most accurate loudspeakers....

Dear Beemerrider,
I totally agree with you. The room and the speaker system have to be one organism. You cannot have one box playing in another box and hope for accuracy. You will only get higher and higher levels of audiophilia in conventional audio, but not accuracy. My two bit, if ever any of you go to Denmark, check out the SLT monitoring system in Focus Recording(www.focusrecording.dk) studios.

Active ATCs or PMCs. Take your pick.

Another way to ask the quesion is, which speaker reproduces the analog signal with the least amount of change? How would you know?? It's all relative isn't it? If you listen to two speakers on the same system, you may think that one of the speakers produces a voice more "accurately" but, how would you know?????

It's easy to define accuracy, hard (impossible) to achieve. Compare the waveform going in to the waveform coming out. The closer the output is to the input, the more accurate the speaker.

Many factors are involved in accuracy. All forms of distortion, harmonic, intermodulation, linear, phase, time. Not all forms of distortion are equally audible or important from a psychoacoustic standpoint. Other factors include, the stored energy of the drivers, the cabinet resonances, refraction, and interference patterns.

Room variations and anomolies must be eliminated for a true measurement. But, of course, the room comes into play for the user. This brings up another factor, power response. How does the speaker 'light up' the room, and how well does the ambient sound field correspond to the direct sound? Speakers can measure very well on axis, and create a miserable sound field that colors everything.

These factors also need to be prioritized. Suggested priority:

1) Linear response
2) Harmonic and IM distortion
3) Stored energy
4) Power response
5) Phase and time

In my opinion the goal is to recreate as closely as possible the illusion of a live performance, and this implies taking psychoacoustics into account. Just because something can be easily measured doesn't mean it's relevant. If low-order harmonic distortion is inaudible, then an impressive-looking measurement in that area is irrevelant.

To give a more common example, as long as we measure the on-axis anechoic response but listen to the power response, there will be a disconnect between frequency response measurements and subjective perception.

Following this line of thought, the most accurate loudspeaker is the one that comes the closest to reproducing the perception of a live performance. We don't presently have a metric that accurately predicts subjective preference - indeed, the audio industry has thus far resisted the adoption of such a metric presumably because there would be far more losers than winners among manufacturers. What we'd need is a psychoacoustically-weighted metric that takes into account the relative audible significance of everything the device does to the signal (the "transfer function).

Of course I have my opinion as to which characteristics are likely to result in a speaker that closely approximates the perception of a live performance, and frankly some of these characteristics are unlikely to show up in published measurements. Among these obscure characteristics that I think matter are dynamic linearity (lack of power compression or at least uniform power compression across the frequency range) and radiation pattern smoothness (related to but not the same thing as power response).

I don't know of any speaker that really "does it all", but my personal leanings are generally towards planars and horns rather than direct-radiator dynamics - even though the latter are more likely to produce impressive-looking conventional measurements.

Duke
dealer/manufacturer