Dialogue Editing for Video Productions - Part 2: Workflow

One of the most frustrating things about preparing a good dialogue track is conveying to the producers how much time is going to be needed to do a proper job. Addressing the challenges outlined in my last post is not as simple as running the dialogue through a noise reduction filter and automating the volume and EQ – although the time and money available for post production sometimes dictate that you do little more than this and hope for the best.

Editing a reel of film (approximately 18-20 minutes of video) requires a lot of separation of scenes and shots within scenes. Within a scene there might be a wide shot with all of the actors conversing, a close up of each actor, and sometimes an angle shot or the “art” shot. Each of these shots will most likely have different sonic characteristics that require blending and smoothing. When you think about how many different scenes are in a 90 minute film you can begin to see how much work has to be done in order to do a proper job and end up with a crisp clean dialogue track.

As written in my previous post a layered approach seems to work best when editing dialogue and may require several layers of work to get a great dialogue track; again time and money will dictate the level of cleanup and editing you can realistically perform.

After taking the time to split scenes and shots to different tracks, my first pass is all about making the transitions from the different shots within a scene sound smooth and believable. Smoothing scenes is a constant trade-off between noise and evenness; this is where room tone becomes invaluable. In an ideal world each scene will have 30 seconds of clean room tone; in most cases you have to carve room tone from the scene itself trying not to get any production FX (clothe rustle, prop sounds, etc) mixed into the room tone you are trying to piece together. A tip I learned was to take small snips of room tone from the shot, reverse it and cross-fade it with itself to create a longer piece of room tone – this works for small transitions – longer transitions might sound like a loop and distract listeners. It cannot be said enough: Room tone is the single most important tool of dialogue editing! Without it you cannot create a convincing and smooth dialogue track. It is used in every stage of editing to transition between shots, to remove noise from clicks, pops and other unwanted human related sounds, and to “place” ADR into the shot. It amazes me at how “clean” a track will start to sound once you have smoothed the transitions between shots.

During this pass I make note of trouble spots that need particular attention – such as poor signal to noise ratio and hunt through alternate takes (if there are any) for replacement and then discuss with the producers about replacing the dialogue with ADR (if the ADR is not already available to drop in). If that is not possible you really have to work hard at creating convincing transitions between room tones and use multi-band expansion along with noise reduction to bring the voices up and the noise down.

Once I have worked through all the scenes and smoothed and added room tone the next thing I tackle are noises – the clicks, the clacks, the whispers from crew personnel, the sounds of extra footsteps from camera men, and the sounds of tools or zippers hitting a boom pole or being dropped. This takes a lot of time, listening, stopping and scrubbing. There are really only two ways to deal with these types of unwanted sound – fill with room tone or replacement. I will also reach for the main tools that are used for most cleaning – EQ, Multi-band Compressor/Expander and the Noise Reduction filters.

The EQ is used at this point to notch out noises such as rumble and buzzes and hum. The MB Compressor is used a broadband processor. Set up as an expander you can remove a lot of noises such as wind, rumble and high frequency hiss. The Noise Reduction I use sparingly – the artifacts, if used too aggressively can ruin a scene. The key at this point is to process less. When the dialogue is mixed in with the Foley, FX and music the remaining noises are usually not noticeable and additional light processing can still be used without over-processing the track

Now that the dialogue track is smooth, balanced and free from strange noises the process of bringing life to the track can start – this is where it gets less boring and tedious. This is the pass where you can bring focus to a specific character or to a group and bring perspective to the scene. It is also during this pass that I create telephone or radio type dialogue splits, placing the regions on their own tracks for further cleaning and processing.

At this point the dialogue should sound pretty good and for an average low budget video this is about as far as I can get (before starting the mix). Going back and reading this post – it seems easy, but it never is. The important thing to remember is that you don’t want to produce one scene and have crud in the next scene – keep a level of quality throughout the entire film. Another tip I learned: Don’t start with the first reel. The first and last reels are the most important in the way of sound quality. The audience will judge the film based on the first few scenes and the will remember the last few scenes. Start editing the first and last reels once you have “learned” the sound of the film.

I probably didn’t cover everything that I have learned or do (blending ADR for example), and as I continue to do more dialogue editing I will post more tips.

Again, one of the hardest things to accept is that you’re not going to fix all the things that you think you need to, especially if you're stuck with less than ideal dialogue. Do the best you can with the limits you face.