In fact, the two most important specifications are Dynamic Range and Frequency Response. They are the specifications whose numerical values will have the most impact on the performance of an audio product and, more importantly, the music you create using that product.

Dynamic Range speaks to the resolution of the sound: how clear, crisp, transparent, punchy and realistic it appears, as well as free from noise and distortion. If the dynamic range is too small, sounds will get buzzy, harsh and distorted rather than louder, and soft sounds will disappear into a distracting fog of hiss, buzz, snaps, crackles and pops instead of resonating and reverberating above a floor of “inky black silence,” to use Rupert Neve’s phrase.

The 120 dB dynamic range of our hearing is a huge range (1 trillion to one, if you recall), and it presents a difficult challenge for equipment designers. Fortunately for the electronics guys, acoustical limitations (the residual background noises of our environment on the soft end and the limits of physiological tolerance of our neighbors and the local sheriff on the loud end) tend to obviate the need for us to have 120 dB of range available all the time. For music-quality audio the 90+ dB range of recorded sounds that 16-bit digital audio permits seems to be pretty much adequate.

Within that 90 dB range, the issues of clarity and transparency, of spaciousness and depth, of punch and impact, and of emotional expressiveness are all issues that are both limited and driven by dynamic range.

Frequency Response has to do with timbre. Timbre is a funny word that means “how-you-tell-two-instruments-apart-that-are-playing-the-same-note-at-the-same-loudness.” A primary aspect of timbre (but by no means the only or even most important one) is the distribution of energy across the spectrum for a given sound. As we change frequency response, we are changing timbre, and such changes are both easily audible and of considerable musical significance.

As I suggested above, in the electronic realm, it isn’t very hard to get excellent frequency response across the audible spectrum (i.e. a range of 10 Hz. to 30 kHz., +/- .1 dB). The current crop of op-amp chips can do this easily. Interestingly, sometimes we get ourselves in trouble by making the electronic response range too great. I’ve managed to inadvertently generate megahertz signals (and DC ones as well -- DC meaning direct current, which is approximately the same as 0 Hz.) that have efficiently and silently destroyed my tweeters, woofers, power amplifiers and bank account contents on numerous occasions. You don’t necessarily want an awesomely killer 4 Hz. component in your audio mix, and if the frequency response of your audio gear doesn’t permit it to happen, that’s probably saved you a fair chunk of change.

In the electromechanical realm, it’s a lot tougher to get good frequency response. Floyd Toole suggests that excellent loudspeaker frequency response consists of a range from 30 Hz. to 20 kHz. +/- 1.5 dB. There is considerably more to the measurement and specification than this, but if you don’t obtain at least this measured performance on axis in an anechoic chamber, the speaker will not be evaluated as excellent by listeners in controlled double-blind tests.

The Zen aspect of all of this is that the essential holistic oneness of a sound (Ohm, Ohm) grows out of its place within the Audio Window. The larger the window opening, as provided by the product, the more palpable and real the sound will appear to be, in terms of its timbre and presence. If the frequency response and dynamic range qualities of the product are as large as the window itself, it will probably sound fantastic. If they are significantly smaller than the window, even if you don't hear the problem on the showroom floor, I suspect you will hear it later - probably as soon as you get the piece home, out of the box and patched in to your system.

Once these qualities are good enough to no longer present problems for you, then you can start worrying about some of the other performance specifications, some of which have really important sonic implications and some of which are nothing but product features. Only you can decide between 32 voice/6 timbre and 26 voice/12 timbre sound modules, and whether or not you need +4 dBM balanced inputs and outputs, for instance.

At the same time, try to read specifications carefully and figure out how they fit into the Audio Window. It will help you separate the useful numbers from the useless and/or silly ones. And that in turn will help you pick better equipment that costs less and gives you more capability for successfully creating your music. Happy tracks!

Dave Moulton is preparing Pink Noise for sale to train your ears (a crass commercial plug). Peter Alhadeff is preparing economic projections for Berklee students, and Alex Case prepared most of this article.