Recording on the AudioShocker

Noise Removal. This is the most troublesome task I encouter every week while editing the audio for our podcasts.

Obviously, the easiest and most efficient way to decrease noise (static, hissing, movements, and shuffling sounds) is to record your audio with as little noise as possible in the background. Regular podcast listeners often listen to me harass Neal during the show when I can hear static on his end of the audio. That's because I don't like taking it out later!!!

Here are my current recording settings:

Recording Device: iMac G4 built-in microphone (in monitor)
Recording Level: the low side of 0.4 (on the Audacity mic input slider)

It's inevitable that noise will end up on your recording. Otherwise, you'd have to be sitting in a sound booth with a high-end microphone and some extremely still participants. Basically, what I'm saying is that you WILL have noise, so you better learn how to get rid of it.

How do you eliminate noise? Even though I have my ProTools running again, I still record and edit my podcast audio in Audacity. It's less of a strain on my iMac G4 (which is a bit outdated and slow). Audacity is great because it's a lighter program and the editing tools are simple. But it's also maddening because the built-in effects are different for Mac and Windows. The best of these filters are, unfortunately, incompatible with the opposite operating system.

Luckily, Audacity on the Mac has the best filter for noise removal. It's listed as Apple: AUDynamicsProcessor and it saves my life week after week.

Here are my AUDynamicsProcessor filter settings:

Compression Threshold: 20 db
Head Room: 40 db
Expansion Ratio: 1.89
Expansion Threshold: -82 db
Attack Time: 0.005 Secs
Release Time: 0.510 Secs
Master Gain: 0.0 db

AUDynamicsProcessor effect in Audacity

The "Expansion Threshold" setting is the most important of all. For quiter audio, you want this to be in the low 80s, high 70s. For more robust audio (and great amounts of static), you'll want to toggle this somewhere in the low 70s or 60s.

You know that you've just got a bad signal when you have to start messing around in the 50s to knock off the static. At that point, I would either rerecord or just cut your losses and leave in some of the noise.

These settings may not work for you, but they've become ingrained in my brain after a year and a half of weekly podcast editing (not to mention countless hours of editing vocals for music as well). If you're running Audacity on a PC... I'm sorry. There simply isn't a tool like this that comes default with the program. You'll have to do trial and error with 3rd party plugins. The filter you want is most easily defined as a "gate," but unfortunately, not all gates are created equal.

Once you've taken care of the noise, then you have to move onto volume levels (a.k.a. compression). But that's a whole other bag, so it'll have to wait for the next edition of Podcast Pointers!

I'd like to discuss a trend that has become an increasing annoyance: the 'delay double' (yes, I just coined that term). We've talked about doubles on the podcast before. Basically, a double is when you record a line a second time, slightly different and layer it back over the original. Puffy explained this really well in an episode of Making the Band. Pretty much everyone does doubles. They are easy to do and give tracks a little oomph.

Proper doubles are subtle and they don't call attention to themselves. Listen to 99 Problems by JayZ. You barely notice it, but Jay doubles during the chorus. It adds a whole new acoustic dimension. This is what Puffy was getting at.

A 'delay double' is when you repeat a certain line or excerpt again for emphasis - but it's delayed a second or two. For example, you may hear something like 'Shorty be the hottest thing I ever seen,' followed quickly by 'I ever seen'. Often it will be whispered or hushed and extended.

Delay doubles are used extensively in hip hop and pop music. The reasoning is simple: it's easy and it simulates a listener repeating his favorite part of a lyric. If you hear a particularly good punchline, the kind that makes you go 'ooooo snap!' and hip hop heads live for, you may be tempted to repeat the last two or three words. That is what the delay double is after. Producers are trying to encode this reaction into the track up front.

A prime example of offensive delay doubling is Bustit Baby by Plies. Take a listen. He does it on every. single. line. First, there are no punchlines worth repeating. Second, even if there were - repeating the last two words of each line is just plain annoying; your listeners are not deaf! Someone please clue me in - how does this sell records? Is this what's hot in the street? Is this what you'd call really hood? Sadly, this is not an isolated incident.

Usher and Jeezy are both guilty of misdemeanor delay doubling on Love in This Club. I think I could cosign the delay if it were used a little more sparingly, but things as they are, I want to slap every artist employing it. Other offenders include: Destiny's Child (Soldier), Mariah Carey (Touch My Body), The Dream's (I Luv Ur Girl) and heaps and heaps of others. Even Mos Def must be brought to task (Ghetto Rock)!

I'm sorry to say it, but delay doubles may one day unseat the vocoder as the worst thing to ever happen to music.