Audio mixing strategy, part 1

Modern nonlinear editors have good tools for mixing audio within the application, but often it makes more sense to send the mix to a DAW (digital audio workstation) application, like Pro Tools, Logic or Soundtrack Pro. Whether you stay within the NLE or mix elsewhere, you generally want to end up with a mixed track, as well as a set of “split track stems”. I’ll confine the discussion to stereo tracks, but understand that if you are working on a 5.1 surround project, the track complexity increases accordingly.

The concept of “stems” means that you will do a submix for components of your composite mix. Typically you would produce stems for dialogue, sound effects and music. This means a “pre-mixed” stereo AIFF or WAVE file for each of these components. When you place these three stereo pairs onto a timeline, the six tracks at a zero level setting should correctly sum to equal a finished stereo composite mix. By muting any of these pairs, you can derive other versions, such as an M&E (music+effects minus dialogue) or a D&E (dialogue+effects minus music) mix. Maintaining a “split-track, superless” master (without text/graphics and with audio stems) will give you maximum flexibility for future revisions, without starting from scratch.

A recent project that I edited for the Yarra Valley winemakers was cut in Avid Media Composer 5, but mixed in Apple Soundtrack Pro. I could have mixed this in Media Composer, but I felt that a DAW would give me better control. Since I don’t have Pro Tools, Soundtrack Pro became the logical tool to use.

I’ve had no luck directly importing Avid AAF or OMF files into Soundtrack Pro, so I would recommend two options:

a)    Export an AAF and then use Automatic Duck Pro Import FCP to bring those tracks into Final Cut Pro. Then “send to” Soundtrack Pro for the mix.

b)   Export individual tracks as AIFF audio files. Import those directly into Soundtrack Pro or into FCP and then “send to” Soundtrack Pro.

For this spot, I used option B. First, I checker-boarded my dialogue and sound effects tracks in Media Composer and extended each clip ten frames to add handles. This way I had some extra media for better audio edits and cross fades as needed in Soundtrack Pro. Next, I exported individual tracks as AIFF files. These were then imported into Final Cut Pro, where I re-assembled my audio-only timeline. In FCP, I trimmed out the excess (blank portion) of each track to create individual clips again on these checker-boarded tracks. Finally, I sent this to Soundtrack Pro to create a new STP multi-track project.

Soundtrack Pro applies effects and filters onto a track rather than individual clips. Each track is analogous to a physical track on a multi-track audio recorder and a connected audio mixer; therefore, any processing must be applied to the entire track, rather than only a portion within that track. My spot was made up entirely of on-camera dialogue from winemakers in various locations and circumstances. For example, some of these were recorded on moving vehicles and needed some clean-up to be heard distinctly. So, the next thing to do was to create individual tracks for each speaking person.

In STP, I would add more tracks and move the specific clips up or down in the track layout, so that every time the same person spoke, that clip would appear on the same track. In doing so, I would re-establish the audio edits made in Media Composer, as well as clean up excess audio from my handles. DAWs offer the benefit of various cross fade slopes, so you can tailor the sound of your audio edits by the type of cross fade slope you pick for the incoming and outgoing media.

The process of moving dialogue clips around to individual tracks is often referred to as “splitting out the dialogue”. It’s the first step that a feature film dialogue editor does when preparing the dialogue tracks for the mix. Now you can concentrate on each individual speaking part and adjust the track volume and add any processing that you feel is appropriate for that speaker. Typically I will use EQ and some noise reduction filters. I’ve become quite fond of the Focusrite Scarlett Suite and used these filters quite a bit on the Yarra Valley spot.

Soundtrack Pro’s mixer and track sheet panes are divided into tracks, busses, submixes and a master. I added three stereo submixes (for dialogue, sound effects/ambiances and music) and a master. Each individual track was assigned to one of these submixes. The output of the submixes passed through the master for the final mix output. Since I adjusted each individual track to sound good on its own, the submix tracks were used to balance the levels of these three components against each other. I also added a compressor for the general sound quality onto the submix, as well as a hard limiter on the master to regulate spikes, which I set to -10dB.

By assigning individual dialogue, effects and music tracks to these three submixes, stems are created by default. Once the mix is done to your satisfaction, export a composite mix. Then mute two of the three submixes and export one of the stems. Repeat the process for the other two. Any effects that you’ve added to the master should be disabled whenever you export the stems, so that any overall limiting or processing is not applied to the stems. Once you’ve done this, you will have four stereo AIFF files – mix plus dialogue, sound effects and music stems.

I ended the Yarra Valley spot with a nine-way tag of winemakers and the logo. Seven of these winemakers each deliver a line, but it’s intended as a cacophony of sound rather than being distinguishable. I decided to build that in a separate project, so I could simply import it as a stereo element into the master project. All of the previous dialogue lines are centered as mono within a stereo mix, but I wanted to add some separation to all the voices in the tag.

To achieve this I took the seven voices and panned them to different positions within the stereo field. One voice is full left, one is full right, one is centered. The others are partially panned left or right at increments to fill up the stereo spectrum. I exported this tag as a stereo element, placed it at the right timecode location in my main mix and completed the export steps. Once done, the AIFF tracks for mix and stems were imported into Media Composer and aligned with the picture to complete the roundtrip.

Audio is a significant part of the editing experience. It’s something every editor should devote more time to, so they may learn the tools they already own. Doing so will give you a much better final product.

©2011 Oliver Peters