Signal Processing:
A Brief Overview

Signal processing is where most of the fun is for folks who try to do their own mastering.

Things like the proper monitoring environment, or knowing what to listen for, often fall behind in second or third place.

Rather than discuss who produces which software however, here I'll outline a few of the actual basic techniques for manipulating audio.

And one of the most basic is volume adjustments – or levelling – or gain riding.

Yes sometimes the best solution is the simplest.

For example, sometimes within the flow of a CD, a song may start a bit too loud when compared to the ending of the previous one. Starting the song just a dB lower and then slowly easing in to its original volume can do magic.

There's times when adjustments need to be as much as 2 dB but unless drastic manual changes are required I would generally never go above that because it might start to sound like - “hey, someone's playing with the levels”(!)

Compression automates those level change processes – but in quite a different way than I just described. Used carefully it can warm up a track, add punch, or smooth things out.

Limiting is a more extreme form of compression. Digital brickwall limiters, especially, are used for:

• preventing digital overs to stop clipping and distortion on playback;
• to make a track louder.

Here is where the most care is needed: make your track too loud and people are gonna stop listening!

Volume changes sometimes need to be made just to particular frequencies in a track. This is what equalization does.

Proper equalization in a mastering context, preserves the artist's original intentions of her song's mix. The attempt is to make it more pleasing to listen to – repeatedly or long periods of time.

Analog emulations are another method of signal processing that attempt to make a track more pleasing. They can involve mimicking tape saturation or tube warmth.

There are also more specialized sonic enhancement processes that can bring out the higher frequencies in a more dynamic way, or 'deepen' the bass region.

This, of course, would be done in ways that would be difficult or impossible to achieve with the more standard procedures described above.

Imaging, or mid-side techniques, involve widening, narrowing, or sometimes correcting a stereo image.

This correction is different from what a simple left/right volume change would accomplish.

Like any of the other processes, mid-side procedures can be 'dangerous' if mis-used. Too narrow and you might end up with a mix that's essentially in mono. Too wide and you'll over-emphasize the out-of-phase elements of your mix, or bring out more reverb than was originally intended.

And when mid-side techniques are combined with any of the above procedures, you've just upped the power of signal processing to a whole new level!

So what processes should you use most?

That, as always, will depend on the individual song.

With me, first, I'll listen.

Then I'll listen again to find what - if anything – is needed.

I, first and foremost, and unless otherwise instructed, try to respect the artist's original intentions, as much as possible.

But if one wants to do this them self.....well that's fine and certainly when all one has left is a 2-track mix. Otherwise - I'd have to ask:

Why not just re-work your mix first until it's the way you like it?!

Return from Signal Processing to
About Audio Mastering Software