Podcode.tv

How to sound like a million bucks

Now you’ve recorded your voice, we need to make a couple of tweaks so that it sounds its best, no matter wherever people listen. There might seem like a lot of information to take in below, but most of this is just background and context – the actual process of doing this takes less than a couple of minutes.

Normalisation

Once your audio is recorded, we need to normalise it. That means bringing the overall volume of the audio up so that the loudest part never exceeds -1dB. (0dB is the maximum, and anything over that causes distortion,, so -1dB is a good safe amount.)

If, after applying normalisation, your waveform got a lot fatter, that means your mic’s gain is turned down too low, and you need to boost it a bit. Try turning it from 12 to 3, re-record your piece, and apply normalisation again.

You want to see a bit of a change in your waveform, because you want to account for any sudden loud parts that might occur now or in future recordings.

Dynamic compression

Different from file compression – which makes large files smaller – audio compression or dynamic compression reduces the distance between the loudest parts of your audio and the quietest parts. When your audio track has been compressed, we’ll see fewer peaks in your waveform, and it will look flatter.

Radio stations add compression to their signal chain so that audio sounds punchier, and can be heard more easily over the sound of a car’s motor. We want the same effect with our podcast, so your voice can be heard equally well in a crowded train car or a smart speaker in a kitchen.

If you load an MP3 of one of your favourite songs from the last 30 years into Audacity, you’ll likely see what looks like a giant slab of blue. That’s because the audio has been compressed so that the quiet parts aren’t too quiet.

The finer points of compression

There are a few fundamental settings we need to know about, but essentially you can think of a compressor as a highly reactive recording engineer, keeping a constant finger on your microphone’s volume slider. When you’re quiet, the engineer turns the volume up; as you get louder, the engineer turns the volume down. We can fine-tune what this engineer does using these settings:

  1. Threshold: the point of loudness at which we want to start to compress, so anything louder than -12dB will be turned down.
  2. Noise floor: the threshold just under which your voice kicks in. Anything below that number is considered noise, not speech, so the compressor won’t activate for anything quieter than -40dB. If your environment is noisier, you might want to turn this up.
  3. Ratio: how heavily to apply the compression, or how much to reduce the loud part by. The higher the ratio, the more the loud parts are compressed.
  4. Attack time: how quickly you want the compressor to engage (or how quickly the engineer should turn down the volume on the loud parts). The attack time should be low
  5. Release time: how long to wait before setting the volume back to normal, after applying compression (or how long the engineer should leave their finger on the volume slider before sliding it back down).
  6. Make-up gain: whether to apply gain back to the finished audio to make up for the fact that the louder parts have now been made quieter.

Limiting

You might find that sudden sounds like Ts and Ks can pop through the compressor, because it wasn’t able to react in time, since the sound was shorter than the attack time setting. That’s normal, so we can counter that by applying a limiter.

A limiter sets a hard barrier to ensure that no sound will ever be louder than the maximum we set.

It’s best to leave all settings in their default state (if you ever need to, you can reset to the default setting via the Manage button), apart from one: the “Limit to (dB)” setting, which we want to set to -3. That’s because -3dB is a good maximum volume for most people’s headphones, and gets us close to the industry standard for loudness.

Normalisation and compression in Audacity

Homework

Complete this lesson