Difficult or not, to not consider the room or, worse yet, to not attempt to make it as good as the speakers, is always going to be sub-optimal or even disasterous. Like it or not, the room is here to stay and you either learn to deal with it or forever hope for a better "system".
The baffle obsevations are purely subjective based hypothesis. If the polar response IS "very" baffle dependent then its a bad design. A good design should not really know that there is an enclosure.
That's partly why dipoles sound good. The radiation pattern is a figure-8 at almost all frequencies, whether by the front and back waves cancelling or by the driver's inherent directionality, so the power response in the room is close to the on-axis response. This avoids the usual monopole problem at the midrange to tweeter crossover point, where mid dispersion shrinks just before the tweeter comes in with wide dispersion, resulting in room power response not matching the on-axis response. Note that Linkwitz says the Orion's sound was improved by adding a tweeter to the back, which adds weight to the room field argument.
Dr Geddes' horns avoid that issue even though they're monopoles by showing even power response through their relatively narrower beamwidth, but the result is the same: the reverberant field has the same response as the direct field.
Of course, Dr Toole covered the importance of room response decades ago. As tg3 says, his white papers are highly recommended; Toole is a major reason Canada has a damned good speaker industry these days.
Dr Geddes' horns avoid that issue even though they're monopoles by showing even power response through their relatively narrower beamwidth, but the result is the same: the reverberant field has the same response as the direct field.
Of course, Dr Toole covered the importance of room response decades ago. As tg3 says, his white papers are highly recommended; Toole is a major reason Canada has a damned good speaker industry these days.
With much respect, I beg to differ. A direct radiator in an enclosure will radiate into a half space (modulo driver directivity) when wavelengths are small compared with baffle size, and into a a full space (four pi) when wavelengths are large relative to the baffle.
Yes I think this is the problem ! But not so easy to have a good polar response.
In addition baffle step on box, back waves cancellation on dipole, I always have diffraction problems in the area 800-2000Hz. If I want a good dispersion in this area I should use small drivers < 5". To minimize this diffraction problem I should offset the driver or reduce the width of the panel. I think there is a relationship between the driver and the geometry. A good design cannot ignore the geometry if you want it disappear, don't have bad incidence on the response.
Some people make a lot of test baffle before to find the best geometry. The baffle diffraction software are not very precise to avoid this work. This is why I have talked about good models
Quite the contrary. The baffle diffraction software I've used is very good at not only the baffle, but also the drivers with both their location and directionality. I avoid making any test baffles entirely by using it to select the desired initial raw baffle response, then minimize the impact of diffraction (both that of the baffle and adjacent drivers) with, in my case, judicious applications of felt. It puzzles me why it's considered to be either difficult or imprecise when reliable and sufficiently accurate software has been available for years. It's also free, by the way.Some people make a lot of test baffle before to find the best geometry. The baffle diffraction software are not very precise to avoid this work. This is why I have talked about good models
Dave
With much respect in retrun, do you think that I am unaware of that? Do you think that it makes any difference in what I said?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.