Learning Space dedicated to
the Art and Analyses of Film Sound Design
What's new?
Site Map
Site Search
Sound Article List
New Books


Editing Dialogue

I'm working on a feature film and the dialogue is not that bad but not fantastic (probably not as bad as I think). There is a scene where 2 people are being filmed and the camera is filming 1 person talking and then the other. The level of volume between the 2 people is different. I know that this is just a matter of adjusting the volume level of one of the actors voices to match the other but I was wondering if anyone could give me some tips on how the best way to edit this would be?

The audio is just on one track so I didn't know if it would be best to separate it in to 2 tracks so that I can adjust the volume easier. I'm using Sonic Foundry Vegas 4 software. This is the first project that I have had to work on dialogue so I really want to make a good go of it and learn as many helpfull tips as I can for future projects.

Jory K. Prum:
I would checkerboard the characters to separate tracks, where feasible. Then create another track where you put your room tone (which should have been recorded at the time of filming). Use the room tone to smooth the differences between edits and level changes.

You can also use noise reduction software (or hardware) to try to quiet down the background noise, if necessary.

Matt Piersall
I would individually gain all of the lower dialogue to match the higher. It definitely doesn't have to be perfect but you could perhaps run through the scene and do a quick pass. I would also split the tracks up (1 character per track) and bus those to a vocal compressor. It's really just a matter of playing with it and unfortunately in my experiences there is no quick answer each and every situation is different and requires a different approach. I guess thats what I love about what we do.

Jory and Matt already gave away the good stuff, so I'll simply add another small tip.

Vegas gives you the interesting opportunity to quickly level clips at high speed with its per-clip volume handle. Use that if you can, but be careful not to go wild on that if you're mixing via faders, since it's hard to remember all you did. Thankfully Vegas has a visual feedback in the clip waveform overview on how much clip-gain you applied. The volume automation is a seperate entity so mixing by fader is not inhibited. Should you have two different tones of the same character (lav vs boom for example) just pop 'em on another track. I sometimes also like to give each character their own compressor.

Do the busses in Vegas now have sends too ? That's one aspect of Protools I would sorely miss(or any $500+ console for that matter). If it's missing that you may have to blow additional CPU resources to insert a compressor in to each voice channel, instead of using it in a bus where you can't send a thing to any fx.

Whats the best way to split the track in to 2? Should I just copy the same track and then edit the dialogue of one actor out and do the same on the other but edit the dialogue of the other actor out so that I have 2 tracks, one of each of the actors? Would it be best to split them in Vegas of should I use somthing like sound forge?

David G.Evans
I feel compelled to note that preparing a good dialogue track is much more than simply splitting off the tracks . The objective of the split is not simply to have each character to appear on a separate fader. Each scene may have its own set of criteria for making the split depending on the technical qualities of a particular microphone set up. Each individual character may have vastly differing sonic signatures and this should predict the reason for the split. For example one character may have very striking differences in sonic quality depending on location or angle. This may require separation for an eq pass that would not be needed on other angles. It would make more sense to isolate that material in a scene than simply placing all the characters sound on track.

The split is made more difficult by the degree to which you are out-taking a scene or character. In many situations the level of ADR will also affect the way you split out the dialogue and fill in the tone around the remaining character or characters. Also one should not forget the sync effects , If they are good should also be prepared and saved for the M+E.. Saving and building up a well recorded dialogue track helps to give the sound track a natural life that should not be ignored. A good recordist may provide loads of good sound that does not necarrily appear in the EDL or OMF. There is often a wee bit of detective work to be done to discover some unused good material . Out-taking a scene and finding alternate takes can save many hours of studio time trying to match ADR to the original. The preservation and enhancement of the original recordings should be a first line priority. You should ask why am I going to split the sound out, above and beyond who is saying the particular line . What is the purpose of the split?

Hi David, thanks. The film I am working on is a student film and is based around 2 charactors all the way through, there is nothing fancy in the movie like sound fx just a bit of foley and dialogue editing, very basic really. There is a couple of scenes involving more than 2 people talking but most of the film is just of the 2 main people. There is one scene in particular that I am trying to sort the dialogue out on where the 2 characters are both in a bar and you can here the air conditioning or somthing but it is much louder when one of the actors is being filmed than it is when the other is being filmed. The air conditioning sound isn't bad, it gives it quite a nice room tone but I'm not sure on the best way to match the volumes of the 2 actors. This is why I was wondering if it might be best to split the dialogue on to 2 tracks (if so, how? e.g. edit out all the dialogue of one character on one track and do the same with the other character on the other track). I could also just keep all the dialogue on one track and adjust it at parts to match by putting in a volume line in Vegas. Please let me know your thoughts on this.

John McDaniel:
I'm interested in the protocol surrounding the use of alternate takes during dialog editing.

Do you make notes about the use of alternates? If so, are the notes for the sound department or do they get passed on? Do director's question the use of an alt? Do they care to know? Is it a case of what they don't know won't hurt them?

Just curious. I've had different reactions to the use of alts; some didn't care --others did

Randy Thom:
The answer to your question, John, is that there are different protocols on every film, depending entirely on the personalities involved. The Picture Editor has often explored using alts to some degree by the time the Dialog Editor comes on board, so the Dialog Editor needs to be careful not to ignore the choices made by the Picture Editor, and therefore insulting him/her. Whenever possible, it's often a good idea for the Dialog Editor to present alts to the Picture Editor before presenting them to the Director.

Many Picture Editors will appreciate this; a few won't. Basically the strategy is usually to do one's best to make the takes used in the Picture Editor's cut work, but to also prepare alts that have the potential to work better, if there are any. It's common to prepare alts for the dialog premix which can be put onto their own tracks.

Too many alts can make the premix too complicated though. That's an issue for the Dialog Editor and the Mixer to work out. Alts can also be cut and offered during the final mix. A good Dialog Editor will have a sense of how receptive the Director and Picture Editor will be to alts. Some will love you for cutting them, and some will think you're wasting their time.... you never know until you've worked with that particular team..

Bob Kessler:
Randy makes some excellent points, as always.

Those of us doing lo/no budget indie films are often presented with truly dreadful location sound, and the job of editing the location dialog tracks is most often an exercise of cleaning up the audio as best we can. The problems I most frequently encounter are:

1. No room tones.

2. No alternate takes.

3. Very low dialog levels in relation to the ambient sound.

4. Distorted words and phrases.

5. Words cut off at an edit point.

6. Directors instructions and crew noises overlapping the dialog.

7. No EDLs.

My first "solution" to these problems is to ask for 20 second handles on the OMF files. The extra 40 seconds on each soundbite, hopefully, supplies me with some short pieces of room tone, and, with luck, alternate pieces of dialog.

My first step in the location sound editing/assembly process is to copy each soundbite in a scene onto a separate track and stretch it out to its full length. I can then find, with luck, those pieces of room tone and alternate dialog parts I need and create a "library" I can call upon.

My second step is to shorten the soundbites back to appropriate lengths. I generally leave A LOT of overlap.

Step three is to re-organize the soundbites by "noise", finding the soundbites that can be treated similarly in the noise reduction process. I create noise reduction pre-sets I can use for later scenes in the same location and a log of how different elements were treated. (I use BNR and SoundSoap Pro and often have to do multiple passes.)

The next step, if I have the material, is to replace unintelligable or distorted words and phrases and cover mouth clicks and other extraneous noises with those little pieces of room tone. I have replaced entire lines word by word and even syllable by syllable in attempts to match the original performance.

Once I have completed the noise reduction, dialog replacement and cleanup process I re-assemble the location dialog track and try to create a smooth track with crossfades, etc.

It is during this process that I create my sound list for the scene; ambience, Foley and sound FX. What I hate the most is that more than half the time I am creating an ambience that will mask the noisy location sound rather than enhance the scene.

I normally don't separate the characters on to different tracks unless I absolutely have to.

I generally don't do anything with EQ until I start the pre-mix process.

Thank you Bob, that was just what I needed, a step by step guide to how someone else does it. I will probably pick up some of my own techniques along the way (hopefully) but I've never done this before so I didn't really know where to start. I will probably think less about separating the 2 characters on to 2 tracks now and try and match the dialogue on the same track. Thanks alot.

Excerpt from thread "Children of Men" Feb 2, 2007 at Sound Design discussion list

Film Sound Question & Answers 

To Film Sound Design


Star Wars Sounds Film Sound Clichés Film Sound History Movie Sound Articles Bibliography
Questions & Answers Game Audio Animation Sound Glossaries Randy Thom Articles
Walter Murch Articles Foley Artistry Sci-Fi Film Sound Film Music Home Theatre Sound
Theoretical Texts Sound Effects Libraries Miscellaneous