Alexander Brandon asked,
Is there a compression process that you’ve found works best when normalizing whispered lines to be as audible as shouted lines?
Let me get something off of my chest: I don’t batch normalize anything. In my experience human ears are better to decide volume. Batch normalization tends to make all of the quiet stuff louder and leave the loud stuff quiet. It seems to cause more problems than it solves. So I work according to how things sound with my ears rather than batch processing.
It’s also worth mentioning that the quiet lines should be audible but still quiet. Varying intensity by performers can be important for creating a believable context and may even help advance the plot. We cheat real world volume in media so we can hear the story, but I always prefer to preserve some volume perspective. Weird volumes can be distracting.
Actors who perform in front of a live audience learn a skill: how to be heard from a whisper to a scream. The folks up in the cheap seats need to hear actors on stage. When those stage actors come into the recording studio, they know how to articulate and project a whisper. I know of nothing that works better than an actor who can perform a legitimate stage whisper. Seriously. I know that’s not what Alexander asked but a believable, projected whisper is worth mentioning because it is incredibly effective.
When stuck with a soft, mumbly whisper I tend to reach for EQ. If the actor worked close to the mic, thinning the low end will usually improve how it sounds, especially with whispers. The old Academy Curve rolled off the extreme lows and highs, then punched up the presence frequencies. That’s a good starting point, but it depends what else is going on during that dialog. After hearing the mastered dialog in game, it may beg for re-mastering adjustments to optimize that whisper in context. Sorry, I don’t have simple answers for this question.
Listen to all of the questions and answers… Dialog Editing for Game Audio.