On Sun, Mar 20, 2011 at 11:21:41AM +0100, Jörn Nettingsmeier wrote: > my question was if this affects the ratio of W vs. the rest, and if > circular harmonics might be a more suitable model for decoder design in > this case.
The answer to that IMHO is no, for the reasons stated earlier. > but part of their logic is broken: you don't need a non-linear medium > for waves to "combine" for all practical purposes, Depends on what you call 'combine'. If the result of driving all elements is different from the sum of their individual results, that is non-linear by definition. > and the superposition > principle does not invalidate the concept of "coupling". What is this 'coupling' ? You mean one driver's response being modified by the presence of others ? This is again a purely linear effect (if not you have other problems). It has to be taken into account when finding the indivual responses of each unit. Again nothing magical, just the correct way to do superposition. > but the huygens principle holds, and of course a line of point sources > can be viewed as coupling to form a linear source, or vice versa. Up to some frequency (determined by the spacing of the point sources) that is a valid approximation. It's just superposition, nothing else. Again what is this magical 'coupling' ? > they contradict their own argument by stating further down that line > array elements should not be used as singletons. it is of course safe to > assume that a single line array element does not create a "slice of > cylindrical wave", but some vendors have waveguiding tricks up their > sleeves that do just that, although single elements still fail because > you are more or less always on the boundary. but it is possible to use > horn reflections in such a way that you will get a hf sound field that > would appear to be made of several hf tweeters, spaced more closely than > they actually are. a hack that does not automatically a perfect > cylindrical wave make, but a damn clever one, and one that moves the > upper coupling frequency a lot higher than the spacing of the tweeters > would imply. Yes, such techniques do exists. Assume for a moment that they work as claimed, and that the complete 3.5 meters of stacked drivers do indeed behave as real line source and produce cylindrical waves in the listening area. It would be unusable: extremely bad FR, even on axis and dependent on distance. And even 10 degrees of-axis it gets much worse again - you'd have a real pencil beam at HF. It's OK-ish exaclty on-axis and at a distance of 40 m or so, but no amount of EQ or whatever is going to make it usable at closer range. To be usable, you'd have to 'splay' the array (and put it on a tower). This will provide a uniform response if done correctly, and at close range it will still drop off less than 6 dB when distance is doubled. But not because of any cylindrical waves or any 2-D effect, but just because such a splayed array emulates a more distant point source. It still needs EQ, close to +3dB/oct. In practice a large part of this is taken care of by making individual drivers more directional as frequency goes up. Which does indeed mean they wouldn't make very good speakers when used individually. Ciao, -- FA _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound
