Hack59.Fix Common Audio Problems

Hack 59. Fix Common Audio Problems

Advice from the experts on how to fix common problems in recorded sounds.

I've rounded up some solutions to common audio problems. In audio you always have many ways to do a single thing. You can use these as a starting point on the way to finding your own solution.

8.11.1. Reduce Wind Noise

Wind noise creates a loud rumbling that is below the 100 Hz level in recordings. You can use a low-pass filter or EQ [Hack #57] to attenuate this effect. However, it's unlikely you will be able to remove the noise entirely. You should aim to reduce it to an unobtrusive background level and go from calling it noise to calling it ambience instead [Hack #64].

The ideal solution is to resample the sound with a windscreen and a filter on the microphone to take out wind noise.

8.11.2. Give Your Voice a Phat Sound

Starting with a good clean signal, you can add some depth by using an EQ to boost the mid and low ranges from about 1.5 kHz down. A slight reverb helps, though you will want to dial this in to make sure you aren't adding so much that the sound feels over-processed. Set the reverb to a small room and set the wet/dry mix very low [Hack #58]. You don't want to hear the delayed signal, you just want to double the existing sound to phatten it out a little.

If you have the chance to do a retake on the sound, have the person move much closer to the microphone. This will make use of the proximity effect, and the microphone will pick up more of the natural depth of the voice.

8.11.3. Obscure a Voice

If you want to preserve someone's anonymity, you can layer a series of effects [Hack #58] on their voice to give them a completely different sound without it becoming too distracting. I recommend first doing a pitch change to drop the person's voice by several whole notes. That will take her out of her normal pitch range.

People talk with a signature cadence. Think of William Shatner in Star Trek. His cadence was so peculiar that imitators don't even go after the voicethey just tweak their cadence. You need to obscure that. So, use a Change Tempo filter to speed up or slow down the pacing, and then take some of the clear recording at the beginning of the file or in between the words and randomly add or remove small segments from the spaces between the words.

8.11.4. Fix a Muffled Voice

If a person moves his mouth away from the central axis of the microphone, his voice will drop and become muffled. Try boosting the signal in ranges between 5 kHz and 7 kHz with an EQ to boost up the clarity of the voice. Then use a gain envelope to boost the signal to the level of the rest of the podcast.

8.11.5. Make Someone Sound Far Away

A person who is far away will sound faint, so reduce his levels. He won't be close enough to a mic to get the proximity effect, so you should cut the low end of the frequency range that would be there if he were up close. In addition, you can add some reverb to make him sound as if he is in the corner of the room.

8.11.6. Remove Hum

Low-frequency hum at around 50 Hz or 60 Hz is caused by ground loops coming from your power source. It's worth finding and removing the source of this signal [Hack #15]. To eliminate the noise from signals you already have, try a parametric EQ with a notch filter at 50 or 60 Hz.

Another option is a tunable noise filter [Hack #57]. Some of these filters, such as BIAS's SoundSoap, have filtering code to handle these power leakage problems specifically.

8.11.7. Remove Hiss

Hiss is high-frequency periodic noise. The high frequency gives it the annoying high pitch. Thankfully, this also moves it away from the voice spectrum so that you can remove it without too much distortion to the original signal. Use an EQ to remove frequencies above 7 kHz. Use less EQ in the closer frequencies and be more aggressive in the higher frequencies.

8.11.8. Simulate a Phone

When you have one side of an interview recorded through the phone and the other side on a clean studio microphone, the result can be jarring to the ear. It helps to take a little of the quality out of the studio microphone recording by dropping off the low end below 60 Hz and reducing the high end above 3.5 kHz. Don't kill them entirely because then you will sound like you are on a phone; instead, reduce them a little to lessen the jarring difference in quality between the two sounds.

Another technique is to use a coffee mug on its side to trap part of your voice [Hack #32].

8.11.9. Simulate a Radio

Growing up in the mid-'80s, one of my favorite songs was Wall of Voodoo's "Mexican Radio": "I wish I was in Tijuana, eating barbequed iguana." At several points in the song, the singer sounds like he is talking through an AM radio, severely boxed and distorted.

You can use digital effects to simulate this by doing a hard chop of the frequencies below 500 Hz and above 3 kHz with an EQ, and then over boosting the top end of the vocals around 3 kHz to add some distracting gritty clarity. Adding in another track of toned-down white noise will also add a little sonic grit.

If you have the time, another solution is to record through an actual radio. Put your recording on your iPod, and then use a Griffin iTrip to broadcast it on an FM band. Tune your radio close to that band. Analog tuners are better, since you can get them close but not quite there. Move the iPod around until you get some randomly slight cut-outs. Then record the signal coming out of the radio with your microphone. Using the internal microphone on your computer will add another level of grit and boxy compression.

8.11.10. Simulate Rewind and Fast-Forward

You can simulate rewind and fast-forward in two ways. The easiest is to use the Change Speed effect in Audacity to speed up the sound for fast-forward. Or reverse the signal using the Reverse effect and Change Speed to simulate rewind.

Another option is to record from iTunes as you rewind using the song position control. Use Audio Hijack Pro [Hack #50] to hijack and record the iTunes output, and then fast-forward and rewind to your heart's content. An advantage of this approach is that it sounds like a person honing in on a piece of content when the rewind slows as you approach the target.

8.11.11. The Voice of God

Apparently God has a deep voice, or sounds like George Burns. I'll take the easy route and go with the deep voice, which is just a change pitch effect. Additionally, drop your own voice and try to talk out of your chest and not your mouth. Your chest gives you a deeper pitch and a fuller sound. Slow down your cadence as well, or use change tempo to slow it down in post. But not too muchGod's voice is deep but not ponderously slow.

8.11.12. See Also

• "Choose the Right Audio Tools" [Hack #50]

• "Juice Your Sound" [Hack #51]

• "Maintain the Gain" [Hack #56]

• "Build a Sweet Sound" [Hack #57]