mixing dialog - basics

[soundslikejoe]

I'm working on a film. The main voice over was recorded poorly and has many different average and peak levels. Peak tends to range from -7 to -25db and the averages just as wide.

I've experimented with the VO by Normalizing the various clips to -7db and it gave the VO track a tighter uniform feeling. When mixing music I always avoid Normalization. Are there reasons that I shouldn't Normalize the VO or does the occasion warrant it's use?

Another thought I had.... I'm mixing in Stereo and the movie will play in small theaters. Should I be concerned about the VO's image placement in the mix? Currently I have it mixed to stereo (hard L&R). Would it be better to use the Center position for a VO? Or LCR?

(I read another thread that mentioned something about center channel issues when translating stereo mixes to 5.1)

Any other dialog related tips would be a bonus and appreciated.

[Henchman]

Keep your VO in the center.

[GooberNumber9]

If normalizing works for you, cool. If it sounds good, it is good. I have found for myself in the past that I get the best sound by manually automating each clip, but that can take a long time. Once I get it close I use a little compression and limiting to smooth out the changes and make sure the final levels are where I want them.

If time is more important than sound quality, I start with more aggressive compression and then go back and automate to fix the spots that are still way off level-wise.

Todd Wilcox

[soundslikejoe]

In an ideal world, I would have tracked the VoiceOver here and wouldn't have these types of problems. It was tracked very poorly without any compression on the front end and the end result was 20db of range.

The director had acutally used automation when editing in FinalCutPro but to me the drawn curves were not precise enough.

[kenn.michael]

This is precisely what I love doing in Nuendo - leveling dialogue... I mix in both Nuendo and ProTools, but I do dialogue exclusively in Nuendo because of their brilliant feature where each region has it's own volume parameter (independent of track automation) that is a handle that you drag up and down inside the region... You literally split your regions into smaller ones, do the level changes (Nuendo also redraws the waveform underneath in realtime to show you the level change) - do crossfades to smooth bigger level jumps and then you have a natural sounding track that is very balanced... Through a compressor on top of that and you don't have to work hard with tons of automation during the mix.

But, if you're not using Nuendo, I would definitely use volume automation to balance a dialogue track before compression.

[soundslikejoe]

I'm using Cubase, which has the same clip-gain handle, and agree totally with your method. I'm acutally using those handles to shape the dialog phrasing and add or remove emphasis to passages.

[Dickens]

Originally Posted by Henchman

Keep your VO in the center.

Would this advice hold consistent if you were mixing only on a stereo LR system?
How about panning dialog when 2 actors can be together but at a distance in the panaramic LR spread...or walking from L to R?

[philper]

Originally Posted by Dickens

Would this advice hold consistent if you were mixing only on a stereo LR system?
How about panning dialog when 2 actors can be together but at a distance in the panaramic LR spread...or walking from L to R?

This is not film convention, since people sit in different places in the theatre and this technique affects intelligibility (and is distracting). I usually only hear this sort of panning in animated films, and then it is subtle. But it your director wants it......

Philip Perkins CAS

[soundboy]

Keep your DX center. Here's why: from cut to cut the pov is going to change, and if you are panning to match, however subtely, it will be a distracting mess. In a LR mix, the dx is in what is called the "phantom center". That point, in a large theater, will drift a little depending where you sit in the audience.

[georgia]

On occassion, and sparingly.... With really bad audio on some indie projects... I pan the DIAG to CENTER and then take a small amount of CENTER % out... just a touch. It "smears" the diag a tiny bit across the front so in a large theatre the noise floor is spread across the entire theatre and not just coming from a "Point" source.
Works great... Just don't go overboard. Not enough to notice the Dialogue is not center only.

cheers
geo

[soundslikejoe]

I'm near the end and went with a combination between diagetic and VoiceOver. VO is totally center and DiagX is soft-pan LR. I'm interested to see how it plays in the theater (sounds good here). I'm hoping that the VO will have a focused middle location and the live action/dialog will fill the stereo image without a loss of claritiy.

It's almost done... !!!

[soundboy]

Originally Posted by georgia

On occassion, and sparingly.... With really bad audio on some indie projects... I pan the DIAG to CENTER and then take a small amount of CENTER % out... just a touch. It "smears" the diag a tiny bit across the front so in a large theatre the noise floor is spread across the entire theatre and not just coming from a "Point" source.
Works great... Just don't go overboard. Not enough to notice the Dialogue is not center only.

cheers
geo

Nice!

[starcrash13]

Originally Posted by joenovice

I'm near the end and went with a combination between diagetic and VoiceOver. VO is totally center and DiagX is soft-pan LR. I'm interested to see how it plays in the theater (sounds good here). I'm hoping that the VO will have a focused middle location and the live action/dialog will fill the stereo image without a loss of claritiy.

I can't wait either. Please post your results. I have a feeling you're going to be in a for quite a surprise at the screening. There's a reason that dialogue has been mixed mono dead center since practically day one (apparently there were a few experiments in the early days of stereo soundtracks that yielded unsuccessful results). As Soundboy pointed out, the edits will cause panning confusion.

I suppose in this case, it won't be quite as bad because you won't have anything playing out of the center channel. The disparity between what comes out in the C channel and what comes out L, R becomes very apparent in a theater compared to monitoring on nearfields or headphones. You might find that voices jumping around to be very strange and distracting.

[soundslikejoe]

Originally Posted by starcrash13

I can't wait either. Please post your results. I have a feeling you're going to be in a for quite a surprise at the screening. There's a reason that dialogue has been mixed mono dead center since practically day one (apparently there were a few experiments in the early days of stereo soundtracks that yielded unsuccessful results). As Soundboy pointed out, the edits will cause panning confusion.

I suppose in this case, it won't be quite as bad because you won't have anything playing out of the center channel. The disparity between what comes out in the C channel and what comes out L, R becomes very apparent in a theater compared to monitoring on nearfields or headphones. You might find that voices jumping around to be very strange and distracting.

That's scary..... if I mixed VO and Diagx to center the movie would be mono, excluding the music. Its a documentary and the Diagx track contains all of the ambience. Since the budget is so low there is no ADR. All sounds are camera capture and tracked with some dialog; trucks, crowd noise, etc.

You say nothing will play out of center. Is that true? Why would that happen if the diagetic is not hard panned? It seemed logical that panning the tracks slightly LR would create some stereo width without leaving the C channel silent.

Soundboy said

and if you are panning to match,

I'm not panning to match screen position. My panning is stationary and doesn't follow on-screen positions. Does that make a difference in terms of confusing the audience?

[starcrash13]

I'm confused. If your sync production sound is mono, how are you "soft-pan"-ing it? I assumed that you were panning back and forth according to who is talking. How are you somehow making it stereo? I would understand if you were mixing 5.1 or LCR and had the narration VO out the center channel and the sync production sound out L, R. But since we're only talking about 2 channel stereo, playing a mono source equally out of L and R (I'm assuming you bussed you dialogue premix to a stereo bus) is still mono and phantom center no matter how much you equally pan each side it in towards center or completely out 100%. If you begin to pan the two channels unevenly, then you will just be steering the entire mono source left and right. This would be no different than if you would've just kept it as a mono track or bus and panned it back and forth since the material in the L and R is identical. This is basic signal flow.

Or, am I misunderstanding your method. How are you "stereo-izing" your mono sync sound?

In most films, the original sync production is captured mono and stays mono in the final mix. This is a film sound convention. In a simple stereo mix, ambience backgrounds are edited in that will be mixed stereo. Foley and hard FX will be mostly mono with stereo reverb (of course, there are exceptions), and music will be mixed stereo.

p.s. Don't frown on mono. Some of the best things in life are mono.

[starcrash13]

Originally Posted by joenovice

You say nothing will play out of center. Is that true? Why would that happen if the diagetic is not hard panned? It seemed logical that panning the tracks slightly LR would create some stereo width without leaving the C channel silent.

The C channel will be silent because there will be no signal sent to it. You're mixing stereo L, R which will create a "phantom center" on the screen. As mentioned, this "center" will move around depending on where you sit. The center channel only comes into play in a surround mix. What's great about the center channel is that it always remains in the center regardless of listener position. This means that dialog and action (production FX, foley, hard FX) stay anchored to the screen.

[soundslikejoe]

Man... your comments are helpful. Thanks.... As is obvious, I'm still learning a bunch of stuff.

The project was delivered to me in multiple mono; ex. VO has two tracks panned hard LR. Starting from that point I had assumed that positioning the VO to center and the DiagX to LR would create a stereo image; seperating the VO from the production sound. I think I see an error in my thinking at this point; dual mono can't be evenly panned/spread in a 2.0 enviroment b/c it will still be mono. Right?

Could the mix be up-converted (UnWrap HD) to be more presentable for 5.1 playback? It will sound like poop to have a big hole in the center of a theater's playback.

Would it be better to revamp the mix to LCR or 5.1? Should I just finish the mix in stereo to the director's likeing, chalk this up to experience, and forget about good theater playback?

Helpful people who know these things and comment ROCK, IMO. I've learned more at this website than I could have alone.

[Henchman]

In the movie "Children of me", there wa quite a bit of dynamic panning going on with the dialog, and I really did not like it.
It reminded me of what I hear in Videogames alot.

[1150post]

Originally Posted by georgia

On occassion, and sparingly.... With really bad audio on some indie projects... I pan the DIAG to CENTER and then take a small amount of CENTER % out... just a touch. It "smears" the diag a tiny bit across the front so in a large theatre the noise floor is spread across the entire theatre and not just coming from a "Point" source.
Works great... Just don't go overboard. Not enough to notice the Dialogue is not center only.

cheers
geo

Great advice! This is really effective, in my experience.

[maria]

Hi,

I always send the dialog as a mono channel to the center.From Geo's post it seems like she usually sends the dialog out of all 3 front speakers.Is this how I should deal with dialog, or did i not understand Geo's post correctly?

BTW, thanks for the smearing tip

[soundslikejoe]

Well... I concerned now. The concept of having a "phantom center" channel doesn't seem appealing.

To clarify my understanding - The playback of a 2.0 mix will not output any sound to the Center channel in 5.1 systems.

If that's correct, can a 2.0 mix be distributed to the Center speaker? Would this require a LRCS mixdown?

Do theaters never recieve movies mixed in 2.0?

[Jorn Lavoll]

i've never been a fan of clever surround panning stuff. i dont think it has ever helped me submerge myself more into a movie.
i recently directed the norwegian dub of an animated film for the movie theatres, and they did too much panning, and in one scene, they "had to" pan the dialoge to the back speakers... the result is just that the kids stand up in the seats to see how is making the noise in the back of the cinema.

[Henchman]

Originally Posted by Jorn Lavoll

i've never been a fan of clever surround panning stuff. i dont think it has ever helped me submerge myself more into a movie.
i recently directed the norwegian dub of an animated film for the movie theatres, and they did too much panning, and in one scene, they "had to" pan the dialoge to the back speakers... the result is just that the kids stand up in the seats to see how is making the noise in the back of the cinema.

That's where I will draw the line as a mixer and warn the client, that if they insist on having principal dialog in the surround's only, Then they need to be aware that there will be cases where that diaolog will not be heard at all.

[soundslikejoe]

Originally Posted by Henchman

That's where I will draw the line as a mixer and warn the client, that if they insist on having principal dialog in the surround's only, Then they need to be aware that there will be cases where that diaolog will not be heard at all.

Do you consider placing dialog in front L, or R, "surrounds" or "creative" panning?

It seems like it would be a nice effect to place distant or creepy dialog in surrounds; for instance, a horror movie.

[Henchman]

Originally Posted by joenovice

Do you consider placing dialog in front L, or R, "surrounds" or "creative" panning?

It seems like it would be a nice effect to place distant or creepy dialog in surrounds; for instance, a horror movie.

Personnally, I find it a gimmick used by poeple with little experience.
It serves no purpose, other than to annoy.
It's a "hey, look what I can do" trick.

I aso have found it very distracting in movies that i've watched that have done it.
When they re-did the Excorcist in 5.1, they had all sorts of dialog stuff in the surrounds, and it was distracting, and on more than one occasion I was turnign around to see who couldn't keep their mouth shut behind me.

IMO, surround should be used as an effect only. Helicopters and planes flying by.
Bullet's shooting past your head, etc.

But, that's just my opinion.
Do you know that Stanley Kubrick mixed all his films in Mono?

[Jorn Lavoll]

phew, i am not alone on this then
for me it is just breaking the illusion. "oh, look! there's speakers in the back of the room i am sitting in!" I don't need to be reminded of that when I am watching a film

if the picture was in surround, then surround sound would be perfect, e.

[soundslikejoe]

Originally Posted by Henchman

But, that's just my opinion.
Do you know that Stanley Kubrick mixed all his films in Mono?

I didn't know that... although I could imagine that would be the case considering he is of the pre 5.1 generation. (Although I just read his Clockwork Orange was the first to use Dolby noise reduction.)

Personally (as a viewer) I enjoy the 3-D aspect of sound. I don't want important dialog to be in the rear but things moving around the sound-scape (fx voices in horror genre) is fun for me. I guess you guys may not view those voices as dialog though.

[Henchman]

Originally Posted by joenovice

I didn't know that... although I could imagine that would be the case considering he is of the pre 5.1 generation. (Although I just read his Clockwork Orange was the first to use Dolby noise reduction.)

Personally (as a viewer) I enjoy the 3-D aspect of sound. I don't want important dialog to be in the rear but things moving around the sound-scape (fx voices in horror genre) is fun for me. I guess you guys may not view those voices as dialog though.

No, dialog is scripted. SFX voices is a completely different thing.

And if I'm correct, the reason the CAt43 is orange, is because it was developed for mixing "Clockwork Orange".

[minister]

Originally Posted by Henchman

Do you know that Stanley Kubrick mixed all his films in Mono?

except for 2001.

Dolby convinced him to try stereo. the amount of channels he found distracting. he went back to mono after that.

[minister]

Originally Posted by joenovice

(Although I just read his Clockwork Orange was the first to use Dolby noise reduction.)

actually, noise reduction had been used on films up to that point, A CLOCKWORK ORANGE was the first to use it starting with the pre-mixes.