Enceintes Magico

[ThierryNK]

(02-22-2024, 02:42 PM)jalucine a écrit : ...
Je ne comprends pas cette approche.
...

De la même manière, ni Alon Wolf, ni Lawrence Dickie, ni l'immense majorité des concepteurs professionnels ne comprennent pas qu'on puisse utiliser des filtre à 6dB, pour un prétendu avantage que la psycho-acoustique bat en brèche en plus des éléments techniques suivants:

Par exemple 
https://www.theabsolutesound.com/article...alon-wolf/

Citation :The notion of time/phase alignment/coherence, at least as it is advanced in today’s high-end loudspeaker marketing schemes, is extremely misleading (I will address the shortcomings of the concept in a passive design first.)


Trying to keep things simple, let me just highlight the two MUST conditions where such concepts are even probable:
1) a first-order acoustical crossover, i.e. a perfect 6dB-per-octave acoustical slope from the designated bandpass;
2) a physical alignment of the drivers’ acoustical centers, which, unless a concentric driver is used, is only possible for one point in space at a time.

Only if both conditions are met is time/phase coherency even probable. Just moving drivers around will not suffice to achieve time coherency. In fact, such designs will ensure a “non-optimal condition” at any point due to the fact that, if a first-order XO is not used, any driver movement will require XO realignment to keep the proper phase relations among drivers at the XO points.

There have been honest attempts at such designs, including some that do meet the basic conditions; however, even if these criteria are met, the compromises needed to be taken to achieve these conditions are detrimental to overall sound quality.

Staggering drivers, in order to align them, in a stepped baffle creates tremendous amounts of diffraction. Unlike time/phase coherency, which has never been proven to be a factor in perceived sound quality, diffraction has indeed been proven to be a big detraction.

A 6dB-per-octave acoustical slope requires a very complex XO, with many parts, which also cause degradation in SQ and by themselves introduce time delays (that is why the actual notion of a truly time-coherent passive loudspeaker is questionable). A simpler XO is possible using non-psionic drivers, at the cost of losing low-level information and increasing distortion due to non-pistonic cone movement. Moreover, with 6dB-per-octave slopes in a typical three-way design, the bass drivers will be only ~18dB down at 2kHz, playing right into tweeter territory. Not to mention the tweeter playing into the bass region.

The unavoidable inherent trade-offs of such [time-and-phase-aligned] design are significant:
1) Big increases in IMD (intermodulation distortion), which clearly affects SQ.
2) Increased 2nd and 3rd harmonic distortion due to shallow crossover slopes.
3) Drivers firing at the listening position asymmetrically—i. e., off-axis (the need to “tilt” drivers to aim at listening position).
4) Limited vertical dispersion.
5) Reduced power handling.

So, weighing all these trade-offs against the fact that it has never been proven that time alignment is essential to SQ, time/phase alignment/coherence as a goal in loudspeaker design is easy to pass.