In response to the questions raised about my 1988 double blind CD player comparison, I would clarify that both players produced the "standard" output of 2.0 volts with a 0dB DFS test signal. They were measured with a high quality VTVM and the difference was only about 0.03V - far less than even 0.1dB. If there had been a substantial difference in level, it should have been easy to differentiate the CD players, and this was not the case. The point of the test was to determine the sensitivity of the double blind test with actual music in a case where significant differences actually existed. I believe that the results show how insensitive these tests can be, even with well-controlled methodologies and experienced listeners.

This conclusion is not a defense of "golden-eared" reviewers seeking to convince others of the superiority of one component over another, it is simply an illustration of how double blind testing is not very effective in revealing differences that may exist. Varying the test conditions and signals used in evaluations may provide better results in some circumstances, but it seems that the routine ABX methodology leaves a lot to be desired.

Over the years it has become clear to me that some electronic components are audibly superior to others, even given similar measured performance. This is even more true in the case of loudspeakers where measurement techniques are more variable and often yield complex data that is difficult to interpret. A thorough evaluation requires both listening and measuring. Comparisons between components are critical; however, double blind tests using routine methodology seems to add fairly little.