the editing room. Use the ADR recording logs to figure out what to do with each line. This way you won't need to listen to all of a cue's takes unless the selected ones don't pan out. Place a check mark on each ADR recording log as you successfully edit the line into your film so you know what has—and hasn't—been edited.
Whether you manage ADR with specialized software, standard office programs, or pen and paper, there are three essential types of paperwork you need to understand.
The ADR cue sheet (ADR call sheet) is the master document, containing all of the ADR lines in the film, in chronological (film) order (see Figure 15.1). As with all film industry forms, the details of the paperwork vary by region and film culture. Adapt your paperwork to the style of your local film industry so that people are comfortable with your forms and you don't come across as a hick or a pedantic jerk.
The information at the top of the cue sheet is pretty obvious. One item, though, will hunt you down and make you cry if you ignore it. This is “Version,” which refers to the picture version you used to spot the ADR. It's perfectly normal for the picture department to keep making “improvements” long after picture locks and even after you spot the ADR. If you don't know which picture version your spotting refers to, you won't know which file to send to the ADR studio. If the recording studio pre-programs beeps, streamers, monitor cues, and the like, based on your detailed instructions, and then you show up with a different picture version—who looks stupid?
In the main section of the call sheet, each line represents an ADR cue.
• Code. Each line of ADR bears a unique code that follows the cue from spotting to mixing. The first digit of the code reflects the reel number and the subsequent two or three digits are the serial ID. Before the number code is a three-letter abbreviation of the character's name, so Rebecca becomes REB and Hanna becomes HAN. ADR lines recorded specifically for a TV mix should carry the suffix “TV,” so the TV version of Hanna's line, “HAN 304,” becomes “HAN 304-TV.” If you know that your film will have reels with more than 99 ADR lines, use a four-digit code.
• Character. If you use sensible name abbreviations in the code, you don't need a separate column for the character name, but including it in the master makes the list more human-friendly. Usually this column is a waste of space, to try to lose it.
• A/R (Add/Replace). This tells you if the line will replace an existing production line or just “sit on top” of the dialogue. Original lines being replaced with ADR require special preparation. Some programs allow you to note the reason that the line was called.
• Call. Whether recording or editing ADR, I like to know who requested that a line be recorded. You may have to drop certain lines in the recording session if there's just not enough time, so it's useful to know who asked for it.
• Start. This is the line's exact start time: if that's an audible breath, Start falls there. (If during the recording session you find the actor is having a tough time getting the timing right with cues that start with breaths, drop them; record some wild breaths and begin each take with text.)
• Stop. This is the end time of the cue, plus about half a second for ringout or a late delivery. Some find this information unnecessary.
• Text. This is pretty obvious. Less obvious is how long a line should be. If the loop line is one short sentence, that's a natural duration. But what about longer text, say a whole paragraph? Remember what you're doing. The actor needs to learn the line and its rhythm, so the spotted lines shouldn't be unduly long. However, cutting text into unnaturally short segments results in an unnatural recording process and can be very time consuming. Plus, you lose the flow of the text if it's acted in very tiny bits. There are a few rules of thumb to remember when dividing text for ADR:
– Break at the end of a sentence.
– Break lines at breaths. Unless an actor is a pearl diver, a breath pause is a pretty natural place for a break.
– Break at an inflection. Most sentences are composed of clauses offset by changes in inflection. Read the sentence aloud and you'll usually know where to break it.
• If a line isn't sync but rather off-camera, behind the back, or in another such hidden place, you can usually spot a longer ADR replacement. Most actors prefer to record complete sentences if there's no concern about lip sync. Plus, you'll likely save time in the session.
Depending on the size of the production, the habits of the supervising sound editor, and the type of software that's used, the call sheets may have more, less, or different information. Figure 15.2 shows one example of the many ways that an ADR management can display and organize ADR information for a project.
Each actor will need a list of his lines in film order. Of course, you could merely provide a sorted copy of the ADR call sheets, but these aren't the easiest things to read, especially in a darkened recording studio. It's better to provide a stripped-down version of the ADR calls, with a minimum of clutter (see Figure 15.3).
You have to take notes and keep track of your takes during the ADR recording sessions. Period. If you don't, you'll face a herculean ADR editing task. And, since there's no time in the session to prepare for efficient note taking, you have to enter the studio ready to write. That's where your prepared ADR recording reports come into play.
Here you'll mark all acceptable takes (the ones at least worth considering when editing) with “Hold,” and for the best take or takes, you'll check “Buy” (or whatever term your local film culture uses to indicate the take). In the comment region for each take you can indicate things like “good ending” or “middle section OK” or “director likes this, I don't.” At the bottom of the
page is a section for notes to the editor (even if it's you), where you can explain complicated combinations such as “Use first three words from TK 7; middle section from TK 1; end with TK 2.” (See Figure 15.4.)
If you don't have a typing maven for an assistant, it's not hard to see the advantages of computerizing your paperwork. Find out what commercial ADR management programs are used in your film community. Failing that, create a custom database, with FileMaker or another database program, which will spit out all the forms you need.3
Figure 15.5 shows a FileMaker data-entry page used to program ADR. This could just as easily be an Excel spreadsheet or even a table in a word processor—anything that allows you to search and sort and print. After compiling an ADR wish list with the supervising sound editor and director, fill in pertinent information about each line. From this you can produce lists for the engineer, the actors, and the production company, and create the ADR recording logs (see Figures 15.6 and 15.7). Much of what's on the form is automatically filled in, so data entry isn't as overwhelming as it appears.
Other important information you'll want to keep track of:
• Sync. I like to know if a shot is on-camera or not. Even before seeing the picture during the recording session I can start planning whether to record the shot wild or to picture. Knowing the balance between sync and wild shots for a recording session can aid in scheduling, since the off-camera lines inevitably take less time.
• Location. If the ADR recording engineer knows in advance the locations he'll be matching into, the session is more efficient. Many ADR studios have wet (reverberation possible) and dry (no reverberation possible) areas to match interior or exterior locations, so the engineer may want you to sort the list by location to make for a slicker session.
• Microphone. The ADR engineer likes to know what microphone was used during the shoot to be able to better match the texture of the production dialogue. You can usually find this information in the sound reports. If not, ask the production mixer.
This is an idealized list; Most of the time you either can't or don't want do create ADR cue sheets with such detail. But this hints at the kind of
information needed. Whether by hand or on the computer, you'll need to produce one set of ADR call lists sorted in film order and another sorted first by actor and then by film order.
ADR sessions were once pretty standardized. You might loop in one studio and record ADR-style in another, but the technology was predictably similar. Today, the workflow in most serious ADR studios is fairly interchangeable but the tools used can differ wildly.
At the top of the ADR studio food chain, you'll find ADR management programs that control the streamers, beeps, and monitors as well as the record and functions. You'll also find automatic voice slates that slug scene and take numbers into each soundfile, and picture being played on one of many types of nonlinear video recorder/player. Very few studios still use tape or mag as their primary recording format, so the recorder can be anything from Pyramix to Pro Tools to Nuendo, or a freestanding multitrack recorder.
At the other end of the ladder you'll find homemade systems built entirely around Pro Tools, in which beeps are prerecorded soundfiles the ADR engineer places in the session prior to recording. Likewise, monitor switching is managed through clever editorial preparation of the guide track or through MIDI. Some workstations, such as Nuendo, Pyramix, and others, offer comprehensive ADR management software within the application.
Some ADR programs allow you to insert the text over the projected video image. For many actors, this is a lifesaver that allows them to focus on their acting rather than being befuddled by the text. The venerated matriarch of this process was known as the Rhythmo Band, a uniquely French process first used on Felix the Cat in 1927.4 Skilled assistants with noteworthy penmanship would listen to lines to be replaced, and then write them on a clear film that would be projected just below the picture during the recording session. Cursive handwriting changed in size and shape to match the original delivery. As moving text crossed a marker, actors knew when and how to deliver the line. Now the process is done with text generated by computers—perhaps more efficient, and certainly more flexible, but much less romantic. Figure 15.8 shows a latter day example of this prompting tool.
You might find that an actor is unable to work in an ADR manner—that is, speaking in sync. You want the line, so you have to be innovative. Consider creating a loop-style session, at least for the lines that torment the actor. Figure 15.9 shows one low-tech way to provide the talent with reference and sample space to speak the line. Inexperienced actors tend to like this approach since it's not so stressful and since he hears the reference immediately before speaking, it's usually easier to repeat the music of the line. However, don't forget that you're recording for film, not radio. Even when the entire room bursts into applause when take 99 is not horrible, calm yourself, drop it into the scene, and see how it sits.
Thankfully, you needn't be overly concerned with the ADR technology that a particular studio boasts. If the studio has a decent reputation for getting the job done and the ADR recording engineer has the experience and talent to capture a good sound match, it's not that important that the equipment is the most modern. People have been successfully recording postsync for a long time. Besides, a dialogue editor doubling as an ADR supervisor usually has little say in the matter of studio selection.
As with the mix, the key to a successful recording session is communication with the ADR studio and particularly with the engineer. Find out what they need to give you the things you need.
Learn a bit about the actors you're going to loop. Any weird habits? Do you have to provide Pop Tarts? Does he drink only Norwegian artisan water? Does she insist that you not be in the studio with her but rather in the control room? Let's not go nuts, but it's good to know whom you're dealing with.
It's common to record ADR with more than one microphone. You may use a boom and a radio microphone to ensure a better match. Or two boom microphones at different distances from the actor to change perspective. Or one boom microphone reasonably close and another in the far corner to provide a natural sense of “room.” These choices will affect the dialogue work flow for the rest of audio postproduction, so talk to the supervising sound editor and the mixer before deciding what to do.
Get to the session early in order to spot-check the programming of the ADR cues and go over the sound issues with the ADR engineer. When the talent arrives, keep your cool. If you're dealing with a “star,” be neither starstruck nor unaware of the imbalance of power. Introduce yourself. Explain what you'll be covering. Find out how the actor likes to work. Sitting? Standing? Does she prefer to work ADR-style or looping-style (or, better yet, a combination of the two depending on the line)? Don't waste too much time “making everything comfortable.”
The best way to size each other up is to get to work. After a few lines you'll begin to understand how to get the most out of the actor. Listen more than talk, but don't be afraid to speak your mind. You're the one who'll have to make all of this material work. Remember that goal and keep at it until you're happy. Don't assume that the actor and the ADR engineer have your interests at heart. Yes, they have a job to do and most likely want to do it well, but everyone is looking to you to know when it's good enough.
When you begin an ADR cue, find the corresponding ADR recording log, which will already include text, timecode, and other recording information. Use it to note your impressions of each take. Also note the director's reaction to each take if she's involved in the session.
If an actor has particular difficulties on a certain line, don't beat it to death. Move on. When you've finished the mandatory lines, go back and try again, this time recording it using looping style on the reprise if you were using ADR the first time around. The results might be interesting.
Be gentle with actors. Never forget that acting ADR lines is horrifically difficult stuff. Try it sometime. It's not easy to walk into a role months after the shoot, and many actors don't react well to seeing their scenes for the first time—they don't like the way they look or they hate the editing—and you may have to absorb some of their disappointment. Be patient and don't buy into the actor's frustration or anger. Your job is to get the line, more or less at all costs, and getting flustered rarely helps. Be kind and polite, but don't let the talent get the better of you. You're in charge.
It's not unusual for an actor to try to talk you out of certain loops. Remember, for her ADR may be as much fun as dental surgery. She'll try to convince you that the underlying noise problem isn't really so bad (tell her that she can't hear the noise on headphones) or that the line requested by the director is stupid (record it anyway) or that she has no problem understanding the corrupted line (it's her voice on the guide track; of course she understands it). Lie, cheat, flatter. Don't be coerced out of a line that you need.
If you're responsible for directing the actor, remember that you're not directing a movie; you're mechanically replacing lines. It's about listening to the guide track—really listening—and getting the actor to mimic it. It's very rare, and almost always inappropriate, to need to dig into method acting (“think about your first kiss” or “reminisce about your long-departed hamster”). Give the actor specific, detailed instructions, usually referring to the guide track. Point out the details that he's missing.
Don't overcue. Of course you want to make sure that you're covered for the mix, but you have to remember you're dealing with actors who in most cases don't want to be there. Prioritize your cue list before the session. Explain to the director and talent that you will hold the less important cues for the end of the session, then go back and record them if there's time left. This will assure you get the best performance for the most important cues that are needed.
Bobby Johanson, ADR mixer
Lincoln; The Sopranos
ADR recording engineers can require a bit of work, too. Of course, they take it seriously and, yes, they're pros. But it's not their film, so they may not be as enthusiastic as you are. Don't be afraid to coax them along, and don't give up if the response to your concern about sound match is “Oh yeah, it's fine.” Trust your gut, even in the face of a more experienced engineer. He may prove that everything really is all right, but don't give up too quickly (at the same time, remember that you may be in over your head, so be prudent). Tomorrow, he'll have another film to work on while you'll be back in the cutting room, kicking yourself for having accepted a sound that you knew was inappropriate.
Most directors will want to be in the ADR recording session. While a reasonable plan, it is not without cost. Remember, before the ADR sessions have come years of anguish for the director—especially a director who doesn't make a lot of films. Aside from the obvious—and not so obvious—filmmaking challenges that have worn down and wrecked him, he's now facing the mortality of his film. He's ignored loved ones, betrayed his body, and alienated his friends—all in the name of this lofty goal. Now the fantasy is dismantling and you're there to share it. ADR is his last chance to bring his baby the breath of life. There's a chance he'll get weird on you. Find a way to give the director what he's looking for, all the while striving to keep the session on time and the actor sane.
When you start a new scene, test how well the new ADR fits into the production dialogue. Record the first few lines. Once you have a good take, tell the actor to relax for a few minutes. Sync the selected take and fill the hole in the production track with room tone from the guide track. Now play the dialogue section against the picture. Listen to the way the ADR “sits” in the dialogue. This is a good time to ask the engineer's opinion. Even though you're in a hurry, spend these few moments to obtain as good a match as possible. If you can't get an almost perfect match between dialogue and ADR in the recording session, you'll never achieve a perfect match in the mix. Check for performance and sync, and remember, ADR is foremost about acting, nothing else.
Pay special attention to the cadence, the “music,” of a line. It's very common for everyone in the studio—actor, director, engineer, and editor—to be convinced that the actor is “nailing” the line when in fact he's misinterpreting its cadence. If the original line is “How now, BROWN cow” but the actor is consistently reading “How NOW, brown cow,” the mistake can easily go unnoticed until you're editing the ADR line, but by then it's too late.
If, despite numerous logical explanations, the actor keeps perfecting the same mistake, try a gimmick you used while editing alternate takes—making up nonsense phrases to describe the cadence and essence of the line. The lack of content reduces the line to pure rhythm and tone, and the silliness adds a bit of levity, which can help you get your point across.
When you finish recording all the lines, review your notes. If you're unsure about a line or two—and you still have time—ask the actor to try those lines again. Ask her if there are any lines she'd like to review. Now that the heat's off, you may get surprising results. When you return to a line, be sure to note the new takes on the appropriate ADR recording log.
There are many software solutions to ADR management and recording. There are film industries worldover. But ADR naming conventions are pretty consistent. Regardless of the method used to wrangle ADR data, the code names with which you christened each line on the ADR cue sheets will become the names of the soundfiles created during the ADR session. Usually, the code number becomes the “root” name of the new file, and the take numbers are added automatically.
If your studio is on the budget end of the spectrum and you can't automate the looping process, you can assign soundfile names by temporarily renaming the record track in your DAW with the root name of the line, leaving it to the DAW to create take numbers (see Figure 15.10). The resulting soundfile will carry the name of the cue. Run a few tests to make sure that your names won't be obliterated or changed when you create and edit new clips from the recorded soundfile.
Rich or poor, remember that it's imperative that the soundfile names reflect the original ADR call names, and that the scrupulous notes you took during the session reflect the realities of the recordings. It's worth the investment of a few minutes to ensure that the names and take numbers on your wellplanned paperwork relate to something real.
Dialogue has room tone. ADR doesn't. When you remove the original (to be replaced) line from the dialogue track, you have to fill the resulting hole with something, typically room tone and motion. Say the shot you're replacing consists of one person sitting quietly in a chair or standing still. Since there's no movement, there's no need to fill the space with motion from alternate takes (footsteps, clothing rustle, loose coins rattling in a pocket, etc.). Just add a bit of clean room tone and you'll be fine.
However, if the shot that was replaced included body motion, talk to your supervising sound editor about whether it will be replaced with Foley or whether you should try to fill it from alternate takes. In general, replacing the missing action with dialogue elements will give you a better fill and will save everyone headaches.
Because a successful dialogue premix depends on flexibility and your ability to move quickly back and forth between the original line and the ADR replacement, you can't simply fill the hole in the dialogue track with room tone (see Figure 15.11). Such a layout affords no flexibility. Instead, you have to construct the tracks so that original dialogue, tone fill, ADR, and X are on their own tracks and you can calmly switch between them during the premix (see Figure 15.12).
In Figure 15.11, notice that all material removed from the dialogue tracks is placed on the X tracks (X, Y, and Z), which are used exclusively for material removed for ADR. Before moving to the next edit, audition this new construction with both playback options:
• Original dialogue →tone fill →original dialogue. Once the ADR is added, this is the “normal” playback for when you want to use the loop line.
• Original dialogue →X track →original dialogue. This option will result in the line in its original form.
It's important that the fades for each option work properly, since you never know which option you'll use in the mix (see Chapter 12 for more on choosing the correct fade type).
Please don't build every take of every ADR line. Pick the best one or two readings. If we have a problem with those we can dig deeper if necessary, but when all of the takes are there at the mix I always feel obligated to listen to all of them, and that can become very time consuming.
Tom Fleishman, rerecording mixer
There are many ways to create the universe in which to edit your ADR. Differences in equipment, local traditions, supervising sound editor preferences, vagaries of the production, as well as personal taste will influence how you get started. But until you find a system that works well for you, try this.
Open a reel of your most current dialogue session and make a copy (“Save As”) with a new name. Beneath the active dialogue tracks (for example, Dial A-Dial M), open four or more new mono tracks—label them “ADR A-D” (or more). These will be your final, edited ADR tracks. How many depends on the density of the ADR, the number of characters, and the preferences of the mixer. Below the new ADR tracks add about ten new mono tracks, labeling them “ADR Work 1–10.”5 (See Figure 15.13.) These are the tracks
onto which you'll initially open your ADR lines— safe places to work without endangering the active tracks. You can delete them when you finish editing the ADR.
Normally you'll edit the ADR by film chronology, rather than by character, since it usually makes more sense to hear both sides of a conversation. If you go this route, organize your paperwork so that you can keep track of the cues by both film order and character. This reduces the chance of error.
If you recorded ADR with more than one microphone, you are faced with the decision of editing and delivering the loops as stereo or mono tracks. The advantages and disadvantages of each are clear. I usually prefer to edit on separate, grouped tracks. Since multiple mics of the same recording carry different information—and different problems—it can be misleading to edit them as one. I tend to roughly cut the mics as a group, and then nip and tuck separately. I can then deliver the tracks however the mixer prefers. Decide how you want to work before you build your tracks and begin editing.
Import the play lists of the ADR recordings, each character with their own list. Organize your note filled ADR recording logs and you're ready to edit.
If an ADR line has a clearly indicated selected take, slip the line directly onto an ADR track if you anticipate an easy fix. If you foresee trouble with this loop, move it to an ADR work track for more serious manipulation.
“Top and tail” the cue, leaving just the line and a bit of ringout afterward. Move this trimmed region to a track just below the dialogue element you're replacing. You may have to temporarily rearrange your tracks. Do your best to cut around headphone leakage that occurred during the recording session.
Sometimes you can align the beginning of the dialogue cue with its ADR replacement and begin making sense from there. Other times the initial attack isn't sufficiently clear to be used as a reference. In such cases, analyze both waveforms and look for notable landmarks. Plosive consonants such as B and P are usually easy to spot, as are stop consonants such as T and K, because they typically rise above the rest of the waveform, begging to be used as guideposts. Find common landmarks in the dialogue and ADR regions and align the two regions from either head or tail, whichever suits you. If they align well for their entire length, consider yourself very lucky (see Figure 15.14).
Listen to the reference and the ADR replacement and make sure the meaning, melody, and attitude of the loop match that of the original. Then watch the replacement line for sync. Watch it again. If you're still happy, move on. In the ADR recording log, mark that the line is finished. Prepare the dialogue line to accept the ADR (discussed in the previous section), and remember that you got off easy this time.
More than likely you won't be so lucky. Most ADR lines need editing, often necessitating a combination of takes, time expansion/compression, pitch shift, and other tricks. It seems a formidable task to compare numerous outtakes and combine them in a way that honors the spirit of the take that you or the director selected in the recording session. Yet with a standardized plan you can extract the best parts from each take without creating a soulless Frankenstein's monster.
Arrange your tracks so that just beneath the reference dialogue track is the ADR track where you will end up putting the cue (ADR A, B, C, etc.), and below that are your ten or so ADR work tracks (see Figure 15.15). Compare the chosen ADR take to the original dialogue. Figure out why it was selected and why it's not working. Is it a sync problem? If so, where does the ADR line fall apart? Does the ADR cue regain sync after the sync-defying irregularity? Determine how much of the chosen ADR take you can use and how much will need to be replaced from alternate takes. This will limit your search.
If the selected take suffers from more than just a sync problem (maybe a section of bad diction or perhaps a vowel held much too long or not long enough), decide what, if anything, you can salvage from it and note what you're looking for. You might be looking for an unusually short “now,” followed by a “brown” in which the vowel is held while the pitch rises.
“How now' BROOOOWN cow.”
Find the qualifying alternate takes (see Figure 15.16). Before importing candidates, listen to them and cull the ones that don't have what you're looking for. No point clogging up your session. A search engine comes in very handy when wading through ADR takes.
Line up all your alternate takes onto the work tracks beneath the reference dialogue line. Listen once to each take just to make sure that your recording session decisions were sound. Maybe, just maybe, one of the alternates will have everything you need. Probably not, but you might as well listen.
When you're piecing together a line from alternate takes, it's best to sort out one problem at a time. Otherwise, you'll easily lose focus and fall victim to the fantasy of quick fixes. From the information provided by the waveform, pick the most plausible candidates for the word or section you want to replace and listen to each. Select the most likely candidate and edit it directly onto
the buy take. See how well this works. Next you must find a way to splice the replacement into your line.
You've already learned the usefulness of consonants such as T and P and B. Add to that list the sibilants, such as S, Sh, F, and Ch. Use the consonants as visual benchmarks to tell you where you stand. Use the sibilants as convenient cutting locations within words. You can almost always cut within a sibilant sound—to shorten or lengthen—and get away with it. When you're anticipating a difficult ADR editing session, hope for lots of sentences like “Sally sells seashells by the seashore.” Of course, as convenient as sibilants are for editing, you'll pay for that convenience in the form of de-essing during the dialogue premix.
Most of the sound shaping for matching ADR into the dialogue (EQ, dynamics, reverb) happens in the dialogue premix, so you needn't bother with that just yet. Your concern is mainly intonation, matching, sync, and performance. Performance trumps all other concerns.
When combining takes, you may find that one is pitched higher or lower than its neighbor. The result is an unconvincing sentence that doesn't make sense. You may be able to whip it into shape with selective pitch shift. Select that part of the phrase whose pitch you want to change. Take a bit more than you need by pulling out the handle and selecting so that you'll have a pitch-shifted handle available for crossfades. Many workstations auto - matically create processed handles when working offline.
Keep in mind that a little pitch change goes a long way, so experiment with the parameters until you have the intonation you want.6 On most plugin pitch-shift processors, there's a button called something like “Speed Correction.” If you want to change the speed and pitch simultaneously so that the processor works like an analogue tape recorder (faster speed with higher pitch or slower speed with lower pitch), deselect this option. Remember, if you don't like the initial results, go back to the original (unprocessed) file. Leave an already processed track alone to avoid accumulating artifacts.
Time expansion/compression processors don't necessarily work linearly, and as you have read in Chapter 14, they present a constant compromise between sync and rhythmic consistency. One way to get around this unsavory compromise is to break the original region into smaller sections. When you dissect a sentence like this, you'll find legato sections and staccato sections, and you'll also realize that there are probably stretches that require no processing; they're just surrounded by words of the wrong length. Don't process the sections that don't need it, and treat the staccatos differently from the legatos, adjusting the “quality” setting differently as needed. Piece together the results and, odds are, you'll have a glitch-free line that has good local rhythm.
Here's one way to safely and systematically use time expansion/compression to make the inserted alternate word—in our case, the word “now”—a bit longer. Remember, always begin by checking other takes for these very specific replacements. As sexy as all this technology is, nothing beats the real thing.
• Move the replacement ADR word's region—in sync—to a free track. This is the region you'll process with the time expansion/compression plugin.
• Copy this region—again in sync—to yet another track. It's useful to have an unprocessed copy standing by.
• Expose a bit of handle from either side of the region you plan to process. Since AudioSuite plugins create a new sound file, you'll want to process more of the region than you need. Without handles, your editing and cross fade options will be very limited.
• Select the region and open the time expansion/compression plugin. There are scores of these available, and there's a good deal of debate as to the merits of each. Ask around, experiment, guess. Or just use the one that's in the workstation you're working with (see Figure 15.17).
• Choose a time expansion or compression ratio to change the length as needed. There are two schools of thought on how to calculate the ratio desired. By switching the workstation time display from timecode to minutes and seconds, you can accurately compare the length of the original phrase with the ADR line you're time-stretching. Then you can calculate the ratio that you'll type into the “Ratio” cell of the time expansion/compression plugin. Frankly, I find all of this a bit tedious. I prefer to give it a try with the time stretch tool (every DAW has one), line up the result, see how I've done, and undo if my guess was wrong. Then I try again.
• If your chosen ratio works well for one part of the word/phrase/region but not for another part, you might have to perform several time expansion/compression operations. This is where the safety copy you created comes in handy. Let's say you expanded a region by 4 percent to match the original. It's a good speed match except for a short section in the middle that's running slow. You'll be tempted to time-flex that section since it's sitting right in front of you. But then you'll be processing an already abused sound file. Time expansion/compression is a nasty business, hardly free of artifacts. So instead of recooking the processed region, go
to the safety copy and see if it solves the problem—part of the ADR line may need no time expression/compression at all. If you need to squeeze specific parts of the line at different ratios, this original region will give you far better results.
• If things aren't working out, take another listen to the other acceptable takes. Once you've started down the processing path, it's easy to forget that better, low-tech options may be available.
If you're satisfied that you resolved the first problem (in our rather simple example, the word “now”), move on to the next issue (“brown”). Follow the same steps. Rarely do you need to go word by word, and it's preferable to find the longest workable word strings to create a more natural flow, minimize the chance for weird edits, and lessen your work. Still, sometimes you have to tackle a line one word at a time.
Piece together all of the resulting regions—newly processed sections intercut with pieces of the raw ADR take. You may encounter some bumps, so don't be afraid to move the edit points a bit earlier or later than the obvious word beginnings. Music editors often use the downbeat (in our case, the beginning of a word) as an alignment reference point but make the actual cut someplace else. In dialogue and ADR you can use word beginnings as sync reference points but you may find that cutting at other, less obvious places gives you better results. Greater flexibility with edit points is the payoff for having pulled out some handle from the region before you performed the time expansion/compression.
Throughout this syncing process, listen regularly to your creation. Waveforms are helpful for syncing and matching, but film dialogue is about sound. It's quite possible that you'll put together a collection of regions whose waveform pattern perfectly matches that of the reference, but the result will be nothing familiar to humans.
Once you're happy with the construction, you can give the whole phrase a nice nip and tuck with word-fitting software. These programs work like time expansion/compression but automatically and on a much smaller, more dynamic timescale.
Word fitting processors enable you to automatically match an alternate dialogue take or an ADR line to a reference phrase (almost certainly the original). All you need to do is select a processing range on the reference and replacement regions; the program then compares the respective wave - forms and locally lengthens and shortens the replacement, ultimately producing a new sound file, in sync, on a different track. A popular word fitting program is VocALign, available as a freestanding application or as a plugin for a number of workstations (see Figure 15.18).7
Word-fitting processors are at the same time miraculous and dangerous. In experienced hands, they quickly coax loops and alternate takes into sync, yielding huge time savings. They are also fantastic at applying a final sparkle to delicately edited ADR lines. Yet scary things happen when inexperienced or lazy editors blithely charge forward with such tools or when miserly producers use them as an excuse for their cheapness.
So how do you most effectively use word-fitting programs? First you edit, constructing the best possible, most natural-sounding phrases. You're probably (but not necessarily) better off building a line from several takes than heavily processing the one selected. If you need to expand or compress
a line or part of one, do so. (I'm not claiming that time expansion/compression sounds better than VocALign; it doesn't. It's just that doing your homework first allows VocALign to do its best work.) Piece together the best possible sentence and then apply word fitting. Since word fitting uses the same logic as does time expansion/compression, you can use the “Flexibility” control to compromise between rhythm and quality.
The processor will do some surgery that you couldn't do yourself, and since you're not asking it to do the impossible you ought to get tight sync without artifacts. While word-fitting programs can be very effective at syncing an ADR line or alternate take to a reference clip, you're still left with the matters of pitch and inflection. Certain processors whose roots are in the music world, such as ReVoice Pro, were designed to create double tracks in musical performances. They can take some of the pain out of ADR editing by tracking vocal attributes other than just timing.
Be sure to intelligently name the resulting region beginning with the ADR code number. Also, place the pre-word fit string of ADR edits—muted—on a track that will travel to the mix. You never know when a previously unheard
If you have a good ADR take but the line is too long, you can often make an edit within a word to put things in sync. In Figure 15.19a the top track is the production line or guide track. Beneath it is the ADR clip that I'm trying to sync to it. First step: copy the ADR clip to the track below.
Decide how much you need to trim from the ADR line and move the copied clip accordingly (see Figure 15.19b). Note how I've tried to match the bottom waveform graphic so that it lines up to the original guide.
Zoom in very close so that you can see the actual waveforms. Note that in Figure 15.19c the waveforms on Edit 1 and Edit 2 are different, out of sync with one another.
Move the copied clip (bottom) so that the waveforms match up, as shown below in Figure 15.19d.
They are now locally in sync with one another.
Next, move the copy up onto the actual ADR track, as seen in Figure 15.19e. Use a very short crossfade to heal the join.
Listen to what you've done and make sure it sounds natural. If it's okay, you're almost finished. If it doesn't sound right, put the join somewhere else.
Tidy up the clip with head and tail fades and you're done (see Figure 15.19f).
Jenny Ward, supervising dialogue editor
The Great Gatsby; Happy Feet Two
processing artifact will surface in the mix or when, as you watch your manicured ADR construction for the first time on the big screen, you'll realize that a word is out of sync. Having the unprocessed string immediately available will make for faster fixes.
Looping a very noisy scene is hard work. The scene may be so noisy that even transcribing the original lines for your ADR call sheets becomes a test of hearing acuity and psychic prowess. The recording session will be difficult because the actor has trouble hearing herself in the guide track, which certainly gets in the way of artistic expression and results in frayed nerves.
Unfortunately, the headaches don't end with the recording session. When you can't hear well and can't discern details within the waveform, you're working with at least one hand tied behind your back. The same goes for word-fitting processors. If the signal-to-noise ratio on your reference track is so bad that the processor can't make wise decisions, the results will show it. Soft transients—the gentle beginnings and ends of phrases—will suffer the most.
To better hear and “see” the reference track, process it with a high-pass filter and perhaps also add a boost at around 2 kHz for better voice articulation. Be brutal. You don't care what the reference sounds like; you just want better resolution to increase the signal-to-noise ratio and provide more detail in the waveform. Ugly, maybe, but you should be able to hear more voice information and VocALign will be able to match waveforms more accurately.
Prominent breathing sounds are usually spotted like any other ADR line. They should be treated as such. However, in many cases an entire shot or scene will be looped in one pass. Not only is this a near-impossible feat for the actor, but it can lead to a bit of laziness during the recording session and while you edit. There's a temptation to lay up the best take, see if it basically works, make a few adjustments, and call it a day. But just as thoughtfully edited breaths bring life to a scene, so do sloppy respirations make the scene tacky. Select the take that best captures the mood, and then get it to work using the other takes. Treat each take as raw nonsync material. So if at the moment you need a soft sigh of a certain type, trawl the raw takes until you find just the right one. Do this for the entire shot. Editing breaths may be one of the most difficult—and most rewarding—parts of your job.
Not all postproduction dialogue recording involves principal characters. Often you need to record the other voices in a scene, whether to enhance the plot, to say more about a principal character, or to add mood and texture. This process is called group loop, and it involves contracting voice actors and a supervisor to study each relevant scene, come up with appropriate dialogue, and record the result in a studio.
Group loop recordings can roughly be divided into three types:
• Specific group loop
Walla is the American term for the indistinct “buzz” created by a background group. The classic example is barroom chatter. Imagine your two main actors sitting at the bar having a beer and discussing the meaning of life. Behind them, seen and unseen, is a crowd of fellow drinkers doing whatever you do in a bar. During the shoot, the extras are only mouthing their lines lest they interfere with the principals' dialogue, so aside from the words of our guys in the foreground, the raw scene is eerily quiet.
You've already looped any critical lines from the principals. What's left is to fill in the background and define the mood—how big and rowdy the crowd and perhaps some clues about class or region. For this, the supervising sound editor—and on jumbo films a group loop supervisor—will spot the scene to find out what's important in the background sound. Are there any critical plot issues that need to be dealt with through off-camera crowd commentary? Are our protagonists noticed by the crowd? What's the mood? This process continues for all relevant scenes. They will come up with a plan for each scene and then hire a group of actors, each of whom can play many roles. Obviously, the group loop talent can't be any of the principal actors in the film, as their familiar voices would confuse the viewer.
During the group loop recording session, there'll likely be stereo and mono passes for the general walla of the scene. The group can be made larger by recording successive passes in each of which the actors change their positions and voices enough so that in the final multipass recording you don't hear them standing simultaneously in several places across the screen.
Group walla is largely improvisational. In our barroom scene, the group loop supervisor might instruct the actors to talk about an imaginary game between the Cowboys and the Redskins or about how much they hate politicians or to complain about this year's pistachio crop. It doesn't matter as long as they don't say something so present-day that it will soon be dated (such as this year's presidential election) or mention something about the film.
The supervisor has to balance the actors as a conductor balances an orchestra. If one or two actors' voices rise above the crowd, it's distracting and makes the track hard to recycle—the ideal walla track should be a relatively smooth din, but with enough language clues to make it worth recording rather than digging into an SFX library. And if the group is speaking in a language neither you nor the supervisor speaks, you have to find a native speaker to do a “stupidity check” to ensure that the seemingly benign walla isn't filled with “I curse your ancestors” or “I really hate this director.”
While much of group loop consists of background chatter, there are occasions for specific mono group loops. The waitress bringing a beer to the customer at the table behind our protagonist might utter “Here you go, hun.” There may be off screen lines for the bartender or for the couple who walks past. All of these lines have to be written and assigned to a member of the loop group, and they're recorded much like normal ADR or as wild lines. Group loop supervisors often develop superhuman skills making up lines that match an actor's actions. Sometimes, a lip reader will be called in to figure out what a background extra is saying so that a sync line can be written.
Callouts are yells mixed into the track at a very low level in order to make an environment feel bigger, or to influence mood. Hearing a disembodied voice from way down the beach yelling, “Hey Chico, throw me the ball!” enlarges the space and livens up the atmosphere.
Editing individual group loop lines, such as those of our friendly waitress, is no different from ADR editing, except that you'll rarely have a reference track. Since extras on a shooting set usually don't make any sounds—their lips just flap—you can't rely on the guide track for sync or for content. Also, with group loop you'll rarely encounter the detailed paperwork that accompanies ADR.
To do justice to group walla, you need to keep a couple of things in mind. First, just because the crowd recordings were taken in a certain order, you don't have to use them that way. Some sound events have to be in sync, such as the cheer of the barroom crowd when the Cowboys (or Redskins) score, but many of them can be used in any way you like. Second, listen to the production dialogue and find appropriate moments for that small burst of laughter or the muffled “Darn governor, he's a crook!” in the background.
You can use these lines to enhance the rhythm of the dialogue or to motivate an actor's twitch or even to make an omniscient commentary. At the same time, if an event in the group loop conflicts with the dialogue—one particularly ugly outburst or a shout that hits at just the wrong time—either move or delete it. If you're cutting the group loop, it's yours to play with. Keeping the tracks exactly as recorded is boring and wimpy.
You may have to repeat a group recording to make it last longer. As when editing backgrounds, you can't merely copy and paste. Open the recording onto a work track and from there build a new sequence. Much of the sequence can be random, but pay attention to what's going on with the picture and in the dialogue to most effectively use the recording. If there are any recognizable beacons, such as a telltale laugh or chair squeak or a notable bit of background dialogue, don't use that moment more than once, even in different scenes.
Don't lose touch with the desired size of the crowd. Groups often sound better, richer, when “doubled up”—that is, when takes are piled atop each other. If this works, great. However, if there are only eight people in the small restaurant, it will sound unnatural when you overlay four group takes, resulting in a crowd of 30 or more.
It's important to know what the mixer expects from your ADR track layout. Should you prepare alternates? If so, how many? Should they be on “active” tracks or hidden away? How should you handle ADR to be panned? There's no such thing as “the right” way to lay out your ADR; it depends on the production, the budget, the supervising sound editor, and the mixer. There's only one “wrong” way: not asking what to do.
One way or another, the ADR lines must be mixed into the production tracks. Loops have to coexist with their production track neighbors, assuming the energy, pitch, tone, and (most importantly) the human essence of the original dialogue. The bulk of this trickery takes place in the ADR recording session. An ADR line that's unconvincing when recorded stands little chance of salvation in the dialogue premix. What, then, is important when recording an actor?
• Performance. It's all about performance. A technically perfect recording is doomed to fail if the actor can't get it right. Cadence, diction, pitch, energy, and emotional message must be right to get a useful take.
• Microphone selection and placement contribute to the color of the actor and the room. All this affects the match with the dialogue.
• Other electronics, such as microphone preamps, console, and any other devices in the chain will influence the voice match.
• Acoustics. Room size, ceiling height, ambient noise, “dryness” vs. “wetness,” and other acoustic issues will determine how the ADR fits into a scene. Acoustics can be controlled with reflective or absorptive panels; many ADR stages include a moderately wet, adjustable area for interior scenes, as well as a dry tent-like area suitable for exterior shots.
• Pitch. When the dialogue editor cut the ADR, she probably manipulated the pitch of certain clips in order to create a life-like phrase. Perhaps she changed the pitch of an entire line to achieve a better match with production sound. So it's likely that further pitch tinkering is not needed in the premix, but you never know.
• Performance. It's worth saying again.
Once recorded, chosen, edited, manipulated, and synchronized, it's up to the rerecording mixer to convince the ADR cues to get along with their older neighbors. As budgets collapse and workflows adapt accordingly, dialogue editors must occasionally premix their own tracks, delivering a “dialogue package” that's ready for final mixing. Like it or not, it happens, so a superficial understanding of the tools of ADR matching is in order. However, just as all production recordings are unique, so, too, are ADR lines. There are no “one-size-fits-all” solutions.
• Equalizer. Most ADR recordings need some sort of EQ adjustment. ADR is often bottom-heavy when compared to original production tracks. A high-pass filter will reduce much of the weight, but be careful not to rely too heavily on it or you risk taking all life out of the bottom. Often what sounds too dark—or too bright—needs to be addressed by a narrow bandpass cut, rather than an all-or-nothing high- or low-pass filter. Get the general shape right, and then pay attention to the density, or lack of it, at specific frequencies when compared to the original. Note particularly microphone peculiarities that add unwanted midrange boosts. Create a narrow boost and shop around to find the troublesome frequency. Then turn the boost into a cut to reduce the problem with minimum damage.
• Reverb. ADR is dry (well, pretty dry), so it's likely you'll need to make interior scenes a bit wetter. This is trickier than one would think; despite a good EQ match, the reverb can give away the loop. It's easy to create a reverb too large for the scene, whose reverb tail is too long and too bright. Shrink it, shorten it, and dull it. This may not work, but in my experience it's a good starting point. There are two basic categories of reverb processors: synthetic digital reverb, and sampled-acoustics convolution reverb.
• Synthetic reverb processors. You're undoubtedly more accustomed to synthetic reverbs. These devices consist of two engines, one creating early reflections, the other, reverb tail. The components can be controlled individually or together to mimic most spaces where you'd like to plant ADR. However, there's one trap: the factory presets. A preset called “Bronx Zoo Lion Cage” is a logical choice when you're trying to match into the scene where the zookeeper struggles to free his assistant from the feline's jaws. This preset may work, it may not. Don't confuse a preset's name with what the processor is actually doing. Synthetic reverb presets—and their names—are dreamt up by clever engineers at the factory. They are combinations of settings that produce a reverb that resembles the idealized sound of a space. Nothing more. This means that you must learn to look past the presets and discover what the controls actually do. Otherwise you are a slave to the well-intentioned but misleading preset names.
• Convolution reverbs. While synthetic reverb processors give you the tools to fabricate almost any space you'd like, convolution reverbs actually place your dry ADR signal into a known space. An impulse response is a detailed description of a space, created by recording the reflections of a noise (a sweep or a loud noise, such as a starter's gun) in a space. This impulse response is convolved (entwined) with the sound you want to process, and voilà, your ADR is in the room, whether Carnegie Hall, the Hollywood Bowl, or the location where a scene was shot. This doesn't always work, and it makes sense only if you have lots of lines in the same space. But if you have the time, the means to make an impulse response, a cooperative location crew, and a good convolution reverb, you can get very good results.
• Dynamics. If your production sounds are compressed, for whatever reason, you'll need to match into it. Also, your ADR tracks may be dynamically out of control, so you may need some compression to contain them.
• Distortion. Sometimes, no matter what you do, the ADR tracks are simply too clean. Try running the track through a noise reduction processor. A guitar amplifier simulator is another decent source of distortion, as is Digidesign's “Lo-Fi” RTAS plugin. A little of this goes a long way.
• Pitch. As mentioned above, some pitch adjustment may be needed in the premix. It's often easier to hear faults like pitch mismatches while in the mix room. Not only can you hear better than in your studio, but you're in a different state of mind; you can listen from a new point of view.
If you're the one to mix ADR into production sounds, do your homework. There are countless discussions—often heated—in the audio postproduction Internet forums. Most importantly, next time you're in a dialogue premix, pay attention to what the mixer is up to when it comes to ADR matching.
1. For the rest of this chapter, let's assume that you, the dialogue editor, are taking on most of the ADR duties.
2. Most people know that Word processes text, Excel processes numbers, and FileMaker does the same with data. However, there are many good alternative products that do more-or-less the same thing. In this section I tend to refer to each of these types of processors by their well-known trademarks, so as to avoid confusion and too many words.
3. To perform timecode calculations in FileMaker Pro, you'll need the “Timecode for FileMaker” plugin. This shareware product is available from Belle Nuit Montage (www.bellenuit.com).
4. See www.emc.fr/pdf/doublage.pdf
5. As you saw in Chapter 10, I use a convention in which all tracks to be used in the mix—the “active” tracks—are sequenced with letters (Dial A-M, ADR A-D, etc.), whereas temporary tracks are identified with numbers (junk 1–6, work 1–4, etc.). This makes it easier to identify tracks during editing and allows anyone who knows my system to prepare my session for a mix. This isn't necessarily a “standard” system, but it works well for me.
6. One reason to be cautious when changing the pitch of a line is that most pitch-shifting processors change the formants along with the pitch. Formants, the peaks in the vowel spectra caused by the various articulators in the throat and mouth, enable recognition of vowel sounds and hence language itself. Pitch variations in a voice can be quite substantial, yet the formants will remain consistent. A small change in formants will make a recording sound like the wrong vowel, or the wrong language.