Postproduction basics

[nycineaste]

you want specifics OP? here you go-- notice how you wont fully understand some of the terms further proving that its a huge subject:

-- start by low cut filtering anything below its fundamental frequency-- what's the fundamental for vocals? check a freq analyzer, depends on teh voice
--cut frequencies before you boost any
-- you have to learn to use a compressor- this is a whole lesson in and of itself- go to google type in how to compress vocals- if you insist on diy-ing it without bothering to learn attack-release-etc use a big dumb compressor vstr plugin like camelphat free and go easy (vegas takes vsts, i dont know if your nle does)

EQ and compression- your two big hurdles google some EQ tutorials find out what a fundamental freq is how to use a compressor it wont happen overnight

loosely:

up to 80hz subbass remove (hi pass) from almost everything that doenst need kabooomyness
80-250 bass
250-1000 low mids, the mud zone beware
1000-5000 boost for intelligibilty in a mix ALL SOUNDS CANT BE UPFRONT!
5000-8000 treble
8000+ super highs add sparkle, unecessary for most sounds

there you just got a wonky awkward start with big giant holes missing but thats what you asked for

go forth!!

 

[KINOKS]

However the OP wants someone to tell him the "secret" so with no experience and the wrong tools he can do a great job. That just isn't realistic.

Well this got skewed a bit! I don't want any secrets, just some usual procedures that are done and what effects are used. What can I do in Avid and what to do in Audition. I'm not a total newb and I learn fast. If something isn't clear to me I read it up. No need to philosophy here. Really. It doesn't need to get "political". Jeeez
My questions a broad because I don't know how to ask targeted questions. Direct me in the right direction or something... Telling me to go read a book or to hire a professional is of little help. And as I've already said if I'll see down the line that I won't be able to manage I'll hire a pro but in any case I'll get a pro to check it out.

Just some basic things to do would be nice. Some basic work that is done, like normalizing, compressing, gaining, de-noising, equalizing...etc. Surely you've got a basic procedure that you do when you get a doc in your hands? I'm not asking for high spiritual stuff here, just a basic workflow of what must be done. That's all. If in all seriousness this is to hard and writing long essays about me building houses with a hammer is easier for you then go ahead. Maybe we'll build a small backyard shed together!

 

[KINOKS]

you want specifics OP? here you go-- notice how you wont fully understand some of the terms further proving that its a huge subject:

-- start by low cut filtering anything below its fundamental frequency-- what's the fundamental for vocals? check a freq analyzer, depends on teh voice
--cut frequencies before you boost any
-- you have to learn to use a compressor- this is a whole lesson in and of itself- go to google type in how to compress vocals- if you insist on diy-ing it without bothering to learn attack-release-etc use a big dumb compressor vstr plugin like camelphat free and go easy (vegas takes vsts, i dont know if your nle does)

EQ and compression- your two big hurdles google some EQ tutorials find out what a fundamental freq is how to use a compressor it wont happen overnight

loosely:

up to 80hz subbass remove (hi pass) from almost everything that doenst need kabooomyness
80-250 bass
250-1000 low mids, the mud zone beware
1000-5000 boost for intelligibilty in a mix ALL SOUNDS CANT BE UPFRONT!
5000-8000 treble
8000+ super highs add sparkle, unecessary for most sounds

there you just got a wonky awkward start with big giant holes missing but thats what you asked for

go forth!!

Great. That's what I need. It's not hard for me to go and read up on how to use a compressor. I just need to know firstly that I need to use it. But regarding the stuff you wrote, unfortunately I already know most about it and have already done it a couple of times
Anything else??
But still, it's great! THx.

 

[KINOKS]

A lot have been said and I think the general message has come across. I'll try adressing one of the questions.


Ok, so you want to know what possible deficiencies to listen for and how they sound. Here are some things you might think about:
First and foremost is the dialogue, which is usually a key element in driving the story. Is the background too loud, making it less intelligible? Are there any distracting noises that should be removed? Does the tonal quality or level change a lot at any given point between actors or between shots? Does it change at all and follow the perspective of the shot (wide shots sounding more distant than close ups etc)? Is the tonal quality balanced, or does it all sound a bit sharp, tinny, or muddy? Ultimately, can you hear what they say without making an effort?
As to fx and music, these are very much in the artistic realm and there are few rules. Does the levels work, or are some things to loud or soft? Is something panned in a funny way? Does the fx seem like they belong to the things causing the sounds (a gun sounds like a gun and not something else) and originate from the filmic world, and if not, is there an artistic reason for it?
Wrong usage of common processing tools like eq, compressors, gating and noise removal are easiest learned by trying it out yourself. Take an audio track and play around with it, and notice what happens when you do too much.

Okey great. Thx! But I know about all this to. I've been working in production for some time now. I've also been an editor on a TV station, so I'm really not a blank slate regarding audio. Levels or I mean perceived loudness must be the same, the voice must be clear, the colour must be the same, removing distracting noises, wide shots should have the voice/sound appear more distant (that's why it's kinda bad when using lavs in wider shots because the sound is right in your face but the picture isn't, right?), yes I know But how to achieve all this? Do you use a compressor to gain then audio or do you first normalize and the compress, etc. I don't know? For what do you use an EQ? Do you use it to colour the voice of let's say for instance you had the same talent but it was shot once with a different mic...?

 

[davefisk]

Here's the thing.

Every project is different, thus every project has its own set of challenges. Knowing how to use a compressor or eq or normalization isn't as important as knowing when to use them, and the only way to get good at knowing when to use them is with experience.

So there is that, and then there is all the tech stuff you need to know...frequencies, rolloff...all the terms and techniques that many of us spent years learning either in school or on the job. I can tell you how to do something, but if you don't know half of the words I'm using it won't be of much help. Yes, you can then go look up each word that you don't understand, but at that point you might as well at least start with one of the books already mentioned, which you said you don't really want to do.

The one thing nobody has really said is that aside from the boom operator on production, the sound editors and re-recording mixers have some of the toughest jobs in film making. It is the job of those people to take what is given to them and make it sound good and add to it. I have to be quite frank and say that I've set up lights, I've run cameras...I've done some of the other things and they come nowhere close to the amount of stuff you need to know to do proper sound post on a film. That's why your camera department will have a few people, but a proper sound post department on a feature film will have an army of people working on it. On The Matrix Reloaded there were easily 30 of us that had our hands on that thing in one form or another.

So, I guess the short of it is, when someone who doesn't know about audio post production asks what's the best way to do audio post production..well..that's kind of like me going to Ferrari and asking "how do I build an Enzo? I don't want to read the books you told me to read or go to school for mechanical engineering or anything like that. Just tell me how to do it. I'll figure out the stuff I don't understand on my own". Even then I wouldn't be able to build an Enzo properly.

 

[KINOKS]

Write, but I'm not asking you how to build an Enzo, just how to tune up my old moped.

 

[davefisk]

Write, but I'm not asking you how to build an Enzo, just how to tune up my old moped.

...and I think that's where the communication gap lies, and I also think it is telling of what you feel really goes on in audio post. What goes on in sound post isn't tuning up an old moped, no matter how basic you want to go. Even then, if you don't know the basics of an internal combustion engine you won't know how to tune it up. Similarly, if you don't know the basics of audio post, it's hard for someone to tell you what to do, especially, like I said, when each project is different and requires different audio treatment done to it.

Even if I said "apply 'x' amount of compression on this track and do 'y' and 'z' on this other stuff" you still have to know the software, and you have to know how to cut dialogue and edit effects, Foley, and then do sound design...and then you have to know how to do pre-dubs and the final mix. You have to know when to use each tool set at your disposal and how to use them, and those are things that you can't just learn on a forum. You learn them by studying them in school or interning with a audio post production facility, and then taking what you learn and spend years perfecting those skills.

I think this is why post production audio is such a mystery to people who don't know it. If you don't know it, then you don't know what you don't know. Plus, there is more than one way to skin a cat in the post audio world. Not everybody is going to do something the same way. This really isn't so with pulling focus or a dolly shot. I know a lot of high level production sound mixers that don't really have an idea what goes on in post, and they are audio guys! This is why "hire a professional" is the go to response for these types of questions. If nothing else, they can explain what they are doing to your audio and why they are doing it, and they can maybe teach you along the way if they are willing.

 

[Noiz2]

Write, but I'm not asking you how to build an Enzo, just how to tune up my old moped.

Well that would depend on what is wrong with your moped.

BTW there probably isn't a single regularly working dialog editor that normalizes the tracks.

Since you know visuals how about you tell me what I should do with my footage. Not everything but just the stuff you do to every clip always.

It would be truly scary if you applied the same filters to every clip of all your footage for every film. So why is it that picture folks always thing sound folks apply the same things to every clip?

Here is your workflow.

Clean up the dialog. Get rid of all you don't need and remove hums and noises. Pull out the clips where they got chopped in your NLE.
Add some room tone if needed to fill in gaps.

Build up the ambiences and backgrounds so they feel "real" and match the world of your film.
Yes you do that for documentaries also.

Add in hard FX that are missing.

Probably add some score.

Balance it all and then mix it so it sounds the way you want it.

In that process you will probably use some EQ, some noise reduction, a lot of editing. Exactly where is very project dependent though NR is almost always at the clip level.

You don't "normalize" because you "mix". You might add some compression on dialog to help it cut through and you will probably add a limiter at the end to help keep you from exceeding the spec.

you may also have to go get a pricy plugin or pricier box to get a Dialog Normal level.

You can't do it in a NLE because it works at the frame level and you need to be working at the subframe level.

You also need a calibrated and tuned room to edit in.

And you really also need five or ten years of experience to not mess it all up, but...

You should have a sound library for the FX and BG's, though on a doc you are usually using the wild tracks you recorded in the field for the BG's. Unless you didn't record any wild FX in the field. In which case you are going to have to "fake it" and pretend you didn't. ie You are going to have to find the recordings of the correct backgrounds someone else did.

There it is. A plan that works for every film.

 

[GaryNattrass]

and after all of the above two posts be prepared for the client to be unhappy and then want to change everything at the last minute, then be prepared for them to re-edit and re-hash the programme so you have to start all over again as they have re-shot the middle scene but sadly it's been done in SD NTSC at 29.97 not HD at 25fps, they also used an on-camera rode videomic but can you get it to match the sanken cos 11's used in the main interviews.

...............then in six months they will normally come back and ask for an M&E as they just re-edited it again and have sold it to botswana!

oh and by the way international sales also require a version for promo's that has all the music removed.

It's all about the prep and doing certain things like split mixes can also be of great help at the initial stages but if your NLE can't handle it all you are always going to be riding your moped in the race!

;0)

 

[KINOKS]

Here is your workflow.

Clean up the dialog. Get rid of all you don't need and remove hums and noises. Pull out the clips where they got chopped in your NLE.
Add some room tone if needed to fill in gaps.

Build up the ambiences and backgrounds so they feel "real" and match the world of your film.
Yes you do that for documentaries also.

Add in hard FX that are missing.

Probably add some score.

Balance it all and then mix it so it sounds the way you want it.

In that process you will probably use some EQ, some noise reduction, a lot of editing. Exactly where is very project dependent though NR is almost always at the clip level.

You don't "normalize" because you "mix". You might add some compression on dialog to help it cut through and you will probably add a limiter at the end to help keep you from exceeding the spec.

you may also have to go get a pricy plugin or pricier box to get a Dialog Normal level.

You can't do it in a NLE because it works at the frame level and you need to be working at the subframe level.

You also need a calibrated and tuned room to edit in.

You should have a sound library for the FX and BG's, though on a doc you are usually using the wild tracks you recorded in the field for the BG's. Unless you didn't record any wild FX in the field. In which case you are going to have to "fake it" and pretend you didn't. ie You are going to have to find the recordings of the correct backgrounds someone else did.

Thanks!

I've already laid down the music in my NLE and I've already put BG/room tone in between the cuts. In some instances it was recorded but in others it wasn't. How can you fake this? And let me ask you this: if or when I remove the noise from the speakers audio is the room tone still necessary between the cuts? Seeing as the noise is the only thing that's possibly heard (in some instances in others there's some other non distracting stuff). And when do you remove the noise? After, before or somewhere in the middle? I would tend to think that you do it before compression as compression can boost the noise up, right?

But still, what do you do firstly? Do you first compress and bring the levels up with that and then use an EQ? Or is it the other way around? Seeing as when using an EQ you basically use gain to boost certain frequencies it's possible to go over your desired dBu limit. So it makes sense to use the compressor after EQing...?

And how far bellow the peak level should the normal dialogue be?

 

[recordingwhiz]

and the correct answer to all your questions is : "it depends on what is needed"

I for one try to compress and EQ as little as possible (depending on the raw tracks im given) and use track automation to level things out, instead using these tools to create a balance between takes and to match other tonalities with other tracks in the same scene and to give continuity through out the project.

Good luck.

 

[High1ander]

I agree with the idea that seems to float around regarding sound that it seems so hard and people are offended by the notion that a newb can do it. It isn't generally the same treatment against beginners wanting to do "video". I don't think anyone here says that anyone can do good sound (because even "good sound" is hard), and it is easy PC, etc. or that there isn't great value in hiring a proper sound girl. I know that in the US, they are more strict with audiolevels and videolevels, compared to here in Sweden. But, I have done plenty of work for Swedish, Norwegian and Danish TV, national TV that is. The requirement has been stuff like, no peak over -9 db. I am no complete newb when it comes to sound, but it is really the area that I feel the least amount of desire to work with (because I have no flashy scopes to tell me if I am dead wrong), and that it requires re-listening to death sometimes on a small soundbite... That said, I mix in some music, narration, background noise, etc. I just make sure narration is louder than the rest, and that is fine, even for commercials. Now I spend maybe 30-60 minutes of tweaking audio for a 1 min commercial/TV-spot, etc. I feel I could spend a good 3-4 hours more to perfect it (well, perfect would require way more), but hey, 30-60 min and no complains. I tweak in Premiere, sometimes Audition and if there is really bad sound I will use other programes and plugin-ins to clean it, I sometimes have no clue what kind/brand of headphones I am using, because they are being switched around all the time at the office. Now, I know for sure I lack a lot of the amazing skills lots of you guys have that work with audio every single day, I prefer having someone else do audio, but it feels pretty good to know I can do it myself and get OK results. It means I can do one-man band type of shoots.

 

[Rick R]

and the correct answer to all your questions is : "it depends on what is needed"
I for one try to compress and EQ as little as possible (depending on the raw tracks im given) and use track automation to level things out, instead using these tools to create a balance between takes and to match other tonalities with other tracks in the same scene and to give continuity through out the project. Good luck.

+1 I normally level things out with volume envelopes and such. Depending on the audio, I resort to EQ and compression as needed on an event/clip basis, be it a multiband comp or broadband limiter. On dialog I normally add a high/low pass filters on the event, track or submaster. NR or other restoration type processing would be on individual events or a track... If severe, I'll export it to a dedicated audio application for 'surgery'. If using processing on the master buss, its usually a brick-wall limiter to catch peaks from exceeding a set value.. for the most part, it rarely kicks in at all.
If this not an answer your looking for, keep posting on many forums and you may get the answer you want... be it right or wrong.

 

[Inveresk]

I agree with the idea that seems to float around regarding sound that it seems so hard and people are offended by the notion that a newb can do it. It isn't generally the same treatment against beginners wanting to do "video".

It appears to me that the OP simply wanted to understand the postproduction process. He's not alone there. I saw no implication that he thought he could the job as well as a seasoned pro. I've been a chartered architect for close on 40 years and I would have no difficulty in explaining to a lay person a model process for designing a building. If a newbie tries their hand at it, they're not going to be able to do it as well as I am but that's singularly obvious and doesn't need stated.

To those who have posted a list of the post production steps and the order in which they'd normally be carried out, thanks for doing so. I find that information helpful and informative. I don't need a litany of what might go wrong at each stage in that process to understand the workflow. That's not to say that knowing how to fix those issues isn't important. It is. It's where experience and education are brought to bear and for that there is no shortcut. But it's no more essential to understanding the overall process than are the various skills and techniques I've developed over a career of 40 years to overcome building design issues.

 

[Noiz2]

...
I just make sure narration is louder than the rest, and that is fine, even for commercials. Now I spend maybe 30-60 minutes of tweaking audio for a 1 min commercial/TV-spot, etc. I feel I could spend a good 3-4 hours more to perfect it (well, perfect would require way more), but hey, 30-60 min and no complains. I tweak in Premiere, sometimes Audition and if there is really bad sound I will use other programes and plugin-ins to clean it, I sometimes have no clue what kind/brand of headphones I am using, because they are being switched around all the time at the office. Now, I know for sure I lack a lot of the amazing skills lots of you guys have that work with audio every single day, I prefer having someone else do audio, but it feels pretty good to know I can do it myself and get OK results. It means I can do one-man band type of shoots.

Well that is another factor that works it's way in. How high is the bar for what you are doing?

On somethings the bar is pretty low and that is the nature of the beast. Doing a full sound post on a local late night used car spot is probably a waste and they are not going to pay for it and probably won't notice.

Most of the ads I did were national spots. We only really did the FX work, the final mix was done at a separate facility. For FX alone we had a two to three man crew for a couple of weeks to a month for a 30 sec. spot. Now usually we were working on a package with cut downs so the shorter versions took a LOT less time and usually we could cover field recordings for the package in a more efficient way. But still you are looking at around 200 - 400 man hours VS your 1/2 to 1 man hour. Actually it's farther off than that since I'm not counting the mix or VO work (usually).

So who is the model to follow? Well as keeps coming up "it depends". We also did a huge stack of bumpers and lower thirds for a network rollout and did 83 in two weeks. They were all animated and most were very FX heavy.

If your job is to knock them out quick and dirty then the big post is not even an option.
If you are doing a feature or a short you want to get into serious festivals etc. then you NEED to spend some time polishing the sound.

OP:

I almost never use compression, I mix to get levels where I want them. I might in the end if it's for broadcast put some compression on the dialog to get it to punch through a bit. It's more likely though for VO.

Backgrounds are not Room Tone. RT is the "air" that you use to fill holes you cut to get rid of the odd noise, etc. On a big film there will be a totally filled dialog track but that is largely a waste of time and done so you can get "into and out of" the dialog smoothly. When it was all analog that was a critical thing, now that it's all nonlinear you can actually just use RT in the few places you actually need it.

Backgrounds (BG's) and Ambience are the sounds that make up the "world" of the film. Other than docs they are almost never recorded "on set", or even close to it.

And while we are at it, since nobody pointed it out, it's Foley not foley. Always capitalized. Someone, maybe your first post? , used it lower case.

You can't do a very good job of sound work in a NLE. It doesn't have the right tools and it works at too low a resolution.

Music should be the last thing you add to the sound mix. I understand that you may want it in for the picture edit but mute it for sound work till everything else sounds good. The music just masks all kinds of problems you WANT to know about. Fix the problems THEN add the music.

Noise reduction early, generally. But it depends on the project and the NR your doing.

 

[KINOKS]

+1 I normally level things out with volume envelopes and such. Depending on the audio, I resort to EQ and compression as needed on an event/clip basis, be it a multiband comp or broadband limiter. On dialog I normally add a high/low pass filters on the event, track or submaster. NR or other restoration type processing would be on individual events or a track... If severe, I'll export it to a dedicated audio application for 'surgery'. If using processing on the master buss, its usually a brick-wall limiter to catch peaks from exceeding a set value.. for the most part, it rarely kicks in at all.
If this not an answer your looking for, keep posting on many forums and you may get the answer you want... be it right or wrong.

Rick and recordingwiz that's great, thanks! It's easier for me to go, well, kinda backwards in the learning process. For instance once you know a subject really good you can point somebody new to the subject with some specific points in mind. The newb then knows the high points of the subject prior to reading/learning it, he then learns the subject much easier, with this points in mind. If you understand what I'm saying that is

Anyway you guys are saying that using a compressor isn't a must? I thought that this was like setting the contrast (black and white level) in the pictures. You know: so you don't have muddy blacks and whites and that you have a healthy amount of midtones. Is there such a thing to be done in post-audio? If you can relate to what I'm thinking...

Volume envelopes means that you are basically using key frames to normalize the tracks, right?

 

[KINOKS]

OP:

I almost never use compression, I mix to get levels where I want them. I might in the end if it's for broadcast put some compression on the dialog to get it to punch through a bit. It's more likely though for VO.

So you use compression to get the RMS of a track up, if it's already good then no need?

Backgrounds are not Room Tone. RT is the "air" that you use to fill holes you cut to get rid of the odd noise, etc. On a big film there will be a totally filled dialog track but that is largely a waste of time and done so you can get "into and out of" the dialog smoothly. When it was all analog that was a critical thing, now that it's all nonlinear you can actually just use RT in the few places you actually need it.

Backgrounds (BG's) and Ambience are the sounds that make up the "world" of the film. Other than docs they are almost never recorded "on set", or even close to it.

Yes, I kinda mixed this two up So basically you get a tone and you lay it under the voice. This is then an added tone to the original track right? But doesn't this just end up giving more unwanted noise to the track?

You can't do a very good job of sound work in a NLE. It doesn't have the right tools and it works at too low a resolution.

Oh, okey. By ressolution you mean 1 frame or 48kHz? ressolution

Music should be the last thing you add to the sound mix. I understand that you may want it in for the picture edit but mute it for sound work till everything else sounds good. The music just masks all kinds of problems you WANT to know about. Fix the problems THEN add the music.

Yes, off-course. It was putt on in the video-edit phase. The music is from a sound bank, it's not an original score. I really love this new sound bank from Universal. Anybody tried it? It's got like a gazillion of great tracks

Noise reduction early, generally. But it depends on the project and the NR your doing.

Ok. So if you remove the noise from the let's say talking heads, do you then still put the room tone in between the cuts? If the room tone is only comprised of noise. Or is this not a likely scenario?

 

[KINOKS]

It appears to me that the OP simply wanted to understand the postproduction process.

More or less, yes. This is a low budget, low profile doc, and it's lucky that it's sound will be treated at all beyond from levels adjustment and using crossfades - basically the stuff that the editor usually does when there is no sound finalisation done. So this is why I stepped in. Otherwise I wouldn't of have. I'm not expecting to do a great job, just an adequate, basic correction/prep job (or what ever you want to call it). And hopefully learn as much as I can in the process.

 

[j1clark@ucsd.edu]

Oh, okey. By ressolution you mean 1 frame or 48kHz? ressolution

For sound, recorded at 48kHz, the required time resolution would be 48kSamples per second.

Ok. So if you remove the noise from the let's say talking heads, do you then still put the room tone in between the cuts? If the room tone is only comprised of noise. Or is this not a likely scenario?

There may be a narrative reason for having 'total silence' in between bursts of sound... but for me most of the time, gaps are filled with 'something'. The most convenient 'sound' to fill the gaps are the roomtone of the location where the filming takes place, provided it is the 'low level' that one has in a 'quiet' room... if the place has various 'impulse' sounds, horns honking, or bursty, such as machinery starting up and grinding away, or continuous loud hum from various sources... unless those sound emitters are 'in the shot', they should be mitigated, if those are the 'natural' sounds of the location, (for filming that is using 'location' sound... rather than constructing all environmental sounds...), or eliminated if they aren't part of the narrative.

Example.
Documentary in a working machine shop... mitigate the grinder noise in the background, since the location is a working shop.
Shooting a narrative in an apartment building, and a plane about 400 feet off the deck is landing at the near by airport (where I lived during a portion of my Student Prince days...)... well time your shots so that sound is not captured... but if perchance it is captured... get rid of it... that's why god invented retakes anyway... even though sound such as the airplane or police sirens are part of the sound set for the location, they interfere with dialog and so would be added later and adjusted to not swamp out the dialog... unless of course that's part of the narrative...

 

[KINOKS]

For sound, recorded at 48kHz, the required time resolution would be 48kSamples per second.

Dddd, off-course! Thx

There may be a narrative reason for having 'total silence' in between bursts of sound... but for me most of the time, gaps are filled with 'something'. The most convenient 'sound' to fill the gaps are the roomtone of the location where the filming takes place, provided it is the 'low level' that one has in a 'quiet' room... if the place has various 'impulse' sounds, horns honking, or bursty, such as machinery starting up and grinding away, or continuous loud hum from various sources... unless those sound emitters are 'in the shot', they should be mitigated, if those are the 'natural' sounds of the location, (for filming that is using 'location' sound... rather than constructing all environmental sounds...), or eliminated if they aren't part of the narrative.

Example.
Documentary in a working machine shop... mitigate the grinder noise in the background, since the location is a working shop.
Shooting a narrative in an apartment building, and a plane about 400 feet off the deck is landing at the near by airport (where I lived during a portion of my Student Prince days...)... well time your shots so that sound is not captured... but if perchance it is captured... get rid of it... that's why god invented retakes anyway... even though sound such as the airplane or police sirens are part of the sound set for the location, they interfere with dialog and so would be added later and adjusted to not swamp out the dialog... unless of course that's part of the narrative...

Yes, I okey. I know this stuff, well, mostly I even watched Barrys DVD about sound. Thx! I call this diegetic and non diegetic sound.
If it has a place to be in the mix then yes, if not - then no! In one interview the crew was (except me) were sitting on chairs that kept making squecking noises through the interview. You know, when they moved a muscle the chairs produced a noise, and the sound guy, well, he did nothing This has to go out, any idea how to? There's probably lots of topic on the net about cleaning audio right?

But, regarding room tone, if there is music in the mix, you don't use it right? Only if it's low enough so that the cuts on the voice can be heard, only then you probably use room tone, otherwise there is no need, right?