Importance of vertical polar response

[gedlee]

How much of a deviation from a circular toward an ovoid cross section, in your opinion, would be tolerable before an ideal areal cross section from a HOM perspective becomes a 'diffraction horn', or at least suffers significant compromise from a HOM perspective?

Its not just the change in cross section, its also the change in slope. I said that before. How much? I have been able to measure very small changes and hence I go to extremes to avoid ANY discontinuities.

The Iron Lawbreaker, as I have previously described in this thread, is an example of a real world solution to this 'problem'. In fact, it takes a fair amount of concentration from any normal listening position to perceive a difference in azimuth for the LF vs HF source of its output since vertical and horizontal lobing and dispersion do not suffer from anomalies through the xover region.

This test should be done with measurements not with listening. If you can't measure the nulls then you certainly can't hear them.

 

[thoriated]

Its not just the change in cross section, its also the change in slope. I said that before. How much? I have been able to measure very small changes and hence I go to extremes to avoid ANY discontinuities.

Well, then, if you were to design an oval cross section horn, how close do you believe it would be possible to come to eliminating HOM by controlling the taper? None? A considerable amount? Completely?

This test should be done with measurements not with listening. If you can't measure the nulls then you certainly can't hear them.

Sorry if I wasn't very clear. I wasn't attempting to use a listening test as a proxy for a lobing measurement. For the Iron Lawbreaker, there is still a vertical offset between the driver centers that could potentially be audible under certain circumstances when the listener is close enough to the speaker even with ideal directional characteristics. That is what I was addressing relative to a concentric driver. However, LF room modes appear to be responsible for several times more vertical offset of the apparent vertical (and horizontal) source locations of bass frequencies than the driver to driver spacing in my listening environment, as I suspect it would be in most.

 

[gedlee]

Well, then, if you were to design an oval cross section horn, how close do you believe it would be possible to come to eliminating HOM by controlling the taper? None? A considerable amount? Completely?

Like nonlinearity HOM cannot be elliminated completely, but it should be possible to do very well. While my techniques for axi-symmetric waveguides are public domain and well known, my techniques for other shapes are not, in particular an elliptical waveguide. There are better and worse ways to do this, but you'll have to glean those for yourself.

 

[Wayne Parham]

I was refereing to coax Waveguides, not coaxial in general, I should have been more specific. Yes, a coaxial tweeter could be made to perform better than a seperate tweeter and woofer or midrange. It is the waveguide issue that I was talking about. In general there is no problem getting a piston tweeter close enough to the other driver such that the lobing is not an issue, but it IS a problem with horns and waveguides, so my comment was specfic to those.

Yes, I agree, and like you, I was only referring to coaxial speakers with horn loaded tweeters when I said, "The problem with coaxials having horn loaded tweeters firing through the center of the midwoofer's pole piece is the short horn is really a compromized stub, too small to get good pattern control."

Like nonlinearity HOM cannot be elliminated completely, but it should be possible to do very well. While my techniques for axi-symmetric waveguides are public domain and well known, my techniques for other shapes are not, in particular an elliptical waveguide. There are better and worse ways to do this, but you'll have to glean those for yourself.

I remember talking with you about a PS horn back in 2005, and I liked that approach. Such an oval device can be made with an asymmetrical pattern, say 90x60, not so different in the vertical and horizontal planes that you compromize your differences in slope. They'd be two different ones, to be sure, but both catenary and both blended. Sounded then like a great idea to me and it still does. The overall shape allows a baffle geometry that keeps the nulls nicely outside for a tall clean pattern, the vertical HF pattern is limited enough to help reduce ceiling slap and your flare shape curviture is right to help keep internal reflections down. Sounds like a good approach.

My question always has been, how much audible difference will this PS horn be from a rectangular horn having catenary curved flare on both axis (top/bottom, and side/side)? And I know from discussions with you and your many posts, that your question is, how much audible difference will this PS horn be from an OS horn of similar size?

The elliptical PS flare is a shape that one might hope to gain both sets of benefits, I suppose.