Above sketches illustrate a setup, but parts may move up/down/etc. since midwoofer radiaton suffers, section shows woofers too close, etc.
Dipole behaviour has many advantages in acoustically small spaces. Especially in the bass compartment (efficiency-loss notwithstanding), but these advantages diminish with frequency rising, and disappear somewhere in the upper voice (2-2.5kHz) where combfiltering starts taking over from the fig-8 radation.
Line arrays have their own advantages, that are poorly researched for acoustically small spaces, since almost all research and consequent products are based on far field situations and their calculations. There IS however some research directed at home applications, where you will find the in-home line source is more forgiving than expected (Keele, Linkwitz, Griffin, e.a.). Shading of various sorts helps, not hurts (amplitude, frequency and physcal shading) to mitigate some inherent and some potential anomalies, while maintaining just about enough of the advantages.
(James Griffin bundled research by all the greats and summarized home-use guide lines).
(Merlijn van Veen also studied and published, putting assumed LA behavior in perspective).
Line sources carry (most) sound further at -3dB per doubling of distance, but this behavior depends on adhering to parameters and decidedly falters below close to 700Hz, which is therefore the point where i plan to cross from one principle to the other. Room modes are said to make up for this faltering. But then we are back at a very close and limited near field!
Or are we? Dipole at bass freqencies typically has a 4.8dB lower reflected energy compared to monopoles at equal on-axis SPL (Linkwitz e.a.). Regardless of said -L. So i'll tune that up slightly, without losing the better near field behavour of 75% less reflected energy. Whaddayasay?
When i cross somewhat steeply at 670Hz, dipole behavior is still fully active, but room modes were just calming down in the upper 335-670 octave of the close-to three octaves this 12" (262mm effective diam.) driver would cover.
Keele's physically- AND amplitude-shaded CBT's have a wide listening area, sitting and standing, from where stereo imaging remains intact and projected somewhere behind the speakers. So in-home, some parameters have other effects than in concert halls and football stadiums.
Keele (who turned 87 just months ago) is still famous for his CBT's belying far field-based calculations, and EVERYONE who has them or has heard them is raving about them.
Martin Logan mixes principles in their monopole bass/dipole electrostat (line-like) and get bad reviews as far as sweet spot is concerned (small!!). But monopole has 300% more reflected in-room energy (when calculating back up, 100% represents 300% more than aforementioned 25%), and has a very wide diafragm, that will beam and not carry very far, about which a the 22" and 50" measurements by Stereophile tell us nothing.
So i plan on hybrid application of two inherently different principles in a single design.
Away from shown sketches, i will keep open the option to shift to:
-line array for the midbass as well;
-exchanging the 12" midbass 100-670Hz for another type;
-putting a second midbass-driver back to back (summing SPL).
And the 670Hz-18kHz upper bodies are swivelling sections of the speaker pair, allowing both toe-in and toe-out up to 23 degrees, for central positioning in a room (avoiding a portion of early sidewall reflections in favor of late rear wall) as well as traditonal triangled or straight orientation.
Ah, and another thing:
Every speaker design needs fine tuning when placed in a listening room IMO, and SOME rooms also need fine a bit of tuning because they deviate too much from average (between dead and live).
But my project is definitely not aimed at one room. I aspire to illuminate wide moving sweet spots for the basic open plan living rooms most people in my country inhabit. (As the saying goes: dress for the job you want, not the one you got.)
As such, acoustic considerations are leading the design, seeking a result that does not depend on DSP to solve issues. Going against physics and then mend flaws by DSP is a path to disaster: Response anomalies filtered out, usually cause new (sometmes worse) anomalies elsewhere.
So DSP will be used exclusively,
-to equalize driver response;
-to implement XO filtering;
-to maintain phase coherence IF necessary;
in that particular order.
Now shoot me dead, win a prize, or at least my gratitude, while doing it.
Show me problems, solutions, questions, answers, research on in-home line arrays, and how i am an idiot.
Tx and cheers!
