Project Mastering Master Class

So for this roundup, well assume you have the basics figured out (you can master at home, but you better know what you're doing), and concentrate on specific techniques that relate to mastering.

Some of these relate to tracking and mixing, too. Also, rather than having our usual review format, well instead pick some cool features from various digital audio editing programs, and show how to apply them to real-world situations.

As to which digital audio editor is best, they all do the job but they all do the job differently. Also, some have unique features that are essential to some people, but irrelevant to others. I'm very fortunate, because I get to evaluate all these programs while doing reviews, then use whatever I want with music projects. And frankly, I use everything. I'll often go through three or more programs to get the final result even crossing back and forth between Windows and Mac.

Of course, having multiple products adds up price-wise, but the price of all these programs adds up to about the same as the reel-to-reel two-track machine I used back in the day. We've definitely come a long way.

SO WHATS THE DEAL WITH GOLDEN EARS?

But first . . . for years, people have talked about the need to use professional mastering engineers, with the usual reasons being well, they've done hit records and have golden ears. But what are the characteristics of golden ears for mastering?

Simple: The ability to detect extremely subtle changes. This is crucial for two reasons. First, applying a processor to a mixed stereo track affects everything if you boost a particular frequency, you're boosting that frequency for drums, voice, bass, etc. This is very different from processing an individual track, where its often desirable to paint in broad strokes.

Second, mastering typically involves lots of little edits, but these add up to a not-so-little result. Every action does have an equal and opposite reaction; alter the dynamics, and you alter the mix. Boost or cut at a certain frequency, and it will make other frequencies seem softer or louder in comparison.

This is where many wannabe mastering engineers fall short, because they apply recording thinking to mastering thinking. Mastering is the art of subtlety, and you have to understand which small changes you need to make for a big result.

THE EQUALIZATION TWO-STEP

If I could only have one processor for mastering, it would be EQ. I actually use two independent EQ processes. The first fixes problems, while the second adds subjective tonal improvements. For fixing, I call up the file in Har-Bal to see whats going on in the overall audio spectrum. The heart of the program is an 8,192 stage FIR equalizer, but it also displays an average of the energy distribution across the audio spectrum in 1/6 octave bands (you can change this, but 1/6 octave is my preferred setting). Looking at the display can provide an early warning system for any frequency response anomalies, although of course you cant make any final determinations without using your ears (and brain).
Common problems are:

Bass doesn't roll off at subsonic frequencies. Cutting everything below 2030Hz can clean up the sound and open up a bit more headroom (also see the section Remove the Subsonics).

Bass range peaks and dips. This is usually due to room issues where the recording was made, but be careful in your analysis there may be a major kick drum that causes an intended peak. However, this tends to be a single blob of energy, whereas room issues cause a curve that looks more like ripples due to multiple resonances.

Too many highs. What with distorted guitars, aliasing that generates weird harmonics, digital clipping, and the like, today's recordings sometimes seem harsh. A little high frequency roll off can tame harshness without reducing the perceived high frequency response.

Midrange issues. Unexpected midrange peaks, attributable to a variety of factors, can sometimes give a honking effect. These may be subtle, but you'll still notice the sound is smoother when you correct them.