Fantasy Audiophiles vs. Objective Audiophiles: Has the hobby changed?
Reply #35 – 2010-04-09 18:25:45
Richard, Thanks for the excellent post. I definitely consider myself an objective audiophile. Differences can be measured when they exist (provided we make the correct measurements with the required accuracy). Likewise, differences can be detected in double-blind tests if they are large enough to create an audible difference. Let me present my perspective as a designer of pro-audio and audiophile equipment: I believe in ABX tests and use them on occasion when developing and testing products. I rely much more heavily on measurements when designing products. If I measure a defect and can cure the defect at little or no cost, I go ahead and fix the defect. It is usually much easier (and therefore cheaper) to fix a measured defect than it is to determine whether or not it is audible. I design with a wide safety margin to keep defects well below audibility whenever possible. Often the difference in parts cost is only pennies. If we were building millions of units, each penny would count. We build thousands of units, and our development costs are a significant portion of our total costs. The few pennies spent on better components, or extra ground plane layers on a circuit board are trivial.As an objective audiophile, I have occasionally been surprised by the unexpected: I decided to test speaker cables to show that the differences are insignificant. I expected to demonstrate that 18-GA zip cord was indistinguishable from high-quality audiophile speaker cable or even the heavy-gauge cables used by the sound reinforcement industry. I was shocked to discover that there were differences, and more shocked to discover that the zip cord performed better than most of the other cables! Let me add that we do not sell speaker cables, nor do we have any plans to do so in the near future – I have nothing to gain or lose from this discussion. All of the speaker cables tested performed well when loaded with an 8-ohm resistor. I substituted an 8-ohm JBL 4410A studio monitor, and the cables performed very differently. The speakers do not present an 8-ohm load over the entire audio band. The actual impedance varies from 1-Ohm to about 16-Ohms. The impedance variations produced frequency response variations. I then set up a demonstration that allowed us to listen to the error signal across the cable, played back through another JBL 4410A at the correct amplitude. We could switch between long cable, and short cable, and cable error signal, and demonstrate audible differences with 100 foot lengths of cable, but no audible differences at 12 feet . One surprisingly poor cable was 10-GA SO cord. The SO cord is the thick black neoprene jacketed cord (with many fine strands of copper) that is used for heavy duty AC power cords. This cord is commonly used in long lengths (100 feet or more) for large commercial sound reinforcement systems. This cord has lots of copper and had the lowest DC resistance, but surprisingly, it had the worst measured performance, and the most audible effect on the music played. The reason for the poor performance is that the cable has far too much inductance, and far more inductance that the cheap 18-GA zip cord that we tested. It turns out that the inductance of the speaker wire is much more of a factor than the DC resistance! The conductors must be closely spaced to achieve low inductance. Telephone or Ethernet twisted-pair wire has very low inductance, but high DC resistance. Multiple twisted-pairs wired in parallel can achieve near-perfect perfect performance at very long lengths. 25 pairs (in parallel) at 100 feet driving 8-ohms are astonishingly good. 10-GA SO cord at 100 feet is surprisingly bad.Final thoughts: I thought the claims about speaker wires were ridiculous, but it turns out that the differences were much larger than I expected. I thought that the heavy-gauge wire would perform the best - it was actually the worst. Like many things, expensive is not always better.Moral of the speaker-cable story: Test before claiming that differences don't exist. Test before claiming that differences do exist. Don't make claims without testing. Don't waste money on claims that are not backed by good test data.