Scene 1. A person goes to a particular room at an audio show to hear a specific component – a DAC. But… how would that person know how much of what was heard could be attributed to the DAC, in an unfamiliar room, with unfamiliar amplification and speakers?

Scene 2. A person searches the audiogon listings for components on the Stereophile recommended components list. Not even necessarily Class A – they cost a packet – but if only a better amp or preamp or speakers or DAC could be found from anywhere on that list, things would be so much better.

Scene 3. A group of people on a message board get all excited over some new discovery. New people join the board and discussion just to learn more. Everybody wants one, the lucky few get one and duly provide glowing reports that make the rest just want it more. (The vendor of said component is probably astonished at the 1000% leap in sales figures that month.)

I’m sure we all recognize ourselves in at least one of these scenes. Scene 2 is me, embarrassingly enough, a decade or so ago. And it sometimes takes a conscious effort for me to step back from getting caught up in Scene 3. It’s the “killer component” syndrome – the belief or desire for a single component to provide a quantum leap in the quality of sound reproduced by our stereo system.

But in reality, it’s the hifi version of Fred Brooks’ “silver bullet” – the magical something that will solve the problems (slay the werewolf, in his metaphor) of software development. It’s the “secret ingredient” that makes the master chef’s dish so much better than everyone else’s. It’s what the Jones’s have that makes their life so much better than yours.

I guess you get the point, no need to belabour it. But, I’ve been thinking about how one could approach a hifi system from the opposite direction. First, put together a system (in actuality or as a thought experiment) using extremely high-value components. Components that don’t make the Stereophile recommended list, but which perform better than you would think. Then step back and look at the system (electrical and acoustic) as a whole, and start to address the weak points, the areas of poor performance. Not by substituting “better” components by guesswork, but by looking at it from an engineering perspective.

What are some examples of likely “weak points”? The interaction of the speakers with the room is a very likely one – poor positioning, early reflections, and long reverberation times can all be measured and improved. Limited dynamic range is another – speakers that are too insensitive or simply too darn small to produce the needed output for transients being the usual culprit. Low end frequency response can almost always be improved – very few people realize that low frequencies are necessary to reproduce transients properly, and not just about reproducing bass notes or explosions. Finally, basic electrical problems with noise, grounding, impedance, and gain, seem surprisingly common.

I wouldn’t normally consider myself one of those “engineering” wet blankets that show up at every online audio party and demand measurements or blind testing or whatever. But without really realizing it, I think I’ve gradually been coming more and more around to the way of thinking outlined above for a while now – address the weak points using an engineering approach, instead of focusing on the next “strong point”. For example, I said nothing at all about amplification in the previous paragraph, not because I don’t think it’s important or that you shouldn’t choose amps that sound good to you, but because the amp is most likely a fair way down the list of weak points that would need to be addressed in this approach to system building.

This approach has perhaps become more feasible than it used to, because the quality of affordable electronics has improved markedly in the past decade (that’s certainly my impression anyway). This is important because most of the budget is going to be spent after setting up the initial system.

Now, nominally there’s going to be a point where the obvious engineering problems have been solved, and one can focus on finer points that increase the subjective enjoyment of the system even further. In practice, you might do both at the same time. Use valve amps, full-range drivers, non-oversampling DACS – whatever floats your boat, as long as you don’t introduce new problems.

How well does it work? I’m going to find out 🙂