Waveguides and horns

[JoshK]

FWIW, I'd look at the 45 degree line vs the 0 degree line. That is the -6db line, which defines constant directivity. Again, its hard to be definitive with that graph and not expanding it a lot in paint or some other program. I would say it is not collapsing, which is why I think its ok with 900hz, just maybe not "ideal". That is my point. Plus if you looked at things like distortion graphs, impedance and phase you'd get a more complete picture.

Never-the-less, i think they are great waveguides and measure up. My own rep was at stake since I argued for a particular design, and am not bashing in any way. I just was arguing a particular interpretation.

[Brett]

Verbal shorthand I picked up from one of the engineers I first trained with; in this case, I meant that the directivity control is weakening.

[doug20]

JoshK and Brett. Im sorry but you are still wrong about any collapse above 1KHz. You definition of collapse is your own but hardly a scientific one, Remember that 15" should hold direcitivity below 1.2K in theory anyways. "Weakened" has no meaning in audio either to me, its either sounds fine or it does not. If its a compromise then its not good.

Many people always will believe what they want to believe so no reason to have a back and forth discussion. Here is the conclusion of this.

If you believe the SEOS-15 should be used to 1.2K then only use it to that point and I do understand why you would argue for the SEOS-18 is okay with 1" (just disagree because you limit the SEOS-18s potential). Many others will be building incredible waveguide designs with XO points below 1KHz and therefore the idea of the SEOS-18 only being a 1" exit is a waste and does not allow for the maximized advantage of the SEOS-18 which is directivity down closer to 700Hz.

I personal have never built a product or offer anything to customers that had a limitation taking away the products full potential but that is just me and how I have run my company.

[bwaslo]

Here's some distortion data - I put together a crossover crossing a DE-250 in an 18" rectangular waveguide (mostly conical, roughly OS with large mouth roundovers) and crossed over at 750Hz, which is just above where the inherent response starts to roll on the driver'/waveguide. The midwoofer is a TD15M. The 750Hz point just happened to be where the setup seemed to want to crossover based on the different depths of the drives (when I go active, I'll be able to deal with delays better). Also, the waveguide level looks (and sounds) like it needs to be reduced about 4dB or so.

The levels shown are actual SPL, obnoxiously loud, at 6ft from the cabinet.

The distortion is visible between 800 and 1800Hz (second harmonic) and 800 and 1200 (third harmonic), but really aren't bad in either case. There is something screwy around 380Hz, too, but I think that might be from furniture buzzing (a CD cabinet nearby wants to sing along at these levels).

I don't know whether this says anything about low-end crossover point for the 1" DE250 in a home setting, but it certainly doesn't look like a disaster at least.

[thorstenlarsen]

Hi Bwaslo !
Did you measure the SEOS-15 in a baffle or in free space ?
Thorsten

[bwaslo]

The SEOS-15 was measured sitting on the edge of a flat pillow which was sitting on a box, so there was some boundary below it, nothing to the sides.

[gedlee]

Bill

Could I get a set of the data so that I could do a polar map on it? I think that without a competent display of its polar response via a contour plot, its hard to say how well it performs. I will add these plots to the list of devices shown in my web app.

[bwaslo]

Sure, Earl.

Do you need a particular format? Right now they are FRD files, 500Hz to 20kHz, mapped onto 1,611 log spaced frequency points, no phase.

[gedlee]

I need the phase. Best is a set of impulse responses at a sampling rate of 44.1 kHz, but the sampling rate is not critical. I usually do 0, 5, 10, 15, 20, 30, 40, 50, 60, 80, 100, 120, 150, 180 degrees. That's because the resolution required is greatest where the rate of change of the polar response is the greatest and that, of course, is on axis. I can interpolate to these angles from any set of angles but thats just another set of calculations. I can also get by without the rear data, but again, the rear data helps to improve the forward data.

[bwaslo]

Sorry, I don't have the waveguides anymore, and didn't save impulse responses. I just had the horns a few days (along with a bunch of others) and just saved response patterns as described, 0 degrees to 75 degrees off-axis at 7.5 degree increments. Wish I had saved IRs but it was hot out where I was doing it, I didn't have a lot of time, and wasn't as thorough as I might have been. The responses should be minimum phase, though, so the phases (and IRs) should be derivable via Hilbert transform, right? Or do you expect some non-minimum phase from the edge reflections (or HOMs)?
I can remap the data to "FFT" format at 44.1k, though data below 500Hz will be absent (0) if that matters. The responses drop pretty well by 19kHz, so not having anything between 20kHz and 22.05kHz shouldn't be an issue, I don't think.