The Dialog Chain: what, why, how? (Part 1)

The dialog chain refers to the series of plugins and processing through which all of your dialog travels on its way out to your master mix.  This is the place where each mixer creates their own special flavor, and is the source of greatest differentiation between different dialog mixers.  A certain amount of it is taste, another amount definitely hocus-pocus.  But there are key things that can be applied globally to your dialog via your chain that really improve things.  Here are a few of my recommendations for your dialog chain in Pro Tools.

First off, you ideally have your dialog cut across 5 or so audio tracks (labeled A thru E, 1 thru 5, or however you choose to identify them).  Since horsepower is easy to come by with today’s computers, I recommend first placing a 7-Band EQ III followed by a compressor on each individual dialog track (I personally like the Waves Renaissance Compressor).  I personally recommend EQ before compression because the EQ can be used to first remove unwanted frequencies (such as low rumble) that would unnecessarily trigger your compressor.  If you started with a compressor it would be triggered by any frequency that exceeds the threshold, even if those frequencies would just be removed by EQ anyway.

Set the individual compressors to a relatively mild setting, just enough to even things out a little bit on each track.  This makes mixing easier, because your fader moves won’t need to be as dramatic.  Next, each dialog track needs to be bussed to an aux track which will function as your dialog master.  Your “dialog chain” will be the series of plugins that you insert on this dialog master aux track.

So what’s in the dialog chain?  Typically it will have an EQ, then some sort of noise attenuation, another compressor, possibly yet another EQ, and then a limiter.  Let’s review each, their placement, and why they’re used.

First set the EQ at the top of the dialog chain to a basic setting that helps the typical issues of rumble and hiss.  That is to say, roll off the lows below 100Hz or so, and the highs gently starting at about 10-12kHz.  Also some slight attenuation around 400Hz and a slight boost around 4kHz  is generally a good idea (gets rid of the honky timbre in the mids and boosts the clarity of consonants) .  Adjust all of this to taste as needed, depending on the requirements of the project.

Second, instantiate a noise reduction tool of some sort.  My primary recommendations are the Cedar DNS (either One or 2000), or Waves WNS.  Both of these are essentially advanced multiband expanders whose function is to push down the unwanted noise throughout the frequency spectrum.  If the budget doesn’t exist for these options (they are indeed mighty expensive!), you can simply use a multiband expander such as the Wavearts MultiDynamics 5 or the Waves C4.

Why a multiband expander?  First, let’s cover why you don’t want to use a regular expander.  Changes in noise draw attention to themselves (like when an air conditioner kicks on).  Constant noise is eventually less noticeable (like freeway noise).  A simple expander turns down the noise in between words, but every time someone starts speaking the noise shoots right back up.  This constant change in noise level is actually worse than steady noise.

MULTI-band expanders, on the other hand, have a distinct advantage; they split the frequency spectrum into 4 or more bands and run an expander on each independently.  Because each band pushes down independently of the rest, the noise never globally pumps in and out.  This still results in the benefit of pushing down some of the noise, while allowing the dialog to come up without swelling all the noise with it.  Functionally you are able to significantly reduce the noise without drawing unwanted attention to the processing.

In part 2 we’ll cover the remaining plugins of the dialog chain!

4 Comments

  1. Hi there, Why is the noise reduction in the aux master? Dialogue 1 may have a different kind or amount of noise than dialogue 3.

    Thanks

  2. Hi Tony,

    you make a good point. In the case of scripted content, provided that the production sound was recorded reasonably well and the dialog edit was done correctly (with proper fill), noise should be relatively consistent within a given scene. Therefore noise reduction is typically a concern on a scene-by-scene basis, not region-by-region.

    Additionally, anti-noise plugins are typically quite resource-heavy, and it’s in your interest to only instantiate the minimum necessary. Also the DNS 2000, a hardware option that’s quite popular, is only a 2-channel device.

    That said, systems are only getting more powerful, and options such as the DNS One are supposedly quite efficient. Still, if you feel compelled to scrub noise on a clip-by-clip basis, you probably have bigger overall issues with the sound.

    Lastly, noise isn’t entirely a bad thing. You don’t want to completely suck all the air out of your dialog, or you’ll end up with unnaturally sterilized sound. Therefore, a single noise-reduction processor that’s processing judiciously on all the dialog is generally sufficient.

    All that said, I have seen some mixers insert a Waves C4 on each individual dialog track in order to attack the noise more aggressively. While not my style, I can certainly see the point! Thanks for posting!

  3. Thanks for a great 3 part article. Probably the most useful thing I’ve ever read about mixing dialog.

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.