Podioracket.com would like to welcome J.D. Sawyer, who is contributing a two-part article on good audio that won’t turn off your listener.

Getting Good Audio
by J. Daniel Sawyer

Good audio can’t save bad writing. You’ve heard it before, and it reflexively sounds true – after all, great special effects can’t save a lame movie, and the same might well be true for audiobooks. The inverse, however, is a little less obvious to most people: Bad audio CAN destroy good writing.

This one doesn’t have a good film corrolary — after all, there are gobs of great movies made on shoestring budget, shot on video or 16mm, that are interesting enough that the low production values don’t make one bit of difference, aren’t there?

Actually, there aren’t. Unless you’re a bad movie fetishist (not that I’d know what that is), any movie that holds your attention is meeting a certain quality threshold with its production values. If you’re not seeing it on the screen, close your eyes and listen: chances are the movie has good, solid audio. Even if it’s not fancy, it’s got a few basic qualities:

1) The voices are clear, rather than muddy
2) The background noise levels in a given scene are consistent, rather than fluctuating up and down
3) The editing is clean – where there are breaths, they are full breaths rather than chopped in the middle.
4) The plosives don’t pop and the sibilants don’t hiss
5) The sound characteristics are consistent – you can’t tell where one take ends and the other begins, even with ADR.

How do you get these? Here are a few quick tips:

Clear voices

Voice is what you’re all about as a podcaster, so make sure that you’ve got a mic that plays fair with your voice. If you’ve got a bass heavy voice, don’t get a mic that attenuates the high end — you need those consonants in there to have a bright, clear sound.

To pull this off, the spectrum analyzer is your friend. Almost every audio editing program has one. Vocal range for most people is between 200 and 2000hz for the tone, and between 4000 and 20,000hz for your consonants. Record your voice doing different things and then look at it in the spectrum analyzer to see where your voice is the strongest and where it’s the weakest. Make sure that your consonant peaks don’t get lost when you process your audio — so don’t ever boost your bass without also making sure you can hear your consonants brightly and clearly.

Give me a clean sound

You must record in a dead room. In audiogeek speak, “dead” means “does not have much of an echo” where “live” means “has echoes like a gymnasium.” When you chose which room to record in, you want something that is, as Miracle Max would say, “mostly dead” (but not all dead – you need a teeny bit of resonance to help bring your vocal characteristics out.)

Most houses are pretty live – the flat, parallel, bare walls and the flat ceilings make for very unpleasant reverb – so you’re going to need to break things up. Throw pillows all over the place. Record with a bookshelf (filled with actual books) behind you. Cover any metal or glass or hardwood tables with reasonably thick blankets. Hang moving blankets or quilts in front of the walls and put throw-rugs down on hardwood floors. Put chairs at odd angles. The object of the game is to reduce the number of straight lines in the room and to introduce surfaces that absorb sound. If the flat surfaces can’t reflect sound back and forth to each other like a ping-pong ball then you’re on the right track.

Once that’s done, and you’ve recorded, you’re going to find out about another kind of floor: the noise floor.

Your noise floor is the noise that you can hear on the track when you’re not speaking – it’s the noise produced by the air moving in the room, fans in your computer, the electricity in the lines, the transistors in your equipment, and any radio frequency interference your mic cable might be picking up.

In professional environments, -70db or lower is an ideal noise floor and -60 or lower is acceptable (as most playback equipment produces -60db levels when it runs.) Anything over these levels is likely to be audible to your audience.

To measure the noise floor, use your spectrum analyzer — select a “silent” portion of audio and have the spectral analyzer graph it. On the x axis of the analyzer you’ll see the frequencies marked off, and the y axis will have the amplitude (volume). As long as your highest peak doesn’t break above -60, you should be fine.

Sometimes, it’s not possible to get a noise floor this low. If you’re recording multiple mics simultaneously, or you’re in a room that just can’t be made quiet. When this is the case, your audio requires special care when you edit. A number of programs out there contain noise removal tools called “gates” and “levelers.” Unless you know what you’re doing, do not use these – most of them will silence the quiet parts of your signal and/or boost your words (including the noise underneath them). This leads to sharply fluctuating levels in the noise floor, which triggers the fight-or-flight response in the human brain and will subtly unnerve your audience. On the other hand, the brain is perfectly capable of adjusting to even a fairly loud constant noise floor — put simply, it’s better to have a slightly dirty signal with a consistent sound than it is a signal that’s only clean in some places.

Stay tuned for Part 2.


J.D. Sawyer will be the host of a special Podioracket.com Blog Talk Radio Extension where he leads a discussion with guests Rhonda R. Carpenter and Heather E. Roulo. If you have questions for the ladies about Podioracket.com, BTR, or either of their books then be there LIVE at 6pm Pacific/9pm Eastern on Thursday, September 17th. Set a reminder!


J. D. Sawyer is author of The Antithesis Progression, Sculpting God,, Down From Ten, the host of The Polyschizmatic Reprobates Hour, and a regular contributor to LinuxJournal Magazine. Through his company ArtisticWhispers Productions, he manages various types of a/v productions and photography projects. Find out more, and get his podcasts, by clicking here.