In the last part we discussed the significance of the main dialog compressor on the dialog chain.  Now let’s cover the last items in the chain: another layer of EQ  and/or a De-Esser, and a limiter.

Why another EQ?  We already have one on each dialog track, and we already have one on the dialog chain.  The reason we’re adding this final EQ is because the anti-noising and compression that we’ve inserted before it have a significant potential to change the timbre of our overall dialog mix in ways that we may want to adjust.

Remember, compressors act by attenuating the signal when it rises above a given threshold.  Of course, at any given time probably only certain frequencies are rising above that threshold, not the entire sonic spectrum.  When we compress peaks that live in a certain frequency range, we’re not only adjusting the level of our overall signal but also its timbrel characteristics. We’re attenuating the frequencies in the spikey frequency range, shifting the overall frequency balance of our signal.  Often this is fine and isn’t noticeable, but sometimes is draws unwanted attention to other frequencies because now they’re louder relative to the overall signal than they were before compression.

In other words, compression can add harshness to our dialog that we may want to use EQ to remove. Also, another consequence of compression is potentially a raised noise floor.  We can use EQ to correct for any excessive rumble or honkiness that becomes a problem as a result of the compression.

Another useful tool here is a de-esser.  A de-esser is a compressor that only acts on the higher frequencies, with the objective of reducing the sibilant harshness in dialog.  Since the compression has the potential to make the dialog more harsh and sibilant, a de-esser can come to the rescue.  I often use the built-in Dynamics 3 De-Esser set to High-Frequency only, slope starting at about 6kHz, and attenuation between 12 and 16dB.  Another good choice that adds the benefit of an adjustable threshold is the Waves Renaissance DeEsser.  Either way you go, you will have added a useful component to your overall dialog path.

Lastly you’ll want to place a limiter on your dialog chain.  Once again I turn to the Waves family, and recommend the L2 limiter.  Depending on the type of content and ultimate delivery format, the setting you choose could vary significantly.  For 5.1 mixes that don’t have a hard-and-fast peak limiting spec, a setting around -2dBFS is probably adequate (just to keep things from clipping).  However, as of the 2010-2011 TV season, networks are frequently requiring that peaks for 5.1 mixes be limited to -6dBFS.  Therefore you’ll want to set your limiter to at least -6dB, and may want to set it even lower.  Why lower?  Due to the rules of uncorrelated constructive interference, your dialog will sum with your music and SFX and result in a higher dB level than any one of those components by itself.  This means that your dialog will probably dig into another limiter after the dialog chain, even if the chain has it set to -6dB (the network spec).  This comp mix limiter will be applied either in layback or broadcast (or both).  These are two places over which you have no control.

In order to minimize the damage done by the overall mix limiter, you can limit more aggressively on your dialog chain in the first place, and mix with that limiter in mind.  That is to say, if you place a -8dB limiter on your dialog chain and mix in such a way that you don’t dig too hard into it, you know that your dialog peaks shouldn’t be over-crushed by the broadcast limiter.

That’s it for this discussion of the dialog chain!