Where should my audio aim to be when looking at the audio meters?

Question

I am working on editing a project and I have audio from some interviews and videos I did, as well as a backing track of music. I typically try to adjust the music lower by ear, but I guess the question is how low? I looked at my audio levels on the side and was wondering if there is an "unofficial" rule for where things should land in regards to the db scale? Where should the voice/main talking audio land? Where should the background music land/peak at?

Answer 1

A good way to do this on a Mac is to set the system out volume at 50% and listen (I use headphones, but you should be careful — very loud audio under headphones can be damaging to your hearing.) [A rule of thumb: Do not cause pain.]
If you're going by the audio levels indicators, you will generally find that audio that peaks between -6dB and 0dB (in FCPX) is "comfortably loud". Anything over 0dB is going to be flat out loud (okay for cinematic effects, noises, etc...).  Anything over +6dB will be clipped (distortion introduced -- will sound like a fuzz box guitar effect.)  Notice the audio waveforms in FCPX. If you see any yellow or orange/red, reduce the dB until they (just) disappear... then work from there. (A few short peaks into the yellow will not make that much difference if you think it's something to worry about — the only real thing to worry about is clipping — clipping is bad — clipping audio can get you fired )
Conversations should peak at around -12db to -10dB, but you will have to use some judgement for "effect", surroundings, etc...
You can increase the apparent loudness levels in audio via compression. Personally, I don't care for it — I much prefer a more natural "dynamic range" in audio, but sometimes compression is a necessity so it's good to learn to use some of the tools available (I particularly recommend the Logic effects in FCPX, they are typically far superior to the others.) If you need to boost conversation from background "noise", I recommend learning the Multipressor effect. You can use it kind of like an EQ except that you can compress audio within specific bands of frequency ranges (up to four). It's very cool
Actual volume should be left up to the end user. You're leaving them with the headroom between 50% volume (comfortable) to 100% volume which should be more than enough power to drive room speakers. (Let them be responsible for their own hearing loss.)