Speakers Meet Mastering

You know why I don’t review speakers? Because I don’t really like ’em. Even the ones that cost about the same as the budget of a small third world country don’t hold a candle to the Real Thing.
Publish date:
Social count:
You know why I don’t review speakers? Because I don’t really like ’em. Even the ones that cost about the same as the budget of a small third world country don’t hold a candle to the Real Thing.

Don’t believe me? Then try this: Have a really good acoustic guitarist play in a room. Then play back a recording of the same guitarist, playing the same guitar. Doesn’t sound the same, does it?

Granted, not all the blame falls on the speaker: The rest falls on the microphone. And I don’t really like microphones, either, because they too are transducers that have the unenviable and extremely difficult job of transforming moving air into electrical waves.

What is amazing about speakers (and mics, for that matter) is that they sound as good as they do, given that they’re based on technology that dates back almost a century. Of course, there have been lots of improvements along the way — better materials, more accurate crossovers, computer-tuned enclosures, and the like. But the concept of using a magnetic field to push a material (paper or plastic) that in turn pushes air seems about as resistant to revolutionary change as, well, the average major record label.


I haven’t forgotten that this is supposed to relate back to mastering, so here’s the deal: One of the main functions of mastering is to make a mix transportable. In other words, it will sound good over any playback system.

What works against this is that playback systems vary wildly. Although we no longer have to worry about all the violence that happens to a signal courtesy of analog playback media (cassettes and vinyl, which offered a nearly infinite number of possibilities to screw up sound), speakers and their associated enclosures are like very mischievous equalizers. For a mix to be truly transportable, it needs to play over:

- Great audiophile systems with flat response, superb definition, and state-of-the-art speakers.

- Boomboxes that have the “MegaGigaSuperBassBoost” button pushed in, which hypes the speaker’s low end to an absurd degree.

- Boomboxes that don’t have the “MegaGigaSuperBassBoost” button pushed in, so the low end is thinner than Kate Moss on a hunger strike.

- Earbuds on portable music devices, whose response is basically luck of the draw.

- Car radios. Let’s not even go there, even though we have to.

- Tabletop radios, which have the same type of issues as boomboxes.

What do all these different systems have in common? Speakers (even if they’re miniature versions that fit into ear buds). How do we compensate for differences among speakers during the mastering (and for that matter, mixing) process?


Edwin Abbott’s Flatland is one of the greatest mathematical fantasy books of all time, albeit in an admittedly un-crowded field. But when it comes to speakers, “flatland” is just that: a fantasy. Look at the response curve for even the finest speakers that money can buy, and you’ll see something that approximates a relief map of the Alps (the part on the Swiss/Italian border, with the really wicked peaks). This response (see Figure 1) undergoes further degradation when interacting with the listener’s room, which is seldom acoustically treated; but let’s pretend that’s not an issue, as it multiplies the variables into the world of Really, Really Big Numbers.

Differences among speakers exist in the lows, mids, and highs (did I leave anything out?). So, over the years, mastering engineers have recognized that the only way to deal with this madness is to create a recording with the flattest, most “average” response possible. That way, it will sound only a little bit “wrong” over every system, rather than okay on some systems and way wrong on others. (The exception is that of the audiophile with the really flat system — who after putting the requisite time, expense, and effort into assembling a great system, should be entitled to the best possible sound.)

It’s difficult to create a truly average midrange response, because that’s just one of the places where speakers exhibit significant differences. (An aside: I always get a kick out of speaker reviewers who breathlessly exclaim that a particular speaker “revealed things I’d never heard before in my favorite recordings!” This isn’t surprising, because any speaker will reveal things you’ve never heard before, as it’s basically EQing the recording differently.)

High frequencies are a different matter. The energy in real music tends to drop off fairly rapidly above 5kHz, so really, what we want is a “sensation” of brightness. You’re not going to get the huge peaks caused by notes piling on one another, because there just isn’t that much energy up there. A little bit of a boost in the “air” range above 10kHz will do wonders for making a mix transportable, as the tweeters open up a bit more. As a bonus, most playback systems have a treble control that can be trimmed or boosted according to the listener’s preferences, based on their acoustics and how shot their hearing is from going to too many concerts without hearing protection.


Perhaps the most crucial frequency range for making a mix transportable is the bass range; there are several reasons for this.

Speakers tend to have the hardest time maintaining a flat response below about 80Hz. Here we’re up against the laws of physics, as bass frequencies have really long wavelengths and to re-create them, you need to push a lot of air. Speakers that fit in the average living room, or boombox for that matter, simply can’t push enough air at really low frequencies. This is one element that separates the big bucks speakers from the pretenders: How low they can go without giving up. This is also why many devices have bass boost switches, although that’s not quite the same as having “real” bass. An analogy: A woman puts on great eye makeup that makes her eyes look bigger, but really, they aren’t any bigger.

To make matters worse, we have two other bass range issues. For one, the bass response of our ears falls off at lower volumes (the infamous Fletcher-Munson curve), so our ears’ deficient frequency response interacts with the speakers’ deficient frequency response. The second is in the recording itself. Unless the studio has really great acoustics, or all instruments were taken direct, there will likely be frequency response anomalies due to room acoustics that cause peaks and dips in the bass range.

So in a worse-case scenario, the bass peak in the recording process doesn’t get caught while mastering, and plays back through a speaker system that has a resonance at that frequency, which interacts with the room resonance in which the speaker lives. Ugh.


There is no solution, so you’re correct in feeling hopeless and depressed. But you can try to come as close as possible to the ideal. If during mastering you can smooth out the bass response to have no significant peaks and dips (except where you actually want them, like a big peak on a techno kick drum), you’ll have gone a long way toward making a transportable master. It’s even better if you can take care of some of these issues in the mix (see sidebar).

Because our ears’ response gets iffy in the bass range, it can really help to have some visual feedback as to what’s going on in the bass range. No matter how good an engineer you are, it’s really hard to quantify a 1.5dB peak at 72Hz solely by listening. I find a good spectrum analyzer that can display an average response is extremely helpful. Why average? Because there will always be natural response peaks and dips. What we’re looking for is a pattern of build-ups and anomalies that, if played back through a system with a peak or dip at that same frequency, will exaggerate the problem even more.

My favorite tool these days for fixing this type of problem is the Har-Bal EQ/mastering program. It provides a very unambiguous look at the bass end, and you can use a simple “pencil” tool to draw out peaks and dips, thus smoothing out the low end (Figure 2). Since doing this, my mastering clients have all — without any prompting — commented that the mixes are more transportable. Granted there are more factors than just bass response that contribute to making a mix transportable, but bass is an important factor.


Another great mastering tool that helps compensate for speaker anomalies is a good multiband compressor (Figure 3). Some subtle midrange compression can bring down peaks and raise valleys that might otherwise be emphasized or de-emphasized by a speaker. In the high end, you can add no significant compression, but just boost the overall level a bit. Meanwhile, in the 300–400Hz range, you can lower the level a bit (without compression) to “tighten up” the sound a bit, as that’s often where you’ll find a bit of “mud.” Meanwhile, adding compression in the bass range helps even out the response.

Eventually, through proper use of equalization and dynamics control, it’s possible to create a master that keeps unwanted peaks and dips under control. And when you’ve accomplished that, you’re in good shape.

As a final reality check, it is worth playing your master over varying systems just to make sure you’ve come as close as possible to the ideal. If your mix sounds a little fatter than usual on systems that normally sound a bit on the thin side, thinner than usual on systems that normally sound annoyingly muddy, and perfect on really good systems, great. No matter what speakers the recording plays back over, you’ve probably done about as good a job as you can do.