The following is based on a series of my postings at DiyAudio.com from 25 March, 2008


I built my first speaker at the age of 16 and I was never fully satisfied with my results since then; even when I became more skilled, and learned more about acoustics, electronics, mechanics, physiology of the hearing system, and human perception in general ...

What is the search about ?

My view has changed a lot in the last few years. I always anticipated that minimizing 'errors' in a technical way alone does not lead to good results. We have to consider the qualities on which our perception system is based. Today I think that trying to build good systems is to explore our perception of things. It makes me wondering every time, when I listen to a certain system, which gives rather good performance subjectively while performing poorly from the 'measurement' point of view. Some systems recognized as being 'good' by a wide range of listeners suffer from obvious technical drawbacks like:

patchy amplitude response v. frequency

high non-linear distortion in a certain frequency range

... and so on.

Those imperfect examples are often more instructional, than the exploration of a system which is 'known' to be unobtrusive from the measurement point of view. This is because it points me to what is important and what is not.

My attempt here is to offer you my current, very subjective, but incomplete list of important and unimportant traits, maybe as fodder for further discussion. I am sure that your 'lists' will differ, but maybe there is some overlap:

Smooth amplitude v. frequency response at a fixed point in the listening room is NOT VERY IMPORTANT

Smooth amplitude response outside the critical bandwidth (say broadband telephony 0,2 … 7 Khz bandwidth)
   at a fixed point in the listening room is EVEN LESS IMPORTANT

Large bandwidth is IMPORTANT

Balanced dispersion of energy over the entire frequency range in the listening room as a whole is IMPORTANT

Controlled directivity which means "no sharp discontinuities in horizontal and vertical radiation angle v. frequency"
   is VERY IMPORTANT

Spectral decay above the critical bandwidth is NOT VERY IMPORTANT

Spectral decay at very low frequencies is NOT VERY IMPORTANT

Spectral decay from mid-bass to the presence range is VERY IMPORTANT

Invariance of all parameters considered important against different power levels (especially against very low power levels)
   which are needed for the perception of subtle dynamic changes is VERY IMPORTANT

Every distinct pattern of "misbehaviour" of the reproduction chain like

a resonance pattern or

a pattern of sharp directivity changes vs. frequency or

a distortion pattern occuring at certain levels of power

makes the transducing device identifiable easily to our hearing system and hampers it in its attempt to compensate for 'minor flaws', which are allowed.