Attention, did you notice that this article is not called "Stereo Sound Mixing for Film/Video." That is because just about everything that you record on the set will be monaural, even though the headphones, mixer, and recorder all use the term stereo.
You may have two (or more) recording tracks, but what you will be recording is two (or more) monaural tracks, not a "stereo" track.
The term stereo refers to capturing the left side and right side components of a sound, so that when they are reproduced out of a pair of left side and right side speakers, or headphones -- the listener will have a sense of left and right balance or space. Sounds may originate on the left or right of the listener, and may even travel ("pan") across from one side to the other, or just park somewhere in between hard left and hard right.
Stereo is what happens when we listen to music, or a fully mixed movie. It is the finished result for the enjoyment of the audience.
But stereo is rarely recorded as such in the field. Instead, we record monaural sounds and wait until post-production is nearly complete to re-assign these sounds to the audience's left, right, and in-between.
Until the film is edited, there is no way to know just where all of the audio elements need to end up. For instance, out on production, it might seem logical to record a car that passes from left to right in stereo, so that you can hear the "pass by" in your phones whoosh from the left ear to the right ear.
If the movie only consisted of that one angle of the car, with no intercuts, then your stereo track would be quite usable. However, in the real world, that car-by is likely to be covered from a multitude of camera angles, and then intercut with a multitude of close-ups and insert shots. Many (if not most) of this additional footage will represent screen center (as opposed to the wide shot of the car travelling from screen left to screen right).
As the finished sequence is built by the editor, it might start with a long shot of a car coming towards the screen, some left to right angles, maybe some more dead-on angles, some shots of the driver, inserts of gas pedals, shifter knobs, angles on the passenger, sheriff's car lurking, woman with baby carriage about to cross the street, and….
You can see where this is going; it is certainly not a simple pan the sound from left to right situation. The editor would not want the audio to ping pong rapidly from screen left, to center, to left, to right, back to center, etc. The flow of the audio has to be paced with the changing perspectives of the shots; with full audio pans carefully and strategically employed for full effect, but not as meaningless distraction!
The solution is to record all of the sounds as monaural (center screen); and to let the editors re-mix and re-assign those sounds to various surround positions during post-production. Also, keep in mind that professional films are not two-track stereo presentations, but are at the least, surround sound six-track (aka 5.1 surround). Audio has to be positioned amongst screen left, screen center, screen right, surround rear left, surround rear right, and sub-woofer effect!
Therefore, if we are recording in two-track in the field, we do not think in terms of left track and right track. Instead, rename these connectors (in your head) as Track One and Track Two. You have two places to store the recorded audio; and the audio in these two places should be different, so that the editor has maximum flexibility to select/adjust one track over the other.
Professional mixing panels intended for film/video applications often lack variable pan settings. Instead of a rotary knob, the operator is given a three-position sliding switch (track 1, track 2, or both/mono). There is no such thing as "panning" a mic so that it is somewhere in-between!
If your mixer has a fully variable pan pot, then get into the habit of assigning each input to either the full left or full right positions. Never allow an input to "straddle" both tracks, since it would make it impossible for the editor to make independent adjustments.
There is one exception to this rule: stereo background sounds. Because all of the dialog and action sound effects will be recorded as single track (monaural), editors often like to work with stereo renderings of a constant background ambiance that they can mix under a complex scene. For example, a beach scene may consist of numerous dialogue shots from several angles, but a constant wild track of beach noise (waves, surf, wind, seagulls) could be played continuously under the sequence to give it some sense of continuity and temporal frame.
Even though the camera angles may jump around from character to character, the editor will take some poetic license and anchor the stereo background effects to give the scene some stability in terms of speaker placement. Sync dialogue will be anchored (for the most part) at screen center, except for some isolated off-camera "wild lines" that may show up as surround, since the characters creating those lines do not suddenly appear somewhere else on screen a few frames later!
So what audio should go on which track of your two-track mix? It is really up to the sound mixer; there are no hard and fast rules. However, as much as possible, try to be consistent -- it really helps the editor or the film lab (transfer house). Keep good notes (aka sound reports).
Think like an editor: What audio elements are most likely to edited or kept as a group and what elements should be kept separate?
If you are mixing most of the scene with a boom mic, and also have a plant mic or lavalier working -- then it might make sense to separate the mics that are most likely to overlap and create a phasing issue. When two mics both hear the same audio, unless one mic is faded down, they will interfere with each others sound waves and create hollowness & echo.
By recording these two mics onto separate tracks and NOT blending them onto the same track, no interference will occur during the recording on set. However, the editor will have to be careful in post to avoid phasing when the tracks are mixed together, but during post one has more time and control to get things just right.
If two actors are both on lavalier or radio mics, it is a good idea to record them onto separate tracks. That way, if one lavalier gets hit with clothing noise or static, at least it won't ruin the other actor's audio.
When there are uncontrollable sound effects, plus dialog, it is logical to keep the dialog mix separate from the effects. The effects are more likely to distort; or the editor may want to dump the location sound effects and replace them with something better later on.
If you have to record an actor who tends to be unpredictable in his/her performance/volume (in other words, goes from whispers to shouts without notice) -- use both tracks to record the same dialogue. But offset the volume of the second track, so that if the actor shouts loud enough to distort on Track One, the editor can simply cut to the non-distorted, lower record setting of Track Two. It is like bracketing your exposure in Photography: at least one of your settings will be optimum for an unpredictable situation.
The same approach works well for recording loud sound effects.
We've covered some of the basics of two-track field recording. Now let's briefly examine some four-track situations.
To begin with, picture editors often prefer two tracks rather than four, six, or more audio tracks. Why? Don't more tracks mean more control for the editor? Well, yes, IF YOU ARE A SOUND EDITOR.
But for many "picture" editors, there just isn't enough time to deal with multiple soundtracks while the Director and studio are exerting insane pressure to see a "completed" rough cut of the movie! Many non-linear editing systems can easily handle two embedded tracks of sound, but increasing that number to four or more takes much more handling during capture/transfer and timeline management. Although a good editor could handle that if he or she had the time… (can you spell "ulcer?")
Which is why the great deity of the universe created Sound Editors! After the Picture Editor has massaged the timeline to the stage where the Director and Studio Henchmen have decocked their pistols, the audio portion of the project is turned over for "sweetening". It is during this sound editing phase of post-production that poorly recorded (or poorly spoken) dialogue gets repaired, replaced, looped, or even hidden under sound effects and music!
How does our multi-track recording scheme fit into all of this?
If we are shooting in HD video, and recording straight to the camera, chances are that the editor is NOT going to have the time nor the opportunity to deal with four independent soundtracks. Two tracks is more than enough for that stressed situation. Four tracks just becomes more of a creative burden, rather than an asset.
But the Producer counts the four inputs to the camcorder, and wants to know why we are not recording in four-track. Never mind that the budget probably does not cover the increased expense of a four output mixing panel, nor additional soundcrew and equipment to take advantage of the four tracks.
Remember how we learned to "bracket our exposure" by mixing one mic feed into two audio channels at offset volume settings… Do that with two tracks of audio output! Send Track One from the mixer to Inputs One and Three of the camcorder, and offset the volume. Do the same trick with Track Two from the mixer, going to Inputs Two and Four.
This will make the Editor happy by not having to do as much audio wrangling/mixing during the basic edit; and provides two alibi tracks in case of unexpected overmodulation/distortion. The Producer sees all four of the HD inputs being used; and the Editor doesn't have to roleplay as Sound Mixer.
What if this is not a low-budget HD shoot? In that case, odds are that we will be recording our audio "double system" onto a multi-track digital recorder. A Denecke timecode slate will be used to provide a timecode reference as well as a failsafe clapstick sync mark for lining up the production audio with the production picture.
If the production company wants audio recorded onto the camcorder as well, it would be for the purpose of screening "dailies" or playbacks; as well as for the basic rough cut of the movie by the Picture Editor. When the picture cut is approved, the project will be turned over to the Sound Editors for cleaning up the soundtrack. At that point, they will want to replace the two-track version of the soundtrack (that we recorded onto the video) with the soundtracks that were simultaneously recorded in multi-track onto our audio recorder.
Some Sound Mixers will deploy two audio recorders on their soundcarts during production. A two-track with timecode, for recording audio for "dailies" and the "picture cut"; along with a four (six, or more) track multi-track timecode recorder to lay down "isolated" tracks, intended for the Sound Editors to use later on.
The rationale that these Production Sound Mixers give for using two recorders as opposed to just one "mega-track" is that they want to keep it all simple and uncomplicated for the lab and the picture editor. If they were to turn in all the tracks at once, there is a greater chance for third parties to muck it up and transfer the wrong tracks (iso's rather than the basic two-track "dailies" mix) or to improperly take advantage of the isolated tracks. By recording onto two machines, the Production Mixer can insure that the basic mix goes where it is intended, and that the complex "isolated tracks" do not end up where they are not intended!
In conclusion, before you record in stereo or multi-track, try to put yourself into the editor's chair. Avoid recording the same information onto multiple tracks. Instead, take advantage of the track separation in order to provide the sound cutter the maximum creative flexibility.