Choosing a balance between soundtrack and voice


I've recently purchased a nice microphone and recording setup. However, my environment isn't as quiet as I would like.

My goal is to make voice tutorials for software developers. I own a fair count of licenses to what I feel is high quality stock music. Am I right to think that putting in a faint background track would be a good way to cope with the noise floor while still providing clear, intelligible speech? Or, will I just make it harder to hear?

My question presumes that a soundtrack is a good idea.

How does one choose the balance between voice and soundtrack, in volume terms? Is there a standard rule of thumb? Is there a way to measure it? Do people just go by ear?

John Haugeland

Posted 2014-10-24T23:20:47.353

Reputation: 115

Be aware that balance in sound mixing isn't just about volume, it's as much about spectrum. Applying a BG track might involve 'cutting a hole' in its spectrum with EQ, that the vocal part can fill. – Jim Mack – 2014-10-25T00:13:11.430

That's interesting and I'm glad you told me that. However, the music in question is specifically meant for voice stuff to be put in, so I think the original authors have this covered for me. Out of curiosity, how would I do that? – John Haugeland – 2014-10-27T07:05:41.033

1The details could fill a book, but the idea is to use EQ to lower the midrange (say 1200-4000 Hz) in the music and boost it in the vocal so they don't compete. Google 'spectral mixing' for more. You can get more specific by using a spectrum analyzer (there are free ones) to identify the primary vocal range. Also, especially for non-pro vocal talent it may help to mildly compress (2:1 or less) to help punch through. Experiment, it's fun and instructive. – Jim Mack – 2014-10-27T10:20:58.007



Personally, I prefer tutorials without music. I also think a bit of background noise is not an awful thing. Gated dialogue or over-baked noise reduction is far more distracting. Choosing a mic you can get close to, thereby increasing the signal to noise ratio will help a lot. Also consider building a makeshift sound booth to block some of the noise out.

Mark Durham

Posted 2014-10-24T23:20:47.353

Reputation: 4 445

Instead of gating the signal I'd suggest using multi band compression with positive gain range, to keep the noise floor at a low level. It's a more careful method than gating. – csaudiodesign – 2014-10-25T16:16:43.750

I got a Zoom H6, and I'm considering getting a DPA 4080 to back it up. However, I'm in a tiny SF apartment, and building an anything is kind of out of the question right now. – John Haugeland – 2014-10-27T07:07:28.827

I wouldn't get a lav mic for that purpose. Try a large diaphragm cardoid condenser and get a reflection filter for it. SE do a good bundle with their X1: which is a bit of a bargain.

– Mark Durham – 2014-10-27T08:38:56.133


You could use a gate to try and remove as much of the background noise as possible. If the background noise is in the low frequencies, try low-cutting the commentary with an EQ or filter. Audacity also has some good tools but it depends on the timbre of the background noise as to whether they will work. As for the levels, you should have the soundtrack quite quiet as your commentary needs to be the main focus. It is best just to play it by ear. Try lowering the volume of the soundtrack until it is as quiet as possible while still being clearly audible. If background noise in the commentary is still audible, try raising the levels of the soundtrack slightly and use EQ to make sure your commentary cuts through clearly.


Posted 2014-10-24T23:20:47.353

Reputation: 237


There are specific Noise-Reduction plugins.

I hope this isn't too spammy, as I've not used anything other than these, but... I've used various of the Waves noise reduction & restoration plugs over the years to remove everything from gentle hum to significant outdoor traffic noise, with reasonable success.

Sometimes, it's still been necessary to put a gentle soundtrack underneath just to mask the bits I couldn't get rid of.
It's not a perfect process, if the background is loud & varying.

Sometimes, on a low-budget movie etc, they simply didn't have the resources to do ADR, so it's a case of fighting the rumble, bumps, knocks & squeaks, mic going off-axis right at the important bit of dialog etc.
In a more controlled environment, you might be able to get the noise-floor low enough to not need to add a soundtrack at all.

If you do add sountrack, bear in mind other comments about leaving enough of a 'hole' in the track that the voiceover can stay cleanly in front.


Posted 2014-10-24T23:20:47.353

Reputation: 5 764