Using A Frequency Analyser to Process A Voice Over
One of the hardest things about processing a Voice Over can be figuring out exactly what’s going on, and what to do next with your EQ, compression and other settings. That’s where using a Frequency Analyser can come in handy. The trick? You want to have a little ‘hump’ rather than a wild looking jagged line… not clear? Read on!
Grab your favourite Frequency Analyser!
I like to use the PAZ Analyser from Waves, but use whatever you like – if you’re using Pro Tools put an instance of this plugin in the very last insert space. Then before you apply any other plugin settings, play the audio back and watch the levels in the Frequency Analyser. Once playback is finished, you should end up with something like this:
This shows the average frequency response across the entire piece of audio I played back and it’s a pretty ragged line. It tells me a lot about the audio file and what frequencies I need to consider when creating my voice chain. It’s important to remember that no information is better than listening to the audio itself, but by seeing the frequency response I can make more informed judgements about what I might want to do to process the audio.
So what can we assume from looking at this Frequency Analyser?
Firstly, lets start with the low end. We can see a lot of response from 0Hz to just beyond 62Hz. Human’s can’t hear a lot of this frequency range (most people can’t hear below 40Hz they can only feel it). We’ll want to use a High Pass Filter (HPF) to tame this area. After that, we can see a little bump around the 125Hz area. It’s likely that we will want to dial our HPF up to around this area, just to avoid any additional bass clash when we add in music and other sounds underneath our Voice Over. Moving forward we can see another hump in the 500Hz region that we’ll need to tame back in order to help add some clarity to the voice. After that things look relatively smooth through the mid range until we high the high range. Around 8000Hz we can see quite a spike, this is likely to be sibilance that we’ll need to tame either with EQ, a DeEsser or both. After this spike the frequency tapers off nicely towards 16,000Hz.
So with this information we can now start looking at an EQ to try and make some of the adjustments as mentioned. Here’s what the EQ looks like:
You can see that I’ve engaged a HPF around the 125Hz region, a large cut around the 400Hz area and some little notches to tame other wild frequencies including a large cut around the 8000Hz mark to help tame the sibilance. With those settings, here’s what my Frequency Analyser response looks like:
Looking at that, you can see we’ve managed to tame a lot of the problems and we’re starting to see a little bit of that ‘hump’ I mentioned earlier. But we can see that there is still a little spike around the 8000Hz region. It’s nowhere near as bad as before, but still there. Listening to the audio at this point I can also hear a little too much sibilance still, so I’ll employ a DeEsser to sort it out. Here’s how I’ve set that up:
It’s settings have really helped to tame the sibilance, check out how the Frequency Analyser looks now:
And there we go, it’s tamed and it’s looking like a pretty decent ‘hump’ right!?
With my EQ sorted, and the voice sounding pretty good I can now get busy with adding Compression, Limiting and other processing I like to taste. As I add this additional processing I will continue to listen carefully and check my Frequency Analyser to make sure it still looks nice and ‘hump’ like!
Give this a shot next time you’re working on processing a Voice Over, I’d love to hear how you get on!
Credit to George Taylor of IMGR who is the original source for this tip.