Your audio journey – Part 45 – Processing voice

In this part of the audio journey, following on from my new bass roll off, I’m taking a look at how I process audio.

I use a multitrack template to produce DJ Drops, Radio Imaging, Commercials and more and on the voice bus I have a set of effects. Here I’m going to look at those effects in the waveform view and why I use them.

First step is to repair the audio. This step shouldn’t be necessary as you should have ensured you recorded good quality audio in the first place. In the event you do need to repair the audio be aware that not all audio is recoverable.

First in my chain is dereverb (switched off by default) as I have been sent audio in the past that has room tone in. Dereverb was much praised when it was first introduced as a native effect in Audition but I really don’t rate it. To remove the reverb (room tones) in audio dereverb degrades the quality of the audio to below a standard I’d call acceptable.

Next in my chain in dynamics and a simple auto gate to trap out background noise. If I needed to use this effect I would scan the noise using the amplitude statistics window and set the number accordingly. Bear in mind that when audio crosses the threshold the gate opens and all the audio is allowed through so if the background noise is overbearing it will also be heard in the audio allowed through the gate.

Deesser comes next to remove sibilence from the audio. This will need to be set to the individual voice you are working with as the combination of different voices and different microphones means this will be different each voice.

Then we’re on to the polish element.

FFT Filter performs the bass roll off (high pass filter). The low section of audio contains no useful parts of the human voice and often can pick up background noise such as air conditioning units, traffic passing by so cut this out the audio.

Parametric equaliser comes next and with the bass rolled off already you’re free to equalise the voice to make it sound great. There is no one size fits all setting here to set the EQ to each voice.

Next comes the single band compressor. My settings here are very light and barely touch the majority of audio. I would adjust the threshold and ratio per project depending on what I was producing.

The final active by default element of my processing is the hard limiter which sits just to ensure no audio exceeds the threshold. It’s set a -0.1db with zero input boost so will most probably not have any effect as mostly the audio will fall beneath this and not trigger it into action.

The final element (switched off by default) is mastering. I may activate this depending on what I’m producing to add some exciter and possibly use the loudness maximiser as well to ensure the audio is powerful enough to cut through sound effects and music.