While you’re editing, there’s a few things to look out for:
Shifts in Ambience
The ambience is the background sound of production dialog. What you need to listen for is:
- Shifts in ambience when switching between mics
- Background noises that come and go unnaturally (like an airplane noise or lawnmower that come and go because of switching takes or mics)
- Ambiences that may been edited in by picture editors but don’t match what’s in there (they’re trying to do you a favor by filling in a hole but use the wrong material to do it)
A savvy picture editor will be listening for some of these problems and correct them before it even comes to audio. But, sometimes they don’t catch it or they can’t fix it. Or, if your speakers are better than their’s, it’s possible you’re hearing things they couldn’t here.
Sometimes these problems can be fixed with editing or possibly with Izotope RX. One thing to consider is that you are only listening to dialog and there may be music or added sound effects which can mask some problems. Some projects have budgets that allow fixing every detail and issue and others are fine with it being masked.
The dialog editor may be responsible for adjusting clip gain. This is a question to ask the re-recording mixer you’re going to be working with. If you are editing and mixing, it might be in your benefit to adjust clip gain as you work. But, if clip gain isn’t set correctly, it can cause more work for the mixer, not less.
Clip gain should be global across an entire scene. For example, a whole scene being brought up 6dB. This allows the mixer to make global EQ and compression changes to the scene.
Sometimes there will be clip gain in the AAF regions and in those cases, it’s better to convert the clip gain to volume automation. What you’re trying to avoid is a mixer having a fader set to zero and the dialog jumping all over the place due to clip gain. The last thing a mixer needs is to be adjusting clip gain problems.
Filling in under ADR
When audio switches from production dialog to an ADR line, you need to re-create the background ambience of the production dialog to play while the ADR line is going. If you’re lucky (like this example below), the handles of the original audio have quiet fill that will match the other dialog in the scene. Izotope RX has a module called Ambience Match module which can be effective at creating fill, also.
If you come across dialog that is going to be treated so it sounds like it’s coming from a low quality speaker, this is called a “futz”. Some examples of futzed audio are dialog coming from a radio, television, walkie talkie, or telephone.
The mixer is the one who does the futz effect and will probably use a plugin like Futzbox McDSP, Audio Ease Speakerphone, or a guitar amp emulation plugin. But, it’s up to the dialog editor to separate what needs to be futzed onto it’s own track (or tracks).
We talked about this some in an earlier section – pfx are production effects, or sound effects that were naturally picked up in production audio. It could be sounds like footsteps, doors closing, breaths, squeaks, etc. It could also be environmental sounds like a car driving by, ambience, or someone riding a bike. The idea is to add production sounds to the foreign version so it doesn’t need to be covered by sound fx later (although some sounds will be replaced or sweetened/added onto with sound design).
In the photo above, you see 4 dialog tracks, 4 pfx tracks, and two mono sound fx tracks. The dialog tracks have spoken language. The pfx tracks have useful sounds without any speaking (in this example it’s skateboarding sounds). The two fx tracks have added sounds that are sweetening the pfx. These fx sounds might be making something more prominent than it is in the pfx (like a skateboard land, a crash, etc).
It’s possible both the sound fx editor and the dialog editor will have pfx tracks. If they’ve been separated out (by the picture editor), it’s possible they will be with the materials sent to the sound fx editor. In that case, it’s the job of the dialog edit to pull out any missed pfx sounds.
It’s common for the dialog editor to do some processing for:
- Pops and clicks
- Thumps or mic bumps
- Other extraneous noises
A dialog editor may or may not be responsible for:
- Wind Noise
- Rustle noise
A dialog editor is generally not responsible for:
- Broadband noise reduction (for ongoing noise in the background)
If you’re adding any processing, it’s really important to leave a copy of the unprocessed audio somewhere the mixer can easily get to. Once you add any processing to a file, you lose your edits, handles, etc (see screenshot below – upper track is rendered, or processed, and the lower track is the original edit). For a mixer, having these edits, fade ins/outs, and handles can be crucial for adjusting the mix but once you’ve processed it, they’re stuck. This is why it’s really important that if you need to do any processing, you keep the pre-processed version for the mixer. There’s a couple ways to handle this. One way is to put the edited/un-processed regions on the track below (and mute it):
Another option is to create a separate track for unprocessed audio (like an additional X track):
If only a small portion of a region is processed (and has handles) and the rest of the region is not processed no copy is needed. In general, you want to make it quick and easy for anyone to get back to the original/unprocessed file.
There’s some Audiosuite plugins that do a pretty good job with clicks, crackle, and hum removal (like the Waves Restoration plugins) but the industry standard is Izotope RX.
Removing sounds with Izotope RX
Izotope RX is commonly used software that dialog editors use to remove problem sounds (RX Elements is a good alternative for just getting started or if budget is an issue). It’s sort of like Photoshop for audio.
Here’s the tasks in RX that a dialog editor will do:
- Spectral Repair (Visual editing)
- Removing pops, clicks, and thumps
- Short dropouts
- Sudden Unwanted Sounds
Here’s the tasks in RX a dialog editor is not responsible for:
- EQ Matching
- Dialog Isolate
- Loudness or level adjustments
Depending on the project, budget, and time a mixer may prefer to do some of these tasks themself. So, it’s possible you’ll get a dialog edit where you’re expected to do a lot of Izotope work and others where you’re asked to do none.
Knowing how to use Spectral Repair for specialized noises can make or break you as a dialog editor. If you don’t do it well, mixers will be unhappy (and they may have to redo your work on the stage). Do it well and they’ll be requesting you for work. And remember, don’t do broadband noise reduction unless your mixer has specifically asked for it (or you are the mixer). If you’re too heavy handed or make an error, the mixer will have to redo your work. This is why it’s so crucial to always keep an unprocessed version of your edit available in case the mixer needs to quickly go back to it.
RX is an area where there are a lot of great online resources and how-to videos (especially from Izotope), so we’ll leave it to them to get deeper into using it.
Noise Reduction on Dialog Tracks
Phase & Sync Issues
If the picture editor did their diligence with the audio tracks before editing, the mics you’re working with should be in phase with each other. But, if you play multiple dialog tracks together (mics that you’re including for the mixer to play together) and there’s a phase issue, it’s the dialog editor’s job to correct that.
For an entry level dialog editor, know that it’s your responsibility to fix severe phase issues – mics that are clearly phasing, comb filtering or have a slap delay (more than a couple frames off). Where you have to be careful is that by moving mics, you are also adjusting sync to picture.
If you are dealing with a mic that sounds out of phase, put your nudge into 1/4 frames and try moving a quarter frame at a time to see what sounds better. Keep track of how much you are adjusting because you may need to do a global adjustment for scene or mic. If you find you are moving more than a whole frame, keep note of what tracks you adjusted and how much.
If you had to adjust more than a couple frames, check if the original placement is out of sync. Anytime you see dialog that’s out of sync more than a couple frames, it’s important to coordinate with the picture editorial team (whether it’s the editor or assistant editor) to let them know you found audio that was out of sync and corrected. The reason it’s important to keep them in the loop is so they don’t slip picture (move it) to compensate for the audio being out of sync. Once you identify any problem related to picture, you also have to clarify who is going to correct it. If you find moving audio a couple frames causes phasing or an audible delay between tracks, then it might be better to request a picture change versus adjusting audio.
A dialog editor could easily spend days adjusting every line of dialog to be perfectly in sync especially if you are looking at waveforms to compare. But, in the scheme of the work, the pay you’re receiving, and the quality of the mics, you also have to pick your battles. If there’s a severe sync issue and phasing and an entire project may need adjustment, this is beyond the scope of typical dialog editing. This is when it’s important to notify the client of the issue because these should be expenses to correct a technical issue that are on top of any fees for dialog editing. As a rule, if correcting the issue could cause you to not finish the work on the timeframe you had set, this is a reason to flag as a technical issue. There always should be a little time planned in for unforeseen technical problems, and if there’s only a few lines out of sync or a scene (and it won’t take long to correct), include it in your normal work.
Editing ADR & Group
When you’re editing dialog, what might be considered “good” or “good enough” on one project may not be the same as another. A major film could hire a dialog editor for months to remove every pop and mouth noise. A low-budget cable television show might give an editor a day to do everything (dialog, music and sound fx) and the source audio is full of problems. In this industry, meeting the deadline is often more important than doing it right.
An analogy would be a studio recording vs a live concert. A live concert is never going to sound as good as a studio recording. The artist will sing some out of tune, wrong notes might be had, but there is an energy and intensity that you can’t capture in the studio. The two are different and because of that, the sound quality can’t be treated equally. It’s apples and oranges.
Part of learning how to dialog edit isn’t just the mechanism of how to actually edit – it’s also learning how to manage your time. If you are given 8 hours to edit 42 minutes of dialog, you have to figure out how much you need to do per hour, which problems to solve and which ones to let go. While 8 hours might seem like plenty of time, you also have to take into account that one or two problem spots could take you an hour and a tech issue in the material could cost you another half hour. If you have any questions for the mixer or just having an off day, next thing you know you’re working after-hours for free.
This video by John Purcell is a must-watch about time management:
- Part 1: Why Learn Dialog Editing?
- Part 2: The Elements of Dialog and Voiceover
- Part 3: The Challenges of Dialog Editing & Mixing
- Part 4: Stems and Specs for Dialog Editors
- Part 5: Prepping for Dialog Editing
- Part 6: Dialog Editing Basics
- Part 7: Dialog Organization for Different Projects
- Part 9: Frankenbites
- Part 10: Delivery