Film - do i compress dialogue?

K

klytus

New member
I'm about to do my first sound mix for film - a 15minute short. I'm really confused about the best way to proceed.

Should I run the dialogue and FX tracks through compressors and limiters? Or just the music? How much etc?

Should I master the music completely, THEN bring it into the mix before mastering it again with everything else?

Or do I leave the music unmastered till the dialogue and fx are all in place to avoid 'overcooking' it?

Please I'm desperate for all the advice I can get!
 
No expert here

I am by no means an expert in what you are asking, but I will tell you how I would try to aproach this. I think I would first get a good ready for mastering mix on the music as well as the dialog together and then get it mastered. That way everything will be blended correctly. As far as compression goes, during final mix down, if you need to use it, you want to do it sparingly. You should not go over 2 to 3 db on the loudest passage.

I hope this helps, Joe
 
much obliged

What you say makes sense, thanks.

Don't spose you know the best way to 'normalise' the dialogue track - some character's voices (females mostly) are significantly quiter than the men they're acting alongside.

Should I individually treat every single line in a wave editor, or leave it till the end and do something 'clever' with compression/normalise...?
 
If each person's voice is on a seperate track, you would be able to control the volume for each. My guess is that they are not. Boy, thats a toughy. I do not have enough experience to know what tools could be used to fix that. If these are in some form of DAW, you could try to automate the volume to control it (sounds like a lot of work). But there are times we got to do what we got to do.

One thought that just came to me as I was typing this was, what are the possibilities of retracking the female voice on a seterate track.

One more thought. If you are using a DAW, copy the dialog track to another track. Then, edit out the male vocal part with a wave editor and then turn it up.

again, lots of luck, Joe
 
klytus

I do this as part of my job. Yes I compress everything. You use compression to fix the dynamics of recording, then after, you can set the levels to suit the dynamics of acting.

I edit each voice track on a daw almost word for word to get all the dialogue at the right level. You just have to be carefull of the background noise (this is why they often post sync the dialogue) if the background is noisy it can be a problem. If you can hear edits or pumping in the background, you can often cover that with music or often we use an atmos track. Basically this is a recording of a room, backyard, beach etc which is put in at low level under dialogue.

If your dialogue background is noisy then you will just have to compromise to get the best balance.

As JC Lives said, I would get the music ready first, you can still tweek it a bit in the mix.

As for fx, it's a bit hard without knowing what you are doing. If you are putting in actual sound fx then they should be at the right level and not require compressing. If you are talking about an atmos track then you should compress as it will be at low level in the mix and needs to be there all the time.

Hope this helps
Brenton
 
Thanks very much both you guys.

Brenton, you mention using compression to 'fix the dynamics of recording.' Would a good start be to find a threshold somewhere between the loud and quiet bits and gently compress the top end down so the fluctuations are less pronounced?

I see you what you mean about background noise - if I cut out each word and boost gain on some, the background hiss is going to really jump out at the listener for that word. Hence the atmos track (our sound man generally grabbed 30 secs to 1 minute of 'room' at each of our locations). But when using an atmos track should I try to cut the words out of the background hiss as tightly as possible, otherwise surely the hiss will 'stack up' onto the atmos track and still produce an audible artefact?

Also (sorry to go on but this is a big responsibility!) when I'm happy with the mix, should I normalise to a peak value or will this harm the mix? And do you apply standard music mastering processes (tube compression, master eq) to your mixes, or would this risk bringing artificial warmth to the dialogue?

Many thx for your help
Klytus
 
Yeah, it's all about what you can get away with. I'm not sure what to tell you because it would change depending on what your audio sounds like at the moment.

If you have an atmos of each room, that's great. Now you can eq quite a bit. Get rid of everything above and below the freq range of your voices. Basically do whatever you can to get rid off the noise, while getting the best sound on the dialogue.

I use noise reduction software and you can usually get up to 5-6 db reduction of background noise before you start to get artifacts. With compression, you only want to lower the louder speech/person and not touch the other stuff.

I don't usually cut out the noise between words, just lower it. If it is fairly noisey background and you are going to definately hear it come in and out, then conceal it, bring it in or out on a vision cut or under a sound fx etc.

Normalise won't affect the dynamic range, it will just bring the loudest part of your material up to the loudest allowable level of your equipment. Feel fee

No I don't do any mastering as such on the finished mix, just the music. Cross referencing on different monitors while mixing has always worked for me.

If all else fails because it is a really noisy background then leave it all in and just compress as much as possible. People get annoyed with having to turn the tv up and down all the time and you don't want a pumping background. Compromise between the two.

Brenton
 
Brenton,
Can you tell us what type of computer, software, and outboard gear you use for audio/video work? I've been asked if I was interested in doing this type work for a friend who films outdoors videos (hunting, etc) but I'm not sure what is involved. What would you recommend (keeping in mind this is a "home recording" environment, not a pro studio)? Most of the work would be enhancing the vocal soundtrack and adding music. Of course, video enhancement would be a plus to make him look better. DAW would be a PIII 700MHz or higher (will start building it in January).
 
Hi DigitalDon

I don't know how much I can help you as I work in a Pro situation, so it will probably all be too expensive for your needs.

I use a Protools Mixplus 24 track system on a Mac G4, no outboard gear (for video anyway) but we have lots of plugins, Waves (eq, reverb, compressors, noise reduction, noise gates etc) lexiverb etc. If you want a full list get back to me. We sync this to the video machines via the Motu Digi Timepiece.

We shoot video on Sony Digicam and edit on either a Media 100 or the Sony Es7 depending on the requirements of the job. the Media 100 is just nonlinear editing and the Sony is a hybrid (both nonlinear and tape to tape). Both of these have a lot of plugins for fx as well.

I really can't help you when it comes to video as I only do the audio side.

Again I don't really have much of an idea about domestic/cheaper stuff. There are quite a few lower end programs for both audio and video that would probably suit you but I don't know any of them well enough to give an opinion. Maybe someone else here can help.

The Digi 001 is now available for PC and is a bargain for the money, so this may be something to look at for the audio side.

Maybe you could search the net for a forum on video production, or check your area for someone doing local video productions for advice.

Sorry I could not be more help. I finish work today for 3 weeks holiday and I won't be on the net, so if you ask me anything else, I'm not ignoring you, I just won't see it until I get back.

Merry Xmas
Brenton
 
Digitaldon,

I'm a newbie at this too, but from what I've heard you might consider taking a close look at Steinberg's Nuendo as the control software for your DAW - it's a full multimedia audio production sequencer which lets you import your movie video into the timeline (I work from a low-res render to save disk space).

Nuendo is standalone, meaning it doesn't require DSP hardware (the advertising literature harps on about this at great length) like Pro Tools but instead relies on the host cpu.

The advantage is of course price - I'm not claiming it's as good as Pro Tools, I really have no idea, but the claims being made are quite something (several indie features have been dubbed using Nuendo).

The downside is you need a fast PC to get a fluid working environment - however I'm assured that 7-800Mhz is quick enough (with lots of RAM and zippy hardisks).

Check it out - for video work I don't think you'll find a more cost-effective solution, as you can use all VST and DirectX plug-ins with Nuendo.

Best
Klytus
 
Klytus - the proccess I use is this

1. work on all the dialogue tracks. This will mean different EQ for each location and each person on screen. I compress all dialogue, some more than others but aiming to achieve maximum loudness. It will mean cleaning up the entry and exit from each location scene. Eqing voice for tv usually involves rolling off the extreme tops (10k plus) and adding mids like 5khz to get presence and rolling off any boominess in the low end. This is all mixed down to a Dialogue Track.

2. Add the fx and music with the fx goint to one track and the music to another. (2 x stereo pairs if in stereo) This allows for the creation of M&E tracks (music and effects tracks) for overseas companies to add their own translations into their native language by simply replacing the dialogue but still using your music and effects. You can now use the atmos tracks to smooth out the edits between different shots within the same scene. Sit the atmos just above the background of your dialogue tracks and it will dominate creating a consistent even background.

3. Don't try to have all the effects and all the music going at once. It becomes a compromise as to what dominates, i.e. either make it work with effects or use the music instead but not both as it gets cluttered. This is determined by your producer/director or the editor you're working with.

Then you can play all tracks back (dialogue, music and effects) and they should combine into a final mix.

It's amazing how many generations TV sound is put through. I estimated that by the time you guys in the US heard the programme Beyond Tomorrow, which I was audio director of, the sound had gone through 15 generations!!!

cheers
john
 
Cheers John, much appreciated. It's really useful to hear how pros like you and Brenton operate, saves me a hell of a lot of faffing around with the basics.

I still have a problem regarding M&S Stereo - have you come across this? Our soundman recorded everything in M&S; apparently this simultaneously records a mono dialogue track AND a stereo ambient track onto the DAT for each take.

Apparently the theory is that you decode the M&S and then have the choice whether to fader in the ambient stereo, or if there's too much background clutter, leave it out and just use the punchy mono dialogue track. Only thing is, I don't know how to go about decoding the M&S when I sample the audio from my DATS. Do I need a special DAT mchine, or mixer, or software?

Best
Klytus



Do you have
 
You must log in or register to reply here.
Back
Top