Rob Bee from B Double E holding a laptop with audio compression showing on the screen.

More essential technical audio production advice for voiceovers

11 months ago I wrote what I said was an occasional series. And I wasn’t joking because it’s taken me this long to get to part 2.

You can read part 1 here.

I also said, “I’m hoping you find this useful, not just in terms of how you run your day to day recordings, but also those times when you’re asked to something slightly unusual or interpreting a client request where a term has been misused or misheard (I still chuckle to myself about the client who asked for files to be supplied as forty four kilo hearts).” And I stand by that – apart from the anecdote about a client, as that memory popped up on Facebook this morning, and it was actually 8 kilo hearts.


Audio compression vs data compression

Which brings me nicely to my first tip. This is possibly a deciphering-what-the-client-wants kind of situation. What do you do if your client asks for ‘uncompressed audio’?

They could mean one of two things – audio compression or data compression.

So what’s the difference?

Audio compression is processing your waveform and squishing the peaks down so you get a tighter sound that’s easier to mix into a project. It’s something that befuddles a lot of people and something I’ll have to write about at greater length sometime in the future.

audio production

Data compression is taking your finished audio and saving it as an mp3 rather than a wav or aif for example. Data compression makes the file size you send your client smaller, but it very often does this by discarding audio data. Done properly you won’t hear the difference, but it can compromise the quality and those bits of data that are missing might be the all-important bits in your client’s project. There’s research for you to do here about lossy and lossless formats, unless you want to wait until I get round to writing about it.

There may be clues within your correspondence regarding which they mean, but you need to know what you’re looking for. Or be willing to risk a guess, and have the alternative saved on your computer so if they come back to you, you can send the right one.


Levels and noise floor

On to another couple of questions that are tied together inseparably.

  1. What levels should I record at? And
  2. How loud/quiet should my noise floor be?

Here’s the short answer.

You should try to record with your peak levels at around -6dB. If your peaks are around -6dB your noise floor should be no higher than -60dB.

Now for the longer answer.

Everything in sound is relative. There is no absolute silence like there is an absolute zero temperature. There are only things we can and can’t hear. We can give rough figures for the quietest sounds and loudest sounds our ears can cope with or the highest and lowest frequencies we can hear, but sound still occurs outside these parameters. All our measurements for sound are comparative levels. The way the human ear works isn’t exactly how sound works, so sometimes our measurements don’t tally with our auditory experience. This is definitely not a topic for now, but it’s something to very much bear in mind over the next bit!

The advice is to record with your peaks around -6dB. This is so it’s a good strong level, but you still give yourself enough headroom for the occasional transient peak to occur without clipping, and you run a minimum risk of your processing clipping. To be honest, this isn’t such a big deal as it’s easy enough for me as a producer to boost the levels at my end if I need to. However, what is important is that comparison between your voiceover recording and your noise floor. There will always be noise in your noise floor. Even if your recording environment is as close to silent as possible, your electrical equipment will make a small amount of noise. The really important thing is that the noise in your noise floor should not be audible beneath your voice.

If you’ve bought good gear and your room is treated well, your noise floor will consist of a low-frequency rumble, and as long as this is below -60dB (presuming your peaks are around -6dB) that’s not going to be noticeable – and can be further lowered with a High Pass Filter. However if your noise floor comprises mainly of hiss, at -60dB it’s going to be very evident, so you’d need to take measures to lower it. The rule basically should be that if you can hear your noise floor clearly, it’s too loud. The caveat on that is that turning your headphones down, or recording at lower levels doesn’t change the relative loudness between your voice and your noise floor, so in that situation even if you can’t hear the noise floor, it’s still too loud! Capiche?


Sort it at source

I know this type of thing has been a thing for ages, but I’ve seen a lot of it over the last few weeks. Someone posts on a forum something like, ‘the sound from the TV in the room next to my booth keeps getting on my recordings. Does anyone know a good piece of software that will get rid of the noise?’ I’m being slightly silly with the example, but I’ve seen quite a lot of that kind of thing. ‘I sneeze every time I say the letter F. Does anyone know of a good de-sneezer plugin I can buy?’ It seems that there is a belief that whatever problem you have in your audio production process you should just throw some money at the issue and cure it in post-production. This is not only not true, but definitely not best practice.

Whatever the issue is it’s always going to be best to solve it at the earliest possible point in the production chain. TV too loud? Turn it down. If that’s not possible, add extra soundproofing to your booth. If you can’t do that try to work around the noisy times, for example. Sneezing when you say F (obviously in reality this would be something like an issue with plosives, sibilance, nasal/mouth clicking, dysphonia or similar) figure out what the cause is and deal with it. Is it something in your mouth/throat? Go to a doctor or get some vocal coaching. If that’s not it, have a look at the way you address the mic or the way your room is set up and see if that can be tweaked.

This may well be the harder road to take in the short term, but you will end up with a more permanent solution, better quality audio than you will if you rely on noise reduction plugins and you’ll get through your work a lot quicker than if you need to edit lots of imperfections. The post-production cures for these issues that do exist should be a last resort.


So that’s it. Part 2 of the (very) occasional series.

If you have any questions about audio production or you’re after some 1-2-1 training then please get in touch. I’m always up for a bit of geek chat.