You’ve shot your film, and now you’re cutting the pictures. This is an important project - one you hope to see make festival runs or to get distribution - and you’ve decided to bring in a sound editor because you’ve realized that post sound is very important. Thing is, you’re a self-taught editor, and you’re used to doing everything yourself within your NLE. You’ve probably formed some bad habits. What happens now?
Let’s back up. Before you get into such a project (if you aren’t already), there are a few things you need to make sure you do. Forming the right habits early can actually make your work more efficient even if it’s just you. More importantly, it will make your sound editor much happier when the project is sent on. Adopt a downstream mindset: who is going to handle this project downstream from me, and how does my work impact theirs?
Truth be told, I often waste quite a bit of (billable) time simply trying to clean up a session before I can truly begin the work of sound editing and sound design. This can be avoided. Here are the steps to take to ensure that your project flows seamlessly downstream to your sound editor.
Timecode Pt. 1: Hitting the Mark
One thing I often see from self-taught editors is a tendency to start the editing timeline at 00:00:00:00. This is often the default starting time of new projects, and defaults are mistakenly accepted as good enough. This starting point, though, can cause issues later on.
When starting a project at 00:00:00:00, there’s nowhere to go ahead of the first frame of content. This means that any leaders have to be dropped in at the head of the project, and the first frame of content is pushed to an odd value like 00:00:12:00, or worse, to a start time that isn’t at an even second. The odd start time can raise questions later when trying to drop the mixed sound file back into the project, and it just isn’t a clean starting point.
Even worse, and I’ve seen this happen, is a project that is edited starting at 00:00:00:00, then later (after picture lock) having to have everything in the timeline slide down to make room for slate and any other leaders that have to go in before first frame. Selecting all and sliding down the timeline risks knocking things out of sync, and sometimes that affects only a couple of bits here and there rather than the entire edit. That’s a lot of fun to try and fix later.
The industry standard is to start the first frame of content at 01:00:00:00. This makes a solid reference point that will always be the first frame of content. This also makes it easier to keep track of things across the timeline, and to make notes for corrections and changes. The project timeline can start at 00:59:30:00 to allow a full thirty seconds of space for leaders, keeping the first frame of content at the hour mark.
And, for good measure, include a 10-second countdown with 2-pop (or 2-pip). The countdown starts at exactly 00:59:50:00, and the 2-pop is a single frame of video (the “2”) and a single frame of 1kHz tone at exactly 00:59:58:00. The 2-pop is another tool to help check sync downstream.
A Tidy Timeline is a Happy Timeline
If there’s one thing that can dominate my first few hours of a session, it’s trying to get sound organized. I’ll often receive an edit where sound is scattered about.
On the video portion of the timeline, it’s easy: the clips on the highest track are the ones seen up front in the image. The clips on the lower tracks are seen behind the higher clips. The audio portion of the timeline, however, doesn’t work like that, and it can be challenging to think differently about it.
Audio clips/events aren’t heard in the mix based on which track they’re inhabiting, but by their own gain settings to make them louder or softer in the mix. When audio is organized in the same layers as the video, things get messy. There’s some music up here among dialog. Some of the dialog is down there with some of the sound effects. The VO is scattered across 3 or 4 tracks in random places. Some of the sound effects are living among the music edits. It can be a scavenger hunt, trying to locate common elements and reposition them on the right tracks.
Keeping audio tracks organized in the picture edit takes a little attention. Some NLEs allow setting target tracks before dropping media into the timeline. Some, like FCP X, allow for setting audio roles in the clips before dropping anything into the timeline. However your NLE handles it, learn to ensure audio clips fall into the right places.
I typically try to keep my tracks in this order:
VO
Dialog (DX)
Mono Sound Effects (SFX)
Stereo SFX
Music (MX) - Keep library music on tracks separate from score, if applicable.
There will likely be multiple tracks for each category, but the idea is to keep them all grouped and in order. If you aren’t already doing this, you’ll find that it makes your own work much easier as you’ll always know where to find what you’re looking for. Your sound editor will also be able to find everything, and can start the actual work much faster this way.
Timecode Pt. 2: The Window to the (Sync) WorldWhen picture is locked and ready to go to audio post, the sound editor will need a copy of the picture. Most importantly, the video needs to have BITC (Burned-In Time Code). This is a text overlay that displays the running timecode of the edited sequence. The video with BITC may be referred to as a “window dub”.
Create a copy of your project, just for safety. Your NLE likely has some sort of timecode overlay effect built in. Make sure to set it to read the project timecode, not the source timecode. The timecode window should be placed so as not to obscure, well, most everything. For letterboxed footage, centered in the lower black bar is great. For full-frame (16:9) images, the top right corner keeps it from obscuring footsteps, which Foley will need to reference. This window is crucial for checking that everything’s in the right place, and eliminating any question about the project’s starting timecode value.
Speedy Delivery
The last, and most crucial part of this is delivering assets to your sound editor. Before I go into any detail, the biggest piece of advice I have here is: talk to your sound editor about which assets are needed, and deliver those. Not every post house operates exactly the same, and not every sound editor has the same preferences. What follows is a basic starting place, but please confirm details with your sound editor.
Typically, the sound editor will request an AAF (Advanced Authoring Format). This file imports into the sound editor’s DAW and opens a session with all the audio tracks and audio clips/events as they were in the NLE timeline. The older, and mostly outdated, option that you still may encounter is the OMF (Open Media Framework), but it’s rare and has some limitations, such as a 2GB file size limit. Back to the AAF, there are options that are important to know.The two types of AAF are referenced and embedded. A referenced AAF generates a folder that contains very small .aaf file and a full media library, as the data in the AAF simply points to the media files. While the AAF itself is small, the need to send all media requires lots of storage space for lots of data. The better option is typically the embedded AAF, which is a single, and much larger file that includes only the audio used in the edit.
Next is to trim or consolidate media (only applies to embedded AAF), which means only the sections of each audio clip/event used in the timeline are embedded, rather than the entire source clips. Along with this is the option to set handles (how much of the media outside of what’s actually used in the edit is included on either end). I typically ask for 240 frames (10 seconds), but this also differs from editor to editor. Trimming media in an embedded AAF provides the sound editor what is needed, with a little extra room to adjust edits and fades, but allows for a smaller data transfer.
As a note to Premiere Pro users: there’s a setting within Premiere that, when a video clip is rendered, its attached audio is also rendered. Disable this option, as it renders only what’s used in the timeline, and eliminates any chance of setting handles. This means the sound editor cannot adjust edits. In any NLE rendered audio cannot be undone later.
Unless requested by the sound editor, do not include video in the AAF. Export the window dub as its own video file, and deliver that along with the AAF file. Also, please check with your sound editor to ensure you’re sending the needed video resolution and format. And make sure that the AAF audio is embedded at 24-bit/48kHz BWF (Broadcast WAV).
Please note: AAF export is a feature available in AVID, Premiere Pro, DaVinci Resolve, and VEGAS Pro. For FCP X users, the process requires a third-party app called X2Pro, which converts a Final Cut XML file to an AAF.
And Then?
Once the re-recording mix is complete, the sound editor will send back any needed mix files, which you will then drop into your video project to export the finished product. That’s pretty much it… or is it?
One more important conversation to have with your sound editor is the intended delivery platform. Whether for broadcast, streaming, social media, or theatrical screening, each destination has its own loudness standard, and the sound editor needs to know which standard(s) to target in the mix, or mixes if there are multiple platforms involved. And the type of mix - stereo, 5.1 surround, 7.1 surround, Dolby Atmos - should be discussed up front. No matter the destination, your sound editor will mix to those specs.
The moral of the story here is that any project that has more than just one person taking care of the different aspects of postproduction needs always to consider the next person downstream and how the project moves on to that person’s workstation. Just as the decisions made in production flow downstream to post, decisions made in the picture edit will flow downstream to sound, VFX, coloring, etc. Filmmaking is, after all, a team sport.
