November 1, 2003

It's All in the Mix: The Secret of Sound Design, Part 2

From Carl Sagan, Ann Druyan and Steven Soter's <em>Cosmos</em>, for which Kent Gibson designed and mixed the sound.

Last issue, I discussed sound design and what you will need to know when considering this important element of the post-production process. Part and parcel with the sound design is the sound mix. The final mix can be a joy as you watch filmmakers hear and feel their movie gel for the first time. Ideally, by the final mix, the filmmaker has already met and discussed his vision with the mixer, and tracks are fully edited and prepared. All that is left is the final mixing, balancing and shaping of the soundtrack, as well as the preparation of final delivery elements.

Following are some things you should be aware of when you go into the mixing process.

Mixes are done in many different kinds of sound studios—often called a "mixing stage" or "sweetening bay." Most frequently for documentaries, re-recording mixes are done in a normal sized room with one or two mixers, and perhaps one or more "recordists" or assistants in a machine room. Multiple machines are typically synchronized using time code and computers. This setup has been termed a "video mix" or "sweetening," however, I do not favor those terms because they demean the process. Mixes can be done for film or tape in any sized room, and often the tasks performed are more important than the term "sweetening" implies. More elaborate mix rooms may be theater-size, and have more assistants. Costs for mixing might range from $200 to $3,000 per hour, so there is quite a range of options. It is wise to consult with your sound house in advance to make sure you are mixing in a room appropriately sized for your project.

Sound tracks are most often prepared on multi-track digital storage systems such as DA88s (digital 8-track tape), DAWs (digital audio workstations such as Pro Tools, WaveFrame, AudioVision), digital multi-tracks (e.g., DASH 3324, 3348) or analog multi-tracks (e.g., 24 track Dolby SR or A). Some mixing stages still use traditional magnetic film (35mm single stripe, full-coat or 16mm). One standard recording medium is 35mm full-coat (especially for printmasters), along with DA88s and the proprietary Dolby DMU MO discs (8-track magneto optical discs).

Scheduling your mix time may the most important part of preparing for your mix.  All too often, filmmakers will schedule only the time they hope the mix will take, not what the mixer and sound house agree is a livable schedule. Mixing stages are very expensive installations, often costing many hundreds of thousands of dollars, and involving expensive labor costs. If you call up a scheduling department and merely book five hours with no discussion, you may find that another client is booked in immediately after you on the same day. After all of the time it takes to set up and get rolling, you may find that you get bumped out of the room, and have to reschedule everything. Even if you finish in five hours, there may not be enough time to play back and create any additional delivery elements. When the tracks are clean, the mix goes very fast.

I counsel clients to tell the truth about time. Try to divine what you really think the schedule will end up being. You may find that the mixer can find a way to allow for your time needs without breaking the budget. No mixer looks forward to attempting a difficult mix and having to rush. This also means that the filmmaker should understand that unless he or she states otherwise, everything he/she discusses is expected to be completed in the time allotted. If you want to extend the time you are willing to pay for, then make this clear.

Some sound houses charge only by the hour and others are willing to work on a bid basis.  Even if you have negotiated a bid or "flat bid" for the project, this typically means that the bid is good only within certain parameters. Get the details explained to you carefully. No sound company can logically commit to an unlimited amount of time for a flat rate. Mixing is typically under the direct control and direction of a supervisor (e.g., the filmmaker, producer, director, editor or sound supervisor) who ultimately determines how long a mix will take. It is quite rare, for example, for ADR (automatic dialogue replacement) recording (looping) to be flat bid because no one can actually predict how long it will take an actor to actually re-record his lines. Sound editing is often bid out (especially where a certain standard is established) because typically, the client doesn't supervise this entire process.

Every sound house has different Standard Operating Procedures (SOPs). Be patient when your mixer explains his special procedures; you may be unwise to force a facility to do things your way instead of the way they are used to working. Networks and distributors have different SOPs, which are elucidated in the "delivery requirements" part of your contract with them. Read your delivery requirements carefully and share them with your mixer early on. For example, I just budgeted a documentary feature that had an extra $14,000 of unusual deliverables. Look out for the terms "fully filled effects, undipped mixes, constant level mixes, special M&E (music and effects) mixes, special tracking requirements, SR/SRD/SDDS/DTS optical sound negatives," etc. You may have contracted for a deliverable that you have not budgeted for. You may be able to re-negotiate your deliverables if you need to, since business affairs may have merely attached a standard set of deliverables for a feature film that does not really apply to your wildlife documentary.

Try to relax during the first hour of a mixing session. The mixer(s) will need to do a lot of preparation, count and mount your elements, patch them to the console, prepare and adjust the outboard gear, listen to the tracks soloed to hear what you have got. It is very typical for some problem to arise, since this is the first time anyone has tried to co-ordinate elements from diverse sources. If the composer's DA88 has bad time code, go out in the lobby for a cup of coffee and let the mixers get their act together.

Remember that often the mixer is listening for the first time to a show that you have been immersed in for months. Ask the editor to prepare a one- or two-page rundown of the show, listing with some detail the sequences in the program, using the "pet" names you use for them with associated time codes. Have a copy of the rundown for all mixers and assistants. Cue sheets are also essential. It is at this point that you may wish you had provided the mixers a VHS of the temp mixed show so they could have seen it last weekend in preparation.

Mixers like to mix on the big speakers because it is easier to hear the subtle sounds, and frankly, the mix is more fun. The small speakers are often used just during occasional playbacks. For TV programs, the mix may also be played through an actual television. There is an acoustical phenomenon (see Fletcher-Munson Curves) that dictates that louder sound levels have more apparent bass. This means that the louder the mix is played, the more apparent the music will be, since music typically has more bass components than dialogue. On small speakers, the music will generally sound softer. For this and many other reasons, mixers often insist on mixing at their standard volume levels. This can cause a problem if you have super sensitive ears, since standard levels can seem loud to some people. Remember that a sound studio is an ideal listening situation; your mix will never sound this good at home. It is helpful to keep in mind that subtle effects tend to go away. It will also help to have a quality listening environment at your home and office. 

At some point, the mixer will explain that it is time to start listening, that the mix is beginning. After this point, try to limit phone calls. Whereas in the beginning of this session your role as filmmaker was to give the mixer space, now is the time to devote your full attention. Make it clear who is in charge of running the mix. I urge the filmmaker to speak out immediately when there is a note. I may not stop mixing at this exact moment, but at least I know that someone has a comment, and I may have to back up. 

Mixes are done using "stems"—individual tracks of dialogue, music and effects. Often there are also stems for voiceover or foley. This means that the mixers can mix the dialogue separately, then go back and mix music and/or effects. It is a "divide and conquer" technique that allows a complicated mix to be broken down into manageable parts. The mixer will usually tell you if he is working on one stem separately than others. Thus, if he is trying to accomplish the perfect music cross-fade, wait to give him comments about the dialogue. If a mix is sufficiently complicated or if there are a large number of tracks involved, pre-mixes may be done in advance. Pre-mixes are often done for effects and dialogue since they may require extensive fiddling, but music is typically saved for the final mix pass.

If the mix is a simple stereo mix (often called TV Stereo), each stem is two tracks. If the mix is 5.1 or AC3, then each stem may be as many as six tracks. This system of mixing in DME stems saves a lot of time and allows changes on one stem without affecting the other stems. Thus, one can change the music mix without having to remix all of the dialogue and effects. Stems also allow making of international tracks, called M&E (music and effects) tracks, so that foreign dialogue may be added at a later date.

Mixing in 5.1 is becoming more and more popular for documentaries, since this format is playable on DVDs. The term 5.1 refers to the five-speaker array (left, center, right, left rear, right rear); the "point one" is the subwoofer.  This type of mixing is also called AC3 and Dolby Digital. The theatrical release formats SDDS and DTS are also usually 5.1. In 5.1 mixes, the five-speaker channels are all high fidelity and discrete. In the older Dolby Matrix system (also called Dolby Pro Logic), there is a left, right, center and surround channel. These four channels are "matrixed" in two channels, called Left Total and Right Total (LT/RT). The surround channel has a narrow frequency range and a sound cannot simultaneously play in the front and back of the room (i.e., channels are not discrete). Even while mixing in a 5.1 format, mixers often check mono, simple stereo and Dolby matrix monitors to make sure the mix will translate to other formats well. 

Mixing in 5.1 format can be more expensive because more tracks are needed and editing time and equipment expenses are higher. Some mix rooms and consoles are not equipped for 5.1 mixes. Consult your sound designer for differences in these formats. It is possible to simulate multi-channel mixes in different ways—for example, reverb and delay can expand a stereo music track into the surround channels. The 5.1 format is ideal for documentaries, since the dialogue (and narration) track is usually played out of the center speaker alone; this leaves more room acoustically for music and effects in the other speakers. In this way, the music and effects can be played louder and have more impact.

What you cannot do easily in the mix. In the old days, once a show was in the mix stage, it was difficult and expensive to change edited tracks. These days, many mix stages have workstations in "virtual" mode, where sound effects can be added, voiceover moved and music cues slipped with a few keystrokes. What is difficult is doing anything that involves a change of even one frame in the picture. Unless the sound house has a music library, it may be difficult to change a music cue, but it may be possible to recycle a cue from somewhere else in the movie.

Push hard for what you want in the mix. It may be hard to accomplish, but any mixer will try to accommodate your goals if there is a way. I find it is easier if you state a goal instead of trying to tell the mixer exactly what to do. Instead of saying "EQ the high end in that speech sharper" you may want to say "I want to understand the dialogue better here". There may be many ways to make the dialogue more intelligible that you may not know about. Also, think about what the loudest parts of your mix should be so the mixer can save up some headroom. If you want the climax with its explosions to be much louder than everything else, it helps if you have clued the mixer to this in advance. Not only will he EQ the explosions to have more "apparent volume," but he will leave some room on the VU meter and he may mix the preceding scene a little softer. Don't harbor an unspoken grudge about some part of the mix you don't like. It will likely be cheaper and easier to fix it now.

Leave room in your mix schedule to playback. It is wise to playback each major section of the mix before moving on to the next section. Even if the mix console has automation, doing fixes is much easier if the settings have not been changed. I love the luxury of saving a playback of the finished mix to the next day; then our ears are fresh and we can set about doing fixes with a clear head.

After the mix is complete, it is time to make all of the versions and deliverables. These include mono mixes and stereo mixes, if done separately; music and effects mixes; undipped mixes with filled effects, etc. In television mixes, the mix is typically "laid back" to the picture master (5.1 mixes may be encoded on a picture master discrete or using Dolby E.). In theatrical mixes, a "printmaster" is made after the final mix, where all of the stems are combined into the appropriate format (Dolby Matrix, Dolby Digital (SRD), SDDS, DTS, etc). The printmaster is often recorded on a magneto optical drive (MO) and/or 35mm full-coat magnetic film and/or DA88 and/or Pro Tools Session. An optical house can make an optical negative from the printmaster to be used during printing of the film answer prints. Take care that all elements of your soundtrack are quality controlled at every step of the process. One bad dub and all of your hard work is sabotaged.

Finishing a final mix can be the peak moviemaking experience of all. Don't forget to invite your mix crew out to lunch or shake everyone's hand at quitting time—including scheduling and back room technicians.


Kent Gibson is the owner of Soundesign in Studio City. He also produces science documentaries.