Hornresp

[David McBean]

Originally Posted by just a guy

I'm gonna take awhile to think about all this before asking if you can incorporate that right into the Export wizard thing.

Hi just a guy,

I can save you some thinking time - the approximation of a parabolic segment using multiple conical segments will not be included in Hornresp :-).

It is a relatively easy matter though, to set up a simple spreadsheet to calculate the required AkAbak inputs - see attached screenprint.

I can email you a copy of the .xls file if you like.

Kind regards,

David

[just a guy]

Originally Posted by David McBean

I can email you a copy of the .xls file if you like.

Kind regards,

David

Thanks, that would be great! I hope you didn't make that spreadsheet just for me though. It would make me feel bad to monopolize your time on stuff that doesn't directly involve Hornresp. But still, thanks!

[Don Hills]

I have some code that appears to work by similar means to David's spreadsheet.

1: Given Y = A * X squared, we have two pairs of X-Y values so we can find the value of A. Because we are working in areas, X squared = area / pi.

2: Parabolic horn curves have more "bend" near the throat and are "straighter" towards the mouth. A good approximation is to make each segment twice as long as the preceding segment as you proceed from throat to mouth. For a 4 segment horn, the segment lengths will be 1x, 2x, 4x, 8x the first segment length. So divide the horn length (distance from throat to mouth) by 1+2+4+8 = 15. The segment lengths will be 1/15, 2/15, 4/15, and 8/15 of the horn length.

3. Using the formula from before, calculate the area at each of the segment joins.

My script is written in an obscure scripting language (REXX). My plan is to find some free time (ha ha) to convert the script to a spreadsheet so it will be useful to others.

[David McBean]

Originally Posted by just a guy

Thanks, that would be great!

Hi just a guy,

It seems that files cannot be attached to emails generated from this forum. Could you please send me a standard email so that I can reply to you with a copy of the Excel file as an attachment. My email address is given on the Hornresp download website - simply click on my name on the site.

Kind regards,

David

[schmeet]

The equation for a Parabola is Y = a.X² + C

We have our two Areas: S1 and S2. From these we can calculate Y1 and Y2.

Y1 = SQRT[S1 / PI]
Y2 = SQRT[S2 / PI]

if we assume that the element starts at 0 and ends at L12 then:

X1 = 0
X2 = L12 (Length)

a and c can be calculated from:

a = [Y2 - Y1] / [X2² - X1²]
c = Y1 - aX1²

We can then calculate the area at any point along the segment by

Y = a.X² + C
S = PI.Y²

[just a guy]

Originally Posted by David McBean

Hi just a guy,

It seems that files cannot be attached to emails generated from this forum. Could you please send me a standard email so that I can reply to you with a copy of the Excel file as an attachment. My email address is given on the Hornresp download website - simply click on my name on the site.

Kind regards,

David

No problem, thanks.

[mwmkravchenko]

As a followup on the update notification:

Works very well.

Still have to figure out what the editor will do.

But very interesting.

[just a guy]

Originally Posted by schmeet

We can then calculate the area at any point along the segment by...

I realize you were talking about something else but this part I've quoted seems like it might potentially lead to something exciting.

Akabak can't give a detailed list of csa vs length like hornresp can. (If it can I don't know how to do it.) Since Akabak won't do PAR segments, there's no real accurate way for a guy like me (that can barely understand the equations you guys are showing) to lay out an Akabak horn accurately, all you can really do is replace the virtual CON sections with straight boards that create PAR segments in the fold plan. It's been historically proven that this isn't a fatal flaw but it's obviously not the most accurate way.

BUT if you can do what I've quoted, but do it with CON sections instead, I could make a spreadsheet that converts an Akabak script (with it's CON segments) to a detailed list of csa vs length like Hornresp does with it's Export wizard.

(This mess created by Akabak's non PAR attitude is the main reason I was hoping for more Hornresp segments. But if I can find an accurate way to lay out Akabak's CON sections that's almost as good.)

BTW, since this isn't about Hornresp specifically I can start a new thread for this if anyone would prefer that.

[David McBean]

Originally Posted by just a guy

No problem, thanks.

Hi just a guy,

You've got mail.

Kind regards,

David

[David McBean]

Originally Posted by just a guy

(This mess created by Akabak's non PAR attitude is the main reason I was hoping for more Hornresp segments. But if I can find an accurate way to lay out Akabak's CON sections that's almost as good.)

Hi just a guy,

Further to my earlier post about keeping things in perspective, I think you might find in practice that it really doesn't make much of a difference if multiple parabolic segments or multiple conical segments are used in the simulation model. To illustrate, the attached screenprint compares the SPL response of a tapped horn having three conical segments against the response of the same horn, but with three parabolic segments. You can see that the traces are effectively identical. Greater inaccuracies are already inherent in the predictions due to the assumption that the driver diaphragm acts as a rigid plane piston at all frequencies, and that there are no losses due to panel vibrations, etc.

Kind regards,

David