Good question. "Flat" off-axis with gradual attenuation is physically almost mission impossible (multi-way dipole and cardiod speakers can do this best). Music spectrum is 20-20.000Hz which means wavelenths 17,2 meters - 0,0172meters. Transducers trying to reproduce this variance are in trouble, and the higher the requency, the harder it gets to maintain wide and smooth dispersion.Is perfectly flat not the ideal result? The only remaining question being should it be attenuated compared to the on axis, and if so how much at each angle.
Is perfectly flat not the ideal result? The only remaining question being should it be attenuated compared to the on axis, and if so how much at each angle.
Revel/Harman has a "anechoic correction" curve for this. It is mild but, to me, makes all the difference.To me the Revels are brighter than I like for music playback.
Good question. "Flat" off-axis with gradual attenuation is physically almost mission impossible (multi-way dipole and cardiod speakers can do this best). Music spectrum is 20-20.000Hz which means wavelenths 17,2 meters - 0,0172meters. Transducers trying to reproduce this variance are in trouble, and the higher the requency, the harder it gets to maintain wide and smooth dispersion.
In listening tests it has been shown many times that people around the world prefer a slightly tilted power response at listening spot. A typical 3-way speaker with dynamic drivers and flat on-axis response at 2m distance can achieve this goal rather easily. If the loudspeaker has different or uneven directivity profile, the task gets more difficult or even impossible.
The listening room always modifies reflections and reverberant energy and our perception has it's own rules for direct vs. reflected sound.
I recently did a blind test comparison between the KEF R3 and Ascend Sierra 2EX. I’ll probably cross-post that thread to these forums (if that’s allowed), but in summary: the Sierra 2EX won easily, by a large margin.
The Sierra 2EX actually has remarkably wide and smooth dispersion all the way up to 20khz, which as you say is quite an impressive accomplishment. The question many people ask is, is this desirable? Some people think it is, some think it’s not. Fewer seem to have tested these hypotheses in blind tests. That’s what I hope to help out with, even if by a little bit.
I think the results of my blind test surprised many, because though there is not a full consensus here, many people seem to just go with how Harman designs their speakers as the model of the ideal speaker. And by that standard, obviously the Sierra 2EX deviates quite a lot by having a drastically broader treble dispersion (more linear, too). Some say this makes it superior, others question that. Well, I figure only a blind test can truly answer this.
The results of my recent blind test is at least a hint that dispersion should be both smooth AND wide to achieve the best sounding speaker. Not proof by any means, but one intriguing bit of evidence pointing in that direction.
Next I am going to be comparing the Neumann KH120 to my Ascend Sierra 2EX. The Neumann’s spin measurements appear to extraordinarily good. One therefore would expect them to beat the Sierra 2EX. If not, this would be another strong data point in favor of the broad dispersion preference theory.
P.S. This time I will (probably tomorrow) post a pre-test thread asking for suggestions and advise for test methodology etc., to try to make my next blind test as scientifically valuable as possible.
I think it's mostly a preference thing when it comes to ideal dispersion.
Some people are also more sensitive to directivity issues. When I compared the Sierra 2 EX to the Salk Silk and Scansonic MB-1B, I felt the Sierra 2 EX had the worst driver matching between the RAAL tweeter and the 6" woofer, where the dispersion was excessively wide at the tweeter axis but the woofer's dispersion near the crossover frequency noticeably narrower. The matching was a somewhat better with the Salk Silk's 5" driver (smaller woofer with wider dispersion and larger ribbon tweeter with narrower dispersion), but ultimately the Scansonic sounded like it had the best directivity matching with the use of a waveguide around the ribbon tweeter and a small 4.5" woofer. If I didn't know I would think I was listening to a typical waveguided dome tweeter, that's how good the directivity match was with the Scansonic/Raidho tweeter + woofer.
When I compared the Sierra 2EX to the KEF R3, the KEF did not sound any more like a point source whatsoever, at least in this room. Both sound like point sources to me, until you get really close (much more so than you should for these kind of speakers).
Is there also a kind of consensus for off-axis, let's say 30° horizontal response of a speaker?
Out of curiosity, what does this directivity matching (or lack thereof) sound like, subjectively?
It often sounds like coherence; you are not aware of hearing the crossover.
It also tends to sound non-fatiguing (assuming there are no other problems), as there is no significant spectral discrepancy between the first-arrival sound and the reflections. When there is a significant discrepancy between the two, it can show up over time as listening fatigue.
In my opinion.