Dialogue Plugin Chain flexibilty and Bussing Preferences?

Question

Sup peps,

One thing that I struggle with every time I get ready to mix a new project is choosing the structure and order of the dialogue processing chain.

Sticking with your choice...

When we mixing for post production we enter the realm of heavy automation. It is therefore important to decide on a processing chain that will be able to address all of your dialogue needs throughout the duration of the mixing and this is a difficult thing. In doing so we realize that we now don't have the freedom to switch our signal processing choice in the middle of a project without it causing complications (correct me if im wrong)... and even if we choose to do so, we would have mirror our processing on all of our dial tracks so as to retain the ability to make use of all of our dialogue tracks should we have to move our audio around... Right?

At present, my default chain is like this for the reasons listed below.

ALL AUDIO TRACKS (usually between 2-8) pre bus

    v

EQ III: This Eq is for corrections. Filters, notches and basic tonality adjustments for if its boomy, thin, nasally or whatever...

    v

Noise Suppression: Mostly waves WNS or their w43... my choice depends on overall quality of the sound from the get go. Even before my first EQ the sound has probably already been exposed to some light noise suppression though destructive processing so I normally don´t need too much of this.

    v

Channel Strip: I have gone with the Channel strip that comes with PT10. I use this for some basic dynamic taming and additional EQ. Expansion for lowing backgrounds a little and some light compression just to tuck in the peaks when necessary. With this EQ i use to get the tone I want. I also like this channel strip as it gives you the option of choosing (and automating) the order of the different processing elements (FILTER-EQ-DYNAMIC or FILTER-DYNAMIC-EQ ext...) in channel strip itself which I find really useful!

    v

De esser: When I have the time, I normally prefer to volume out my esses but when time is of the essence (no pun intended) this is faster.

    v

Waves C4: I find that from time to time, I need to remove some of these "in your face" mids that can become unconformable to listen too. Its also good for controlling boomy dialogue. I use the C4 as a multi-band compressor to accomplish this.

    v

Waves Ren-compressor: This is my go to compressor for finding the correct dynamic balance in the dialogue I am working on. Smooth and transparent and I find that this has actually been known to solve tonally difficult dialogue at times.. don't know how... just has! :P

   v

DIAL BUS: After my audio track has gone though all of the processing above, it goes straight though to this dialogue bus.

   v

L2: This is the only processing I have on the bus. It normally keeps the same settings throughout so things don't get out of hand but its no stranger to automation either.

Thoughts, suggestions and comparisons would be most welcome!

Answer 1

That sounds similar to what I use, although I don't use an of the RTAS noise suppressors (except for the C4, which I have saved a noise suppression template for which I like to use). To my ears the WNS sounded nearly identical to the C4 but with a whole lot less parameters to control.

I always run Q10s an Trims on each DX track, but all other processing is done on my sidechain bus, so the processing is happening to a summed signal. I find that the EQIII isn't tight enough to get in there for particularly nasty notch filtering. At the beginning of my summed sidechain bus I always do a hard rolloff at 80Hz - nothing down there except sonic material which could interfere with compressors or noise suppressors, and it prevents the risk of any unexpected mic bumps from hitting the LFE. From there I run my plugin chain and after that it gets bussed out to my DX Stem Submaster - the only thing I put on it is Maxim as a preventative brickball (sound similar to your usage of the L2). Then the output of that is sent 1:1 to the DX Stem, followed by 1:1 to the printmaster.

My usual rule of thumb has been to do frequency-based and gain-based (corrective) processing at the discrete track level before the sidechain, and do dynamics-based and noise-based processing on the summed sidechain itself.

This is just my 2 cents. Your chain sounds similar to what I and likely others use (assuming the chain is processing a summed signal versus track-based).

Answer 2

Hey Coca,

That chain sounds pretty typical for me as well, although I often use the high band of the C4 instead of a dedicated de-esser.

Depending on the project, I often create a separate set of editing tracks every time the scene/location changes. Supposing that I'm working on REEL 2 SCENE 3, a set of tracks looks like this:

• 2.3 Dialog A
• 2.3 Dialog B
• 2.3 Roomtone
• 2.3 Fix
• 2.3 ADR

all routed into

• 2.3 PROC (mono aux track)

These edit tracks contain my corrective EQ, and, like Stavro, the summing aux track contains my dynamics plugins (usually a C4 for multiband expansion and a C4 for multiband compression) and any aesthetic EQ. If I'm mixing for a small-speaker medium such as an internet video, I'll add another compressor (I, too, love the Waves RComp).

Even if multiple scenes take place in the same location, creating a dedicated editing chain for each lets you customize for the scene as needed, and makes it simple to return to the edits if you need to make a corrective change later, without worrying about affecting the sound quality of another scene. Once you get the settings you like, you can send the output from the summing aux track to your DX stem, bounce out your edit, and deactivate/hide the set of edit tracks.

To reinforce Stavro's rule of thumb, it's also best to write your volume automation on the individual track level, before the sounds hits your dynamics processors on the summing aux. This will allow you to smooth out the source audio and will keep your compression/expansion more even and predictable.

Cheers,
~Matt