Post processing recorded dialog tracks - tips and tricks ???

C

c3hammer

Well-known member
I've been using a DPA 4061 lav and an Oktava 012 with a hyper cap into an H4n. I'm getting very boomy vocals and need suggestions on how to prevent it in the first place and how to post process the dynamics with compression, limiter, EQ and amplitude. I use Sound Forge Audio Studio through Vegas and am very familiar with Audacity.

Are there some low pass filters or EQ I can set in the H4n to tailer the tone and "roominess" with my mics?

What exact steps in the editor do folks usually use for each dialog track?

My work flow has been something like the following, but I've very unhappy with the results of any modifications I do other than to normalize the peak amplitude always make it sound worse to my ear.

1. Import wave file
2. Normalize to 0 db
3. Attempt to run some level of compression and limiter to even out the peaks (rarely helps)
4. Attempt to add some EQ to taste (rarely helps)
5. Re-normalize to -3 db for dialog
6. Save new wav file.
7. Drop new file on time line in Vegas and sync with the in cam audio

Suggestions on how to record, compress, limit and EQ dialog would be greatly appreciated.

Cheers,
Pete
 
How do the raw tracks sound? If it's boomy from the source, about the only thing you can do is band-aid it.

Was the boom kept within 18" of the talent?

Did the room have lots of hard surfaces, walls, floor, and ceiling and was this addressed with sound blankets? And I don't mean a couple, I mean 10 or 15 in close proximity to the talent?

Where you booming from below? Indoors, this often creates a monster boomy sound.

About the only thing you can do is eq a bit, and soft gate it.

Typically, normalizing and compressing are avoided during dialog mixing. Levels are automated (sometimes on a phrase by phrase or even word or part of word to bring the levels in check with the intent of the story). Only after the entire scene is checkered, leveled (via automation), and eq'ed (again using automation to only apply eq to the phrases and sylables that need them) can you begin to think about very mild compression, typically multi-band and with a very gentle hand, to polish.

What's different with dialog sound for film and music recording is that you can never make the dialog better than the original spoken word (and the spoken word is one of the most commonly heard sounds around the world so everybody knows what a natural voice sounds like). About the only think a dialog edit can hope to do is not screw it up and more than necessary.
 
Very hard to give a general reply to a specific problem ...

... but ...

... a cut at around 240Hz is a good starting point to reduce the "room effect" or "boominess" in post.

I also have a habit of rolling of at 80-100Hz when recording speech in an unforgiving room in the first place.
 
Thanks for the quick feedback gp. I now have quite a few things to work on before I ever get to post.

I did have the Oktava 14" out and below the jaw line aimed up at a 30 deg angle in my last test! All my other tests have been with the DPA clipped on the lapel 6" directly below the chin and I haven't used any sound blankets.

Any other suggestions for placement of the lav mic so it might be able to better mix and match with the boomed Oktava on another subject? Under the chin lav's just seem to have a radically different color than boomed mics.

Cheers,
Pete
 
Cutting between 80 and 240Hz seems to help a bunch! It's still crappy but much better. Thanks ceejay7777
 
c3hammer said:
Thanks for the quick feedback gp. I now have quite a few things to work on before I ever get to post.

I did have the Oktava 14" out and below the jaw line aimed up at a 30 deg angle in my last test! All my other tests have been with the DPA clipped on the lapel 6" directly below the chin and I haven't used any sound blankets.

Any other suggestions for placement of the lav mic so it might be able to better mix and match with the boomed Oktava on another subject? Under the chin lav's just seem to have a radically different color than boomed mics.

Cheers,
Pete
Click to expand...

Generally a boom should be above the camera line firing down aimed directly at the talent's throat with the mic axis at about a 30 to 45 degree angle off the vertical. Aiming up from below the camera line increases the chance of background noises intruding into the mic's pickup pattern and can also increase the pickup of room slap. Aiming down from above the camera line puts a lot of potential interference into the mic's pattern nulls.

Why should the mics on two different talents match each other? Their voices aren't the same are they? There's no reason a lav on one person should match the timbre of a boom covering a second person. A constant room tone under both can unify the scene when you cut them together. A boom and a lv on the SAME subject should intercut smoothly but that's very difficult to achieve. Generally a boom and lav are a choice of one or the other, whichever sounds better with the unused mic's track discarded.
 
Last edited:
I've found that this mic placement is of utmost importance :)

I had my mk012 on a boom pole laying on the table aimed up to simply do some tests. Big mistake! It radically changes the tone and feel of the shot vs. booming from overhead. Learn something new every day :)

@steve, I am trying to mix the recording of a lav on the cameraman with the boom on the subject out front. The timbre from the lav just doesn't fit the space the boom has and it's awkward sounding to my ear. I haven't played with room tones under both yet, so that may change my impression. Cutting down low as ceejay7777 suggested has helped immensely though.

Thanks for everyone's help. I'm getting there !!!

Cheers,
Pete
 
I've found the MK012 to be a boomy mic to begin with, so I never use it for dialogue (but others would disagree). Boom from above.

In terms of dealing with what you currently have. Don't normalize. You are just making the quiet noise louder. I never normalize dialogue unless it's for video games, and that's a whole different can of worms in terms of workflow. Import the dialogue. Roll off everything below 200Hz. Hope for the best. I don't like using a gate on the dialogue, as that tends to be really obvious when it kicks in. Crappy audio sticks out more when you hear it come in and out. It's better to just let it be crappy audio the whole time. Believe it or not, people won't notice it as much that way. Yeah, it's weird, but the brain is good at filtering out crap it doesn't want to hear...at least for people that aren't audiophiles.
 
This is just a random question, but do most of you change the dialogue that you recorded significantly in post? If it was recorded and sounds good, wouldn't the volume level be most of the "editing" you would do to it?
 
Yes I change it a lot by filtering out unwanted frequencies and adding sparkle and compression if required, if mixing for TV there is more compression but if it's for cinema then the dialogue tends to be more narrow band with more dynamic range.
 
movieman1005 said:
This is just a random question, but do most of you change the dialogue that you recorded significantly in post? If it was recorded and sounds good, wouldn't the volume level be most of the "editing" you would do to it?
Click to expand...
Most editing in general is cutting. Levels would mostly be "mixing" but those rolls are heavily intertwined these days.

As a rule I edit to clean out junk. I do full backgrounds so I don't spend a lot of time filling in with RT till I get BG's in and find problem areas. Admittedly a lot of the clean up is patching nearby "air" over pops and other weirdness so there generally is a fairly full RT, I don't just leave holes all over the place.

Once it's cleaned up I will pull out and deal with takes that need help fitting in. Sometimes with cutting between shots and takes the sound quality will shift and you need to EQ or? to make it all sound like it's the same take.

I will use automation to get the levels where I think they should be and I may use track EQ to make voices sound "better", clean out low end mud etc.

I generally put a mastering limiter on the aux track (that all the dialog tracks get sent through). This acts like a compressor on the loud parts with out boosting the quiet parts. I find it a much more natural sounding way to punch up the dialog if it needs it and you don't get "pumping". It also is a brick wall limiter so if I set it to -6 nothing is going to pop over that. That tames those few short hits you often get.

I will also have a master limiter on the master Aux channel (that the whole mix goes through) as a safety. Often projects are going to places whee there are peak specs and I can set the limiter to keep me safe and adjust the mix so that I don't hit the limiter often. Also it lets you compress the mix for video when they want it "louder" and you still of course need to respect the peak limits. You are effectively compressing only the loudest parts, which to me is a lot less objectionable than compressing the whole mix.

I generally put a low roll off on all dialog tracks. I try to keep that between 80 and 120. I find that much higher tends to affect a lot of voices. I will also often put a low roll off at 20 on the master Aux to kill any DC offset than may have happened with FX work or music. You can pitch shift stuff etc and get a bunch of subsonic info that you can't hear but eats up your bandwidth and will cause distortion. If I'm doing some big crazy low end s#%t in the FX I may slap that on the FX master also or even destructively process individual clips (regions). You really don't want the sub harmonic content unless you doing a thrill ride or a remake of "The Tingler" ;~)
 
That's a good idea using limiting instead of compression.
What do you do for really noisy situations where the background sounds are constantly changing between cuts? Like in a car. Other than ADR of course.
 
movieman1005 said:
This is just a random question, but do most of you change the dialogue that you recorded significantly in post? If it was recorded and sounds good, wouldn't the volume level be most of the "editing" you would do to it?
Click to expand...

This depends greatly on the project and the budget. If it's a low budget thing where I'm doing everything, then I try to clean up as much as I can during editorial. If I'm just editing the dialogue, and there will be a dialogue pre-dub and then a final mix, the clean up is left to the mixers. It's my job to give them dialogue that has no gaps in it.
 
Last edited:
Tim Joy said:
That's a good idea using limiting instead of compression.
What do you do for really noisy situations where the background sounds are constantly changing between cuts? Like in a car. Other than ADR of course.
Click to expand...

Spend way more time than I am getting paid for. ;~(

basically you do what you can to minimize and then when you add in a car interior BG you hope that most of the changes are masked by the BG track. You keep going between minimizing and masking until you get the best you can.

I'll also give the "don't record in a moving car" lecture. There is almost never a necessity to record in a car moving with it's wheels on the road.

If you have the bucks you can put it on a trailer and shoot it moving or with about ten minutes of shot planning you can shoot all your dialog takes with the car stationary and engine off and then create the "driving" in post with editing. You could go crazy and green screen it, but generally the CU on dialog don't have enough detail out the window that you can't "fake" the motion. Films with budgets NEVER shoot sound in a car that is moving on it's own power or on it's own wheels.

Actually there are probably a couple of exceptions , but don't mention it to the director.... because it's just a really bad idea.
 
The TV show Modern Family really looks like they shoot while the actors are actually driving. If they are faking it, they do a really good job of it. Their car interiors sound really good.
 
davefisk said:
The TV show Modern Family really looks like they shoot while the actors are actually driving. If they are faking it, they do a really good job of it. Their car interiors sound really good.
Click to expand...

THey are almost certainly doing a trailer shot if they aren't doing it in a studio. A tell will be where the shot is. If there is any shots from the front it pretty much has to be a trailer or a set. Thee is this cool "airport" around LA that is one huge set complete with sliced in half planes so you can light and get decent angles. I was at a talk by one of the editors on LOST where he was talking about how they were using motion tracking (of some real plane motion) to impart realistic bounce and shake to static set shots in the plane.

If you bet that every shot that looks like it's "real" was faked in any film with a budget you would come out way ahead. There is a great "extra" on the DVD of a small film called "Pipe Dream" (that I really like BTW) where they show the various takes of the last shot in the film. It's a simple shot of the two leads walking out the door of the studio and walking away from the camera down the street. Sounds great looks great.
Then in the extra you realize this shot was under a freeway and they are getting no production for most of it and what they get is just hammered with traffic. But also that the door is just propped against a solid wall. They are waiting behind that and walk out and it looks perfect, but it was a total fake.

There is also a great extra on the John Adams mini series that shows how little set was actually built. About 90% of the "sets" were matte work.
Then there is the doc on the making of Gone with the Wind where you find out that a LOT of that was matte work also as in all of Tara is a painting except the front doors, which are actually the sound stage doors. None of the rooms had ceilings, they were all painted in, in post. Most of the big epic vistas are Matte paintings, etc.
 
You must log in or register to reply here.
Back
Top