baffle reflection

[DBMandrake]

and it makes a convenient handle. But why is off axis response so important? You are usually supposed to control the reflections in a room anyway so the issue is not only about how the speaker radiates the sound.

Depends on who you ask.

Personally I don't think the very wide off axis response is all that important - I don't really care whether my speakers have significant response dips 60 degrees off axis at high frequencies for example. Most full size speakers do unless extreme measures have been taken in the design to avoid it, but those extreme measures usually result in significant compromises in other areas of performance. (IMHO)

There's a lot of perceptual research that shows that the listener axis response dominates our perception at high frequencies, and although overall power response at high frequencies does have some effect on the perceived "liveness" of the room by affecting the direct to reflected ratio, it's not critical in the same way that the on axis response is, again IMHO.

(There's no one "right" power response at high frequencies, while you could argue that there probably is one correct on axis response)

However the near off axis response I think is very important, as is a smooth polar response out to about 20-30 degrees horizontally and perhaps 10-15 degrees vertically. Consider this the "listener window" - the range of angles that listeners are likely to be in for critical rather than casual listening.

You want your speakers maximum (and flat) response to be on axis, with any variations off axis out to the above angles being a gradual, smooth monotonic fall off in response - eg no ripples in the polar response, or off axis having greater output than on axis etc.

This is particularly important above about 2Khz where stereo localisation comes entirely from amplitude changes. Anything that causes significant amplitude fluctuations with relatively small angular shifts on the horizontal axis in particular is going to cause instability of the stereo image with small listener movements, and unfortunately that's precisely what diffraction does at treble frequencies - relatively rapid shifts in comb filtering over small angular ranges.

(There are some that also believe that the comb filtering caused by treble diffraction can add harshness to the sound due to the peaky response that it can cause - I'm not fully convinced one way or the other and haven't done enough testing to form a solid opinion, but I think it has some merit)

With symmetrical driver placement the largest response shift occurs between directly on axis and a few degrees off axis. Offsetting the driver on the baffle may reduce this effect but so long as the diffraction is present there will still be undesirable ripples within the polar response in the listener window.

Far better to eliminate the diffraction as much as possible - by using waveguides, curved baffles, baffles with large radius edges, absorption etc.

So yes, off axis response is important but I think good behaviour across the "listener window" is more important than trying to achieve extremely wide dispersions free of power response "holes". Many people would not agree with that.

 

[Professor smith]

(There's no one "right" power response at high frequencies, while you could argue that there probably is one correct on axis response)

There was a very long thread on here about the correct power response. I didnt bother to read it all. So what was the conclusion?

However the near off axis response I think is very important, as is a smooth polar response out to about 20-30 degrees horizontally and perhaps 10-15 degrees vertically. Consider this the "listener window" - the range of angles that listeners are likely to be in for critical rather than casual listening.

Is this so that you can move around and have a wide sweet pot?