The production sound you’ve captured makes up a surprisingly small percentage of the total film’s soundscape. The work sound designers and foley artists perform is hugely underrated and it’s often one of the things documentary or non-fictional films spend too little time considering. Here’s a brief look at the other elements present in good sound design.
Production Dialog (Sync check and cleanup being handled by the production dialog editor).
ADR (Dubbing over when production dialog doesn’t work. Or dubbing everything if you’re Bollywood.)
Foley (Artificially wonderful, usually massively exaggerated, effect sounds.)
Music (Timing, conform, and final sync being handled by the music editor).
Mix (Generally divided into two or three person teams based on dialog, music and effects).
First, per clip, use automation on dialog to visually “compress” it a bit. This can often be done in the editing workflow of any NLE. Then use a compressor or limiter to bring it up to your target and a workable level. Mix it in with your music and effects to taste more than anything really and then make sure the whole thing hits LUFS-wise where you need (-14 LUFS for web on average seems great).
Learn the importance of declick. It combats the high EQ shelf which, while it makes clarity really clear, brings out the mouth noise. De-essing is also important for a similar reason. Apply a bit after your EQ. Pull mid-lows out of music because they really compete with your voice. Ctrl+T gives you the aux track in the viewer so you can use automation. Resolve (and ProTools) are both cool in that they let you adjust clip gain automation before the track level automation.
Correct phase issues for symmetric waveforms (otherwise a compressor will reduce headroom falsely)
• Peak normalize all your dialog clips to get peak loudness in the same ballpark (-2 dbFS works well)
• Compress to bring the peaks down and use makeup gain to raise the overall loudness to hit your LUFS loudness target (or you could use the gain on a limiter for simplicity)
Busses are simply paths to AUX tracks.
Some Tips for Mixing
Mix dialog at 63dB
Resolve doesn’t sync picture and sound from separate folders, nor does it waveform sync reliably. I recommend timecode.
A cinema mixing environment has an “X Curve” which is a rolloff or attenuation of high frequencies above 2kHz. Plush seats and walls absorb most mids and highs.
You also can have nearly a second of reverb time (warm, bass reverb in a cinema).
Calibrate your speakers first to 85 dB. Now you hear the right “loudness” based on where you’re seated.
Alt+L for “listen again”
-12dbfs for dialog average (not peak which is represented by the horizontal line that floats and sticks above)
You’ve got to balance dialog on a per-line-level so that your master faders for the stems are just general automation and you’re not using them to micromanage dialog.
Right-click to normalize audio clip to -12dbfs
You’re choosing where you want your peak to be. So she sets peaks to -10db.
Don’t do this with a complex dialog scene since it looks at peaks.
Keep you dialog within a 5dB range.
-23 is target for loundess meters
Use range selection tool to mark clip I/O points
Peak level tells us nothing about loudness
RMS average tells us more
LUFS tells us most
LUFS is similar to RMS (same reading on pink noise)
Integrated Loudness is loudness average over time (more global loudness) (-12 LUFS is normal for an album)-16