Blog Relations

The biggest sin for podcasters, or any recorder of audio, is to allow “clipping”. This happens when your sound wave peaks higher than your microphone is able to record. The top part of the wave is lost for good, because it was never recorded in the first place. The resulting distortion sounds like a speaker overloading, even when the volume on playback is turned down.

It’s very easy to commit the clipping sin, especially as most people like to have a nice loud recording. The other problem is that people tend to talk quietly and then more loudly as they start to “perform”. Singers also like to really put their all into the emotional bits.

If you are recording several voices, then they will almost certainly talk/sing/shout at different levels. The first rule to avoid clipping is to wear a good pair of headphones. Do a sound test. Get everyone to say what they had for breakfast. Obviously if it is music or a performance of some sort, get them to do a few lines. Try to encourage them to speak at their natural level. Listen out for any distortion during the test.. And of course, keep listening while you are recording.

It also helps if your recorder or recording software has a meter. Watch it like a hawk to see if it is going into the red, or above zero. If you are recording several people in a studio, or what passes for your studio (podcaster’s garden shed), then it is a great benefit to have a sound mixer and each microphone plugged into a different input with its own volume control. I prefer a mixer with sliding faders, because it’s much easier to pull them down or shove them up as the volume changes. Even so, I find it really difficult to “ride” the recording.

When you do your sound test, leave some “headroom” – don’t record at the maximum level available to you. There should be some space for the soundwave to peak before it clips.

The next problem comes when you encode your recording into an MP3. This involves “compression”, and the higher and lower registers are pushed towards the middle. Quite often, even nicely recorded sound gets clipped in the process. The answer is to “normalise” to a peak level, which is jargon for a post-production technique that reduces your highest sound levels. If you normalise to -1db there should be some headroom even over your highest peaks in sound. Obviously do it before converting to the MP3 format. This leaves some room for compressing into the MP3.

Most good sound recording programs, such as Sony Sound Forge or Adobe Audition have normalisation tools.