Dialog Editing

T

trueindie

Active member
Just wanted to start a dialog editing thread with a few basic questions.

I was watching a tutorial a few days ago, where the tutor was adjusting the volume differences between different actors on a clip by clip basis. Yesterday, I found out what a compressor was. I was wondering if the same thing couldn't be done more efficiently by using a compressor on a dialog sub-mix. And then further slight adjustments could be made on a clip by clip basis as necessary. It seems to me that using a compressor might alleviate dialog volume level problems somewhat before a final adjustment. Is this a bad idea?

Does anyone have a favorite compressor for dialog editing, as opposed to voice compressors for music production?

Please take it easy on me, and feel free to be elaborate in your answer. I didn't understand what the hell a bus was before yesterday. It's possible I still don't know what a bus is.

- Also what do you guys think about the idea of boosting the frequencies around 160Hz to add power to dialog. Do any of you guys do this? Or should I stay away from this idea?

Thanks,
Aveek
 
Never use a compressor to replace mix level automation. A compressor doesn't just affect level; it also affects dynamic range. This means that it can make the background/noise floor sound louder in comparison to the dialog in varying levels, which will make for a very inconsistent edit. That will also creat hell in trying to smooth out edits with room tone.

Light compression may be used to help catch sudden peaks or to help an overall dialog edit cut through the mix, but it is NOT your go-to for level mixing. Use it gently and sparingly. It won't have the same setting from source to source.

As for EQ, there's no universal formula. Yes, there are specific frequency ranges that are used to add or remove certain characteristics of various sources, but they'll be used differently on each source. Especially if you are not monitoring on accurate monitors that have been calibrated, and in a treated room, you don't really hear what you're doing. And that's the first, most important thing you can do if you really want to be working with sound: ensure your monitoring environment isn't misleading you.
 
I always tend to edit clip by clip and adjust them all to be about the same levels.

Some people like to checkerboard for each actor but I prefer to checkerboard on a scene by scene basis unless special treatment is required such as telephone effects etc and those clips will be on a different track.

I edit each clip from the picture master OMF and overlap and add mono atmos to fill where required so that each scene has a continual sound track for the dialogue.

I would then pre mix my dialogue and it is then that I would add any compression via the dialogue sub group, these days of full autometed mixing allow virtual mixes to be done but by starting with a clean dialogue pre mix you can iron out any problems with EQ and any bumps in the sound track.

I would then add the stereo atmos and music along with mono spot effects and noises off etc and final the Foley tracks as and when required if they are available.

I would lay out my stereo sub groups and buses as follows for a stereo mix:

Group 1: Dialogue
Group 2: Effects
Group 3: Music

Main mix stereo (1+2+3)
Main mix music and effects (2+3)

Sometimes I also do a third mix bus which is dialogue and effects for promos etc where they will edit ad put their own music bed on but that would be sub groups 1+2.
 
It kind of depends on what you are editing dialog for. Features (professionally) is significantly different than say ENG or even no budget filmmaking. Are you doing it all or just dialog?

As a general rule a compressor is not a good choice. Neither is "normalizing", but it also depends a bit on what you are doing. Have 1/2 hr to polish up a 1/2 hr segment for broadcast and maybe a compressor or normalizing has a place. But then that isn't really doing a "dialog edit" either.

On a bigger budget film you are not going to be doing things like compressing, the mixers will generally. They, especially old school mixers, will often run the dialog through the "container" which is the pre processing for the matrixed stereo. It does a bit of EQ and a bit of compression, but that will happen after the levels have been adjusted. You can speed up the process by getting a control surface and actually "mixing" with level automation. And then tweak it after and deal with special cases.

A lot of mixers have a dialog "chain" (series of boxes or plugins) that they like for dialog and will use some variation pretty much always. As a rule mixers don't like editors to use plugins and such, but that is not universal.

If it's a One Man Band post and you are both then you kind of need to separate the editing from the mixing, because they have different goals. As an editor you are cleaning up and getting tracks ready to mix and when mixing you are balancing them against the other tracks. It's usually a help to do at least rough level adjustments when editing because it makes the mixing go faster, but it's a waste to spend a lot of time on that till you have the rest in place. You are going to be working on every clip/region anyway so doing a level adjustment is not a big deal as you edit.
 
Thanks for the warning Alex. I fully appreciate the control that audio post guys and gals want over every millisecond. But I'm open to do a few tests.

Gary, do you have a favorite compressor, or a couple of favorite compressors. It would be great to know.

Scott
Yes sir, I'm a one man band right now for post production, whether it comes to color or audio, or anything in general. My shoot was a low budget one, and there was no mixer on set. We had 1 boom and two lavs for almost every shot.

So just wanted to understand this a bit further:
- The dialog editing and the mixing processes are separate jobs with separate goals.
- In dialog editing, the job is to make sure that any clicks or unwanted noise is gone (not NR, but de-breaths, or unwanted thumps etc.), and cross fades and room tones are where they're supposed to be.
- In Mixing, the goal is to make sure volume levels are where they should be in terms of dialog (dialog pre-mix) / bg music / fx.

Am I roughly okay in my understanding? I started reading a dialog editing book, and haven't got to that stage yet, but I think I get more out of asking dumb questions on forums, than by reading a chapter in a book.

I know I'm making it simple, but I'm just trying to get a sketch for what my approach to the work should be, and I'm flexible, and have no religion regarding workflow or process, and am open to anything.

And yes Scott, I agree, going clip by clip is not a big deal at all :)

It's a lovely day here in Toronto outside my window. I wish you guys all a great day too.
Aveek
 
There is only one book on Dialog editing that I know of and it is really good. The only think I would flag about the book is that he is talking about dialog editing on big budget features. You can do it that way but there are some things done on big budgets for reasons that don't really apply to no/low budget films. The dialog editing book may not get much into dialog mixing, can't remember, but on a big budget film those tasks are almost always done by different people. And the splitting up of tasks in the workflow is part of why some things are done the way they are.

For instance, the reason most mixers don't want the editors to much about much with the EQ etc is because they don't want processed sounds. If they get a file you "fixed" and it sounds off they don't know if it's because your system or your hearing was off or because that is the way the file sounds. When the director get's stinky because it doesn't sound the way it should "and it didn't sound that way before" you have no real response because you don't know. And the mix is happening in a calibrated room that has been tuned for the big screen and dialog editors are often working with headphones two or three to a room, so they are very poor choices to make sonic decisions.

None of the above applies to OMB posts. You know what it sounded like and you know what you did to it, no surprises. And you are going to mix in the box 99% of the time on the same system you are editing on.

I would check out the stickies, I believe there is a fair amount of info there that will balance out the book. None of the practices are a bad idea but some are a bit of a waste of time on a no/low film, because they are more about making sure the next link in the chain can easily take over and do their job and less about the most efficient way to do something.

The two big problems with trying to use a compressor to "level" dialog are that you don't want it all the same level, you want on average to be about the same loudness. The compressor will attempt to squeeze everything to the same level and that can suck the life right out of the dialog.

Listen to just about any top 40 pop song of today and then one mixed in the 50's or 60's, you may not like the music but you will feel the dynamicness of not having everything compressed to death.

The second reason is you are going to be boosting the noise periodically and that can make the edits pop, or the background "pump".

It's used in the mix to push the dialog forward a bit on a track that would play fine with out the compressor.
 
I've got a whole sackful of external compressors, and now do everything in the machine, and I really don't have a favourite - although I probably use the Cubase Dynamics VST when I need a compressor as it has lots of tweaks and adjustments - but they are for when all else fails, there really isn't a need for them as a regular tool - fader automation, groups and mutes work pretty well for me.
 
In the old days we used Neve analogue compressors and you could do anything with them and they sounded great, most of my TV and Film work has been using AMS Neve Logic 1,2,and 3 consoles and the current DFC type that are used at Skywalker and Warner Brothers. Their compressors were similar to the Calrec analogue ones and I helped design some of the features in the early 90's.

I use Pro Tools at home and mainly use the Waves plug in compressors and they have a great simple one : http://www.waves.com/plugins/renaissance-vox

Compressors have always been used in my 35 years in film and TV but it is all about if you can actually hear them sucking the life out of everything or using them to improve things and make dialogue punch through the soundstage.
 
Last edited:
Instead of using compressors, I ride the hell out of my faders. It's much more natural. Compressors lower the dynamic range while bringing up the noise floor. If I do use compression, the threshold is set really high (with a fast attack and medium release) to catch transients that I couldn't deal with while riding the faders. However, if there are a lot of busy sound effects, like heavy ambiences, weapons, and music going on in the scene, I tend to use compression to help the dialog cut through the mix.
 
Ryan,
You warned me against using the leveler in RX before and you advised me to "ride the faders." I'm now assuming that's because the leveler works like a compressor. I've been googling to find out what that "ride the faders" really means, and it seems to me like it is manual compression, or adjusting the gain manually. If that's the case, wouldn't raising gain in certain places raise the noise floor in those areas also?

The problem I have is that my main actor speaks at almost a level of 8 to 10db lower than everybody else in the scene. It's crazy how low her voice is. We realized it during the shoot, but what can I do? I can hear the noise floor go up when I use the leveler or use manual gain adjustment.

So that's my problem. I was thinking, if using a compressor raises the noise floor across the board, maybe that's okay, if I do NR after the compression, since the noise print being used for the NR would at least now be representative of the raised noise floor.

Anyway, I'm going to go back to editing and stop thinking about audio for several days. It's driving me insane.
 
I never did compression during the dialoge edit, even if I was doing everything myself, including the mix. I personally like to do things in stages. It allows me to keep track of what I've done better. If I have to run a compressor on the dialogue, I may put it on the dialogue sub mix bus, but that's it. I hate having to use it. It squishes the dialogue. Maybe only during really heavy action scenes where there are a lot of sound effects happening.
 
trueindie said:
Ryan,
You warned me against using the leveler in RX before and you advised me to "ride the faders." I'm now assuming that's because the leveler works like a compressor. I've been googling to find out what that "ride the faders" really means, and it seems to me like it is manual compression, or adjusting the gain manually. If that's the case, wouldn't raising gain in certain places raise the noise floor in those areas also?

The problem I have is that my main actor speaks at almost a level of 8 to 10db lower than everybody else in the scene. It's crazy how low her voice is. We realized it during the shoot, but what can I do? I can hear the noise floor go up when I use the leveler or use manual gain adjustment.

So that's my problem. I was thinking, if using a compressor raises the noise floor across the board, maybe that's okay, if I do NR after the compression, since the noise print being used for the NR would at least now be representative of the raised noise floor.

Anyway, I'm going to go back to editing and stop thinking about audio for several days. It's driving me insane.
Click to expand...

Yes, "riding faders" is sort of like compression but you don't get the artifacts (pumping, distortion, etc.) that compression can introduce. Plus, part of the fun is using faders! I think a lot of times people overuse and use compression for the wrong reasons and just because you "can" do it doesn't mean you should. I always ask myself, does this particular piece of audio need EQ? Does it need compression? You don't always have to use NR, EQ, compression. If you have a great recording, then why try to mess with it? I've mixed many things in the past year without the need for compression.

You must be dealing with some extremely noisy recordings. Try to do NR before it hits any processing. It might yield better results for you. I try to make an imperfect recording manageable first, before doing any dynamic processing or subjective EQ. So If I open an OMF/AAF and one scene has horrible traffic noise, the first thing I do is try to NR that scene or denoise it. I try not to NR too much if I'm working with another mixer, but if I'm doing the mixing, then I'll NR the full way since it'll save time for me during the mix. I don't think I've ever used NR after compression - since after compression, the audio you want and the noise floor are now closer in level to each other and that's the last thing I would want to do. I actually try to push the noise and usable audio further away from each other so that there's a higher signal to noise ratio. The tools that help me do that are expanders (multi-band and single band), some EQ, iZotope, and other NR tools.

I advised using Leveler since it "bakes" the clip gain into a rendered audio region in Pro Tools. You can use "Clip Gain" in Pro Tools instead. In fact, that's what I do and what many editors do as a "premix" so that they don't have to touch the main Pro Tools faders, which is what the mixers would use if it's an "in the box" mix. If there are really fast and powerful syllables in the dialog that I can't bring down with faders, then I'll go in and use clip gain to lower it for just that syllable. But for me, I'm all over faders. My volume automation lanes look insane after I'm done with a mix.
 
When faced with really low levels I have at times used the GAIN plugin to boost the level of the clip so that there is some room to actually mix.

In some ways level riding is like compression, the biggest difference is that you get to control when and how much. On some clips the compressor may be great and on others not good at all, if you just go and do everything some things will get better (maybe) but others will get worse.

Personally I tend to end up using NR after things like GAIN. If the noise is really bad maybe before and after. The problem is NR is very time consuming to do well, partially because of the "is it better or just different?" question you have to do as you play with the settings (for each clip). And second because you really want to do it before you do a lot of volume automation etc. So you need to pull out the handles so you have some extra to fade with do the NR and then trim it down and put the fades back. I also always save a before version to a UNUSED track that is disabled in case I made a poor choice and need to undo later.

However often you don't need NR. If you are doing a full sound post what you need depends a lot on the "world" the scene takes place in. If it's under an over pass some noise on the production is going to completely masked by all the cars you are going to layer in.

You will have to ease the higher noise floor in and out or do it on cuts that make sense and that may mean you need to add some noise to the "cleaner tracks". But again the noise may not be an issue depending on what else is going on.

This is where you have it all over a big budget dialog edit where you need it to be able to play smooth on it's own.
 
You must log in or register to reply here.
Back
Top