Look at any audio forum or magazine that has technical leanings and you will see an ongoing debate regarding “objectivity versus subjectivity” or the supremacy of measurements over listening. I think that these discussions in themselves are pointless because neither side will ever convince the other, but that wouldn’t particularly annoy me if I felt that one side was ‘right’. Neither side is.
- the people doing the listening (the “subjectivists”) have no solid criteria against which they make their claims. We only have their word that they have Golden Ears. Judging by some systems I have heard that are raved about by others, this means nothing.
- the people advocating measurements (the “objectivists”) have no solid criteria against which their particular claims can be demonstrated to correlate with sound quality. The measurements they brandish might be slightly useful to a technician on a work bench but are virtually meaningless in trying to determine a system’s overall quality.
As such, the two sides are not even having an argument, but merely talking to themselves.
The listening-only people cannot win the debate, but at least their argument is coherent. They can say: “If the whole aim of this audio malarkey is to stimulate the ears in a pleasurable way, then we can say that this has been achieved – at this moment and in the current circumstances for this particular listener even if it was all in his mind”.
In contrast, the only thing the measurements people can say is “If the aim of this audio malarkey is to reproduce a wiggly line faithfully, then in some respects we can show that this system does it better than some other systems – when judged against certain criteria that may or may not be important to the sound”. It is a very feeble argument.
Measurements can only cover a tiny fraction of the overall ‘information space’ that an audio system processes. To argue that sparse measurements of a few parameters using a few specific test signals under specific environmental conditions somehow encapsulates the performance of an audio system is completely bogus. It can only work if it is also argued that the device is utterly predictable and, in effect, that we know exactly how it works – in which case the measurements are superfluous: we could simply study the circuit diagram and simulate the whole thing in software. Using measurements to establish a design’s basic validity (which is what we are implying we are doing if we think that a few measurements can characterise a device fully) is a throwback to the time before computers.
And then there is the basic misunderstanding of what measurements mean. Simple two-dimensional measurements (that might provide limited information about the operation of an amplifier, say), are frequently abused in an attempt to tell us how speakers will sound in a room – which is a problem of multifarious dimensions and yet is probably much less critical than many people make out: over-simplification and over-complication at the same time, and never conclusive. Factors that defy simple measurement are the real determinants of sound quality. For example, compare the sound of an active speaker with that of a passive multi-driver speaker. It’s a finite difference, but not easy to measure and summarise in a couple of numbers. I would ask: why bother even measuring it? Just make the speakers active and forget about it!
There are other such factors that we know are detrimental to the sound without even measuring them and which are fundamental to antiquated designs:
- non-DSP corrected speakers introduce large phase shifts and/or timing misalignments that are simply not in the original signal. Objectivists (those who know anything about it) accept on faith that phase shifts are inaudible, based on some dubious experiments carried out long before it was possible to even create a hi-fi speaker without phase distortion.
- bass reflex speakers introduce many strange distortions to the signal, but are near-universally accepted as hi-fi by audiophiles of all persuasions
- speakers that cannot reproduce bass fully, change the whole balance of the sound in a very artificial way, yet are embraced as High End nevertheless.
These problems can be eliminated at a stroke at the design phase. There is no point in agonising over them.
The religion of measurements is a red herring. A great system can be designed on paper without using measurements at all. When it is built it can be perfected using measurements of the right kind, but they are purely ‘tuning’.