Mixing multiple mics with different room tones in post?! Help!

E

ericsf

Member
I'm a total newbie when it comes to audio mixing in post need some serious help! I only have 2 days to solve this problem.

I'm editing footage from conference room scenes with 3 cameras and 2-5 actors per scene. Each actor had their own lav and there was a boom over the conference table. The problem is that if I only use each's lav mic, cutting btwn them is jarring because each mic has a slightly different room tone. It doesn't help that they are also placed slightly differently on the actors and the levels aren't all the same.

I normalized the levels so the peaks are no higher than -4.5db and voices are close to equal in loudness. Bec I had to boost some of the levels more than others, the room tone now is also relatively a lot louder on some, and it's very apparent and sounds bad across cuts. Not only is the room tone different on different mics but it's also louder on some tracks. I would love to just go strictly with the boom track but it was far away from some of the actors so the ambient room noise is kinda loud relative to their voices and it doesn't sound clean.

We recorded 20 secs of room tone from every mic. My question is how do I smooth out the room tone differences across cuts from different mics??? Do I:


  1. Use each actor's lav and the lay one room tone track down across the whole scene? If so, which mic's room tone?
  2. Lay multiple room tone tracks from all the diff lav mics?
  3. Fade in and out of room tone?
  4. Use a track for each person's lav and a boom track even though it sounds more echoey?
  5. Do some combination of these or something totally different?
  6. What level should the room tone be at?

I'm frustrated and any help is greatly appreciated being that I only have 2 days to sort this out. Thanks in advance!!!
 
If you have enough separation between your noise floor and the talking, you can use a noise gate to isolate each of the voices in the mics. From there add a slight reverb to blend and smooth the talking back out a bit since the noise gate will chop it up a bit. From there you can add back in your desired room tone from any of the mics.

As far as level goes, well you'll probably have to automate that manually, I wouldn't use a compressor.
 
ericsf said:
I'm a total newbie when it comes to audio mixing in post need some serious help! I only have 2 days to solve this problem.

I'm editing footage from conference room scenes with 3 cameras and 2-5 actors per scene. Each actor had their own lav and there was a boom over the conference table. The problem is that if I only use each's lav mic, cutting btwn them is jarring because each mic has a slightly different room tone. It doesn't help that they are also placed slightly differently on the actors and the levels aren't all the same.

I normalized the levels so the peaks are no higher than -4.5db and voices are close to equal in loudness. Bec I had to boost some of the levels more than others, the room tone now is also relatively a lot louder on some, and it's very apparent and sounds bad across cuts. Not only is the room tone different on different mics but it's also louder on some tracks. I would love to just go strictly with the boom track but it was far away from some of the actors so the ambient room noise is kinda loud relative to their voices and it doesn't sound clean.

We recorded 20 secs of room tone from every mic. My question is how do I smooth out the room tone differences across cuts from different mics??? Do I:
  1. Use each actor's lav and the lay one room tone track down across the whole scene? If so, which mic's room tone?
  2. Lay multiple room tone tracks from all the diff lav mics?
  3. Fade in and out of room tone?
  4. Use a track for each person's lav and a boom track even though it sounds more echoey?
  5. Do some combination of these or something totally different?
  6. What level should the room tone be at?
I'm frustrated and any help is greatly appreciated being that I only have 2 days to sort this out. Thanks in advance!!!
Click to expand...

What DAW are you using?

I would not normalize the levels. It's really hard to answer your questions without hearing everything. There are really no rules, but you just want to make everything as smooth and as unnoticeable as possible. I would not lay multiple room tracks across the entire scene as that'll just be a ball of noise floor. But hey, at least you have 20 seconds of room tone from EVERY mic. That's great!

You'll probably want to separate each lav into a different track since they probably all have the same type of tone. While there are no rules for this... but try utilizing long fades as you checkerboard between each lav. In Pro Tools, the type of fade I use is "Equal Power" with "Standard" checked.

Now that you have those lovely room tone recordings from each lav, you can use those to add some "handles" or room tone before and after each line of dialogue starts for each character. You can make these "handles of noise" (for lack of a better word early this morning) as a way to cleanly transition between each lav. Like I said, you'll probably be using long fades, but just make the fade long enough so it sounds good. There are no rules however, longer fades between two different room tones calls for more noise in the soundtrack. Shorter fades means less noise but maybe more audible edits. It's a compromise.

I would not use the boom (combined with the lav) at all if it sounds "roomy" and "echoey".

There are ways to "push down" the noise floor. One way is using an expander (the opposite of a compressor) set at a really low ratio. The expander will allow you to push the noise down and potentially make the transitions even smoother as if there was no difference in room tone in each lav at all.

I know you only have 2 more days, but all of this I've learned from this book

Like I said, it's hard to describe through text and it's much easier to see it in action in the book.
 
George Tsai said:
If you have enough separation between your noise floor and the talking, you can use a noise gate to isolate each of the voices in the mics. From there add a slight reverb to blend and smooth the talking back out a bit since the noise gate will chop it up a bit. From there you can add back in your desired room tone from any of the mics.

As far as level goes, well you'll probably have to automate that manually, I wouldn't use a compressor.
Click to expand...

A noise gate will be too dramatic. A noise gate is basically an expander with an extremely high ratio. An expander will certainly work a lot better.

I would still add a touch of MONO reverb to add a bit of decay to ease the edits a bit. Works like a charm.

If you're using a DAW, then just automate everything. Use a compressor sparingly as you don't want to raise the noise floor back up again.
 
The real answer is to do a real sound post. That is with a sound person and enough time to do a good job.

What you have (hard to really tell with out hearing it) sounds like what one would expect from that set up, and not a really unusual situation for post.

What a sound person would do (at least one approach) is to NOT normalize but to clean out the ambient noise as much as possible from all the different mics.
Add in some ambience/ room tone to fill in all the holes so you don't have gaps in the "noise'.

AND create the background sound for the scene. At that point the background you have created should "eat" the "noise" in the dialog track and your differences should be unnoticeable.

You will then MIX the sound so the levels are where they should be. Unless there are some seriously low levels that generally will not cause problems with RT changes. If it does you have to deal with it on a case by case basis.

You can't do the above in a picture editor, at least not well.

It takes time. Two days sounds doable if you get someone experienced. Hunting and pecking your way through?????

There is no one button "fix".

-------------------

The above is aimed at your specific question. To get a decent sounding mix you also need to do FX and probably Foley. But you only asked about making the voices smooth.
 
Thanks so much for the speedy responses guys!

To further expose my audio mixing newbiness, I had to google DAW haha. I'm just working on a macbook pro in Final Cut. I've never even had to go as far as opening Soundtrack, so being that I'm short on time, I'm hoping I can do everything I need within FCP. I've never had this complex of an audio setup before and have really only adjust levels before in the past, so I'm not sure about adding reverb, decay, expander, etc. and if it's time efficient for me to try it for the first time now. but if it'd make a significant difference, I'd be willing...

Why wouldn't you normalize levels? I used the automatic 'Apply normalization gain' in FCP, which in my understanding, just applies a gain filter to bring the levels of the peaks to no more than a designated amount (-4.5dB in my case). How would you most efficiently get levels btwn voices about the same and up to a useable output level then? Different mics were at diff levels and actors talk with diff volumes. Again thanks for understanding that I'm a complete audio newbie.

ryanjf, thanks for the suggestions on adding room tone 'handles.' I could see that helping. Sometimes the dialogue back and forth btwn actors is pretty fast, but other times, an actor might talk for 10 sec or a little more. I guess when there are quick back and forths, maybe, using the room tones of all lavs of the dialogue might be ok. And then when one actor talks for longer, I could fade in and out the room tone handles.

At this late stage in the game, I'm just looking to make it watchable without jarring audio on the cuts. I don't need hollywood-clean sound, nor do I have time to attempt to get it. If it has to be a little noisy throughout to cover the cuts, so be it as long as it's within reason and not ridiculously noisy.

Thanks!!!
 
Noiz2, thanks for that info. It's helpful to understand all that. Yeah I would love to do it the right way with a proper sound person, but unfortunately, this fell into my lap and the 11th hour and I gotta make my best newbie hack at it. I know there's never a one-button quick fix to anything, esp not audio, but suggestions/tips that would allow me to get the audio to be passable at this late stage and given my lack of experience would be more than helpful :).
 
I wouldn't normalize levels because it just takes the peak of the file and raises it to whatever you set it at. Say you record dialogue and the average is around -20dBFS, but in the same file, there's a hand clap or a hand slam on a wood desk that peaks at -3dBFS.. and you normalize to -2dBFS. Well, the dialogue (which is what you want to normalize) will only be raised by 1dB due to the fact that a higher peak is in the recording. My point is - it doesn't do anything really. You're better off doing automation and the art of MIXING, which is how you get levels smoothened out. It also depends on the quality recording too. If one mic was recorded at a really low level where the signal to noise ratio is horrible, and you need to raise it up to match the other lavs, then all you're going to get is noise from that one mic. Sure, you can try to "fix" it but even things like that we would just replace or spend a lot of time using various tools to attenuate the noise.

As Scott said, there's no "fix" button that will mix for us (well there are technically a few tools out there, but that's a different thread).

Oh good, there are some fast back and forth between the actors. This is great because the initial attack of dialogue might just be jarring enough to disguise the room tone difference. In this case there is no need to use the room tone. It really really really all depends.

Are you getting more and more of this work? Or is this a "one off" thing?
 
ryanjf said:
I wouldn't normalize levels because it just takes the peak of the file and raises it to whatever you set it at. Say you record dialogue and the average is around -20dBFS, but in the same file, there's a hand clap or a hand slam on a wood desk that peaks at -3dBFS.. and you normalize to -2dBFS. Well, the dialogue (which is what you want to normalize) will only be raised by 1dB due to the fact that a higher peak is in the recording. My point is - it doesn't do anything really. You're better off doing automation and the art of MIXING, which is how you get levels smoothened out. It also depends on the quality recording too. If one mic was recorded at a really low level where the signal to noise ratio is horrible, and you need to raise it up to match the other lavs, then all you're going to get is noise from that one mic. Sure, you can try to "fix" it but even things like that we would just replace or spend a lot of time using various tools to attenuate the noise.

As Scott said, there's no "fix" button that will mix for us (well there are technically a few tools out there, but that's a different thread).

Oh good, there are some fast back and forth between the actors. This is great because the initial attack of dialogue might just be jarring enough to disguise the room tone difference. In this case there is no need to use the room tone. It really really really all depends.

Are you getting more and more of this work? Or is this a "one off" thing?
Click to expand...


I know it always depends and there's not blanket rule, so thanks for trying to give suggestions around that and without being able to hear what I'm working with.

This is kind of a one-off thing. I'm a one man band video producer and am chipping away at learning the ridiculous amount of knowledge in all the disciplines that encompasses, post audio obviously being a huge hole haha. This project is bigger than my normal scope, where I assembled 3 more people for my crew and is def more complicated on all technical levels. I thought I was gonna have a sounds guy to help in post, but it fell through last min, so here I am. All that to say, I'm not sure how many of these situations I'll find myself in in the future, but I want to learn at least the basics, so I can hack it if necessary...

Yeah with respect to normalizing levels, what you described is what I thought was happening. Not perfect, but at least a place to maybe start before going in and adjusting levels by hand? and yes, there is one mic that was not placed so great and the voice is quite so when I boost the level, the noise is really loud. That was the one mic where I was thinking of supplementing with the boom track, but not sure how that's gonna work. I'm not gonna be getting much sleep tonight...
 
A lesson to be learned here is also the importance of a sound person ON THE SET as you shoot. You mentioned that you had a boom as well as the multiple lavs but it was unusable because some of the actors were off-mic and the overall timbre had to much reverb in it. Well, an experienced sound person would have made sure that the boom was positioned close enough to the actor speaking and aimed directly at him and that it was re-aimed between each line in order to follow the dialog flow between the actors. A boom operator is a skilled member of the crew, far more than just a living mic stand, while a mic on a stationary boom, more often than not, produces results such as here that you have realized are marginal.
 
There really isn't a good way to make good sounding dialogue tracks in FCP based on what you've said (and this is based on just my experience). An experienced dialogue editor working in Pro Tools or Nuendo could probably hammer this out for you, with a good dialogue pre-dub, but it seems that isn't going to be the case here.

For what you have to work with: Throw an expander on the lav tracks, checkerboard the audio, get rid of as much room tone on the dialogue tracks as possible, create a new room tone and have that be constant underneath the dialogue track. If you start cutting between two different room tones between dialogue tracks, and if they are different enough, it is going to be really​ noticeable.
 
you know that this kinda things are "basics" for soundpost-people ?

"cleaning up dialogue track"
" filling up holes with roomtone"
" replacing dialogues with other takes to make it sound better"

Is this something you have to do completly yourself or can you ask for practical help ?
 
You must log in or register to reply here.
Back
Top