Dialogue Transfer and Editing
Up to this point in our series on how audio can help your digital video production we’ve been focused on preproduction and production. The overview in part 1 laid out the case for taking audio seriously in the first place, the preproduction checklist in part 2 set us up for success, and the location sound primer in part 3 spelled out what to listen for on the set. From here we move on to the post production process, and specifically dialogue editing.
Dialogue Editing: an overview
As with location sound, entire careers have been built solely on the art and science of dialogue editing, and it is an aspect of filmmaking and digital production that can take a lifetime to truly master. If you are doing the dialogue edit yourself and have even a moderately complex project I would highly recommend the book Dialogue Editing for Motion Pictures: A Guide To The Invisible Art by John Purcell. It offers excellent insight into the practical aspect of making production dialogue work and spells out the dialogue post workflow in a straightforward and understandable manner.
Even if you are not doing the dialogue edit yourself, you’ll want to be familiar with the specific technical aspects to watch for and the potential trouble spots that can crop up in modern post production workflows. Successful digital filmmakers understand the weight that the dialogue carries in their productions, and spend time preparing and testing those workflows well in advance. Your ability to navigate the following parts of the post production workflow will often be the difference between something that is watchable and something that is spectacular.
In the days of analog the transfer process was one of the most frequently fouled up steps that the production audio had to pass through. It was common for stereo sources to get summed to mono, for well-recorded sounds to be distorted, and for excess noise to be added unintentionally and unnecessarily. Fortunately since the digital revolution everything has become file based and your precious dialogue won’t have to pass through a mis-routed 4 channel mixer on its way into the editing workstation.
The digital workflow has its own quirks however, and the transfer process can often be the source of a new batch of audio gremlins that didn’t exist before. Sample and bit-rates should be matched up between the acquisition and the editing workstation. Some workstations will import audio with multiple sample rates and not convert them, instead just playing them at the wrong speed - so it is very important to know what the sample rate setting is for your edit session as well as for your source audio. If the two don’t match you’ll want to convert them all to the same sample rate before or during the import.
Remember that once you import the audio, sync it and link it together with your pictures you will be listening to the primary source audio that your final audience will hear. Take a few minutes to be sure that all of the front end work and effort put into your sound during the shoot are being retained during this critical step. If something doesn’t sound right, raise the red flags right away and treat any new problems that crop up seriously.
Sync is one of the trickiest and most essential parts of the transfer and editing process. Any number of elements can cause your audio to slip out of sync and you have to be ever vigilant to keep it all in line. A common problem is sample-rate mismatch, which was addressed above. If your edit session is set at 48k samples per second and your source audio is 44.1k samples per second then you’ll start to notice sync issues within a 60 second clip.
Another common source of problems is the difference between film speed (true 24 or true 30 frames per second) and video speed (23.987 fps or 29.97 fps). Some cameras have documentation or settings that state 24 or 30 fps (indicating film speed), but in reality they run at 23.987 or 29.97 (video speed). Others actually can switch from film speed to video speed. Audio recorded at 48kHz or 44.1kHz assumes that it will be run at video speed (23.987 or 29.97), so if your camera and your edit session end up at film speed then you will end up with drift until the location audio is pulled up by .1%. This difference of .1% can cause sync issues within a 3 or 4 minute clip, and after 10 minutes or so the drift can cause sync to be off by nearly a second. To avoid this most digital video projects should be shot, edited and post processed at video speed. Either 23.987 or 29.97 work great, but remember that your documentation and menu settings may be a little misleading if the manufacturer resorted to shorthand when setting up what the parameters look like.
The third and trickiest of sync issues is often created by the display that you are using itself. Modern flatscreen displays such as LCDs, plasma, and DLTs have been shown to introduce non-trivial amounts of delay in the picture. The Dell LCDs that I work on induce about 1/2 frame of delay by themselves. Further, all digital video interfaces including firewire and USB add a layer of delay into the display chain - often as much as 4 full frames. The way to deal with this is to measure the delay in your specific display chain and then offset the audio playback by the same amount. The playback tab inside of system settings in Final Cut has a field called “Frame Offset” that does exactly that. One easy way to check display sync is to set up a countdown leader that has one-frame sine wave pops that play back as the numbers change. Experiment using the frame offset until the sound and pictures line up exactly.
Its important to address sync issues early and seriously, as they can cause huge headaches if you let them get out of control. Most people can perceive sync shifts as small as 1 or 2 frames, and people are far less forgiving of sync shifts that place audio before the picture than the other way around.
The process of editing will vary greatly depending on the quality of your production audio and the nature of the project that you are working on.
An important aspect of editing is to listen at a high enough resolution to get a real feel for what you’re actually doing to your audio. Headphones are often ok for editing, though they can become fatiguing pretty quickly. If you are using speakers be sure to listen at a high enough volume that you can hear any clicks or pops that your edits may be causing. Remember that the quieter your environment the less loud you’ll have to turn up your speakers to achieve this. Laptop speakers are usually not going to put out enough resolution to do any real editing. If you do rough edits using just laptop speakers then be sure to double check everything on a better sounding system (or good headphones) before turning the project over to audio post or before mixing yourself.
For documentaries and interviews that are recordings of real world people talking in real world speech patterns you will often be dealing with people stuttering and stammering, droning on and on, and otherwise saying things that just won’t make the final cut. Do your best to get rid of all of the unflattering “Um"s and “Uh"s and cover with B-roll. When editing disparate parts of a long speech down into a more concise thought, be sure to make the track flow like a normal speech pattern. Allow a little space in between thoughts and just use the natural environmental sounds or room tone to fill in the gaps. Make liberal use of crossfades to help with smoothness and to eliminate pops that come from DC offsets.
For scripted productions you wont have as many compromises to make with regards to “Um"s and “Uh"s, but you still may encounter a few. You will probably also have better production audio than a documentary style shoot, so the focus should be on smoothness of the edit. Use room tone or any other ambiances that were captured on the set to create a solid, smooth dialogue track that can stand on its own without needing to be hidden in music and sound effects. If there are sections that are just unusable then either mark them for ADR or consider altering the edit to work around them. Refer to John Purcell’s book for a deep look into the specifics.
This is the part of the process where all of the preparation and care that was taken during location scouting and production recording will pay dividends. If you have solid location recordings then the dialogue editing process will be a creative process where you can focus on content. If everything coming in sounds horrible then the process (and the entire project) can quickly devolve into a nightmare of problem solving and compromise.
The audio transfer and edit are a pivotal part of your production. There are always technical hurdles to be overcome and the standards of editing and delivery are constantly changing, so this is a step that must be taken seriously. Careful transfers will relieve innumerable headaches during the edit and mix (and the final screening). Poor transfers can cost time, money, and in the worst cases they can sink the entire project.
Your dialogue edit will be a much more creative process than the transfer, and will go hand-in-hand with your picture edit. Creativity in removing the superfluous sounds and thoughts are rewarded with a tighter, more professional end product. This is the true art of storytelling in your project, and the time devoted to it pays dividends quickly.
In part 5 we will discuss noise reduction and dialogue mixing.
This post is part of the series: How good audio can help your digital video production
An in-depth step by step guide to audio for your digital video production. Learn about audio pre-production, location sound and post production. Learn what to listen for and what questions to ask, even if you’re not the audio guy on the set.