Originally Posted by Speedskater
I wonder what the "impulse response" would look like if the test mic use moved 10 or 20 degrees to the side and up and down.
10, 20 and 30 horizontal it still looks exceptionally good with perfectly maintained time coherence (that's why the tweeter is ofset in the first place!). On the vertical axis -as any non-coaxial design it will deteriorate when going exessively off-axis.
Ingvar Öhmans floorstanding 2(.5)-way speakers have been critisized (and routinely rediculed and dissmissed by self proclaimed "knowitalls" at many hifi-forums) for years and years for the offset tweeter and the severe diffraction effects this *must* cause, as well as for the huge directivity missmatch between the 1" tweeter and the 8" mid/bass drivers that simply *must* exist due to gemoetrical reasons, the small voicecoil etc. etc. etc. The list goes on.
Now, in Sweden, (where his "floorstanders" have been available for the last 15-20 years or so) they have been scrutinized, measured and put to the test in absurdum and they measure and perform pretty much imaculately.
All in all, sure there is diffraction happening (as with any edge) just no where nearly as much that one would first expect
Now, first thing to understand about this particular case is that due to the acoustically close proximity of the point source to the edge, the wavefront from the tweeter is is not straight -it is significantly curved, and at the distance of the edge, it is curved with a radious that is acoustically "small" compared to the wavelengths the tweeter emmit. This in turn will cause the continuum of diffraction reflections along the horizotal edge to "fire out of phase" with eachoter which will significantly reduce the diffraction effect within the listening window.
The other thing to realize is that with "normal" edge diffraction for a baffle with sharp edges, the soundwave travel from the speaker along the baffle (the edge of the wavefront is fully supported by the baffle) to the edge where the support abruptly disapear. The more abruptly the baffle transfer into full space (the sharper the edge), the more diffraction effect. A way to significantly lessen baffle edge diffraction is to angle or round the edges of the baffle.
Now, if you look at the tweeter again, the part of the wavefront is hitting the edge at an angle -it is not travelling along the horizontal part of the baffle to the edge, it is arriving at the edge at an angle, and hence the impedance change is much less abrupt than it might seem at first, reducing the diffraction even more. The specular reflections caused by the horisontal baffle parts are a much greater problem (potentially), if not properly taken care of.