If there is a preference score then presumably the perfect directivity with a score of 10 must be known? Can someone please post what it is?
I don't understand preference score. How/why did I use something I don't understand? I used it to learn when the score is useful to warrant smoothing the Power + DI curves in a simulation prior to cutting into plywood - which is expensive these days. From the posts so far, it looks like it isn't a big enough difference to matter so I should look at different things.
My experience with speakers having variously different DI curves is that a smooth DI and power response matters up to a point. The power response should smoothly decrease with frequency, and the DI curve (by definition) should smoothly rise with frequency. However, once the power and DI curves are within +/- 2 dB of being "smooth", any further refinement does not seem to make much difference to me... or to be more precise, I may be able to hear a very small difference, but I can't say that I prefer one curve over another.
Thanks. Helps me learn when close enough is enough.
I don't understand preference score. How/why did I use something I don't understand? I used it to learn when the score is useful to warrant smoothing the Power + DI curves in a simulation prior to cutting into plywood - which is expensive these days. From the posts so far, it looks like it isn't a big enough difference to matter so I should look at different things.
The directivity not only matters it dominates the perceived difference between well designed speakers. The problem is that the optimum directivity is a significant function of ones preferences and expectation w.r.t. imaging and spaciousness (e.g. concert hall orchestra vs studio rock vs ...), the details of the recording, the room response and no doubt one or two other things. There is no perfect directivity or single number to measure the goodness of a directivity at least without a lot of qualifications.
So a single figure cannot reliably rank speaker's with a good directivity except possibly with a lot of qualifications to assign specific values to all the variables involved (assuming the calculation of the number includes all the relevant variables in a reasonable manner). What it can do reliably though is reveal speakers with poor directivities if the objective is neutrality rather than introducing attractive sound effects as seems to be the case with a fair few expensive home audio speakers. So take notice of the low numbers but less so the high if you are seeking a neutral sound.
Phewwww!
Generally human ears distort (THD) at 70 dB SPL about same amount as average speakers on 90 dB level. Ears distortion can be measured with beat tones. For example you play with headphone 1 kHz to one ear and 3005 Hz to to other ear, then 3005 Hz level where 5 Hz beating is biggest, is actual ear 1 kHz signal 3rd harmonic level.
Last edited:
I have been asking the same question myself, but never got to test it. @kimmosto advocated for tweeter to be less directive at XO than woofer to compensate off-axis vertical nulls. Can't find more elaborative post, but here is his basic rules for 2-way XO: https://www.diyaudio.com/community/threads/vituixcad.307910/post-6538511
Revel and JBL seem to compensate in a somewhat different manner, employing diffractive widening, as done e.g. in 228Be:
Off-axis nulls at 2 kHz are compensated by diffractive widening in the same region; while Q of the nulls and peaks can't be matched exactly, this compensation still results in rather flat DI.
However, Revel Salon 2, which has comparable bass extension but shows virtually no diffraction, was preferred in a blind testing by not-so-large, but statistically significant margin. Although, maybe, this preference came simply from wider dispersion of the Salon 2 or some other, less obvious reason.
(see attached PDF for test details and measurements)
Greg Timbers seemingly jungled rather narrow band peaks and dips in his designs, as in Studio 530:
https://www.spinorama.org/speakers/JBL Studio 530/ASR/index_asr.html
Even more peculiar is Perlisten S4b design, whic compensates ceiling ER dip with floor bounce peak, leaving horizontals intact:
Is it some CEA2034-specific optimisation, or, as 2034 represents typical listening room, such lobe aiming will work in most real environments?
Overall, I can't come to a definite conclusion. On one hand, it is indeed common to cross-compensate reflections in some way; on the other, lasting sucess of Ultima Salon 2 and similar designs may support a preference for unmarred horizontal radiation patterns.
Revel and JBL seem to compensate in a somewhat different manner, employing diffractive widening, as done e.g. in 228Be:
Off-axis nulls at 2 kHz are compensated by diffractive widening in the same region; while Q of the nulls and peaks can't be matched exactly, this compensation still results in rather flat DI.
However, Revel Salon 2, which has comparable bass extension but shows virtually no diffraction, was preferred in a blind testing by not-so-large, but statistically significant margin. Although, maybe, this preference came simply from wider dispersion of the Salon 2 or some other, less obvious reason.
(see attached PDF for test details and measurements)
Greg Timbers seemingly jungled rather narrow band peaks and dips in his designs, as in Studio 530:
https://www.spinorama.org/speakers/JBL Studio 530/ASR/index_asr.html
Even more peculiar is Perlisten S4b design, whic compensates ceiling ER dip with floor bounce peak, leaving horizontals intact:
Is it some CEA2034-specific optimisation, or, as 2034 represents typical listening room, such lobe aiming will work in most real environments?
Overall, I can't come to a definite conclusion. On one hand, it is indeed common to cross-compensate reflections in some way; on the other, lasting sucess of Ultima Salon 2 and similar designs may support a preference for unmarred horizontal radiation patterns.