Help interpret graphs?

[markusA]

I'm trying to simulate a speaker with a jx92s driver using Hornresp and MJK's worksheets.

Before I can really make anything out from the sims I need to know what I'm looking at.

I'll start with uploading 4 images, two from each program.

How would you expect this speaker to sound?

I'm not saying this is the final draft or anything like it.
It's a first stage experiment while learning the software.

The graphs look very uneven and scary, but how would one expect it to sound?
#1 = Hornresp terminus
#2 = Hornresp combined
#3 = MJK SPL plot
#4 = MKJ Combied SPL + diffraction (baffel step)

[freddi]

they sure don't agree - I'll assume Hornresp can't model the effects of damping material (?) "If" it acted like the Hornresp model then the "hit" region would be emphasized - perhaps too much (maybe not?) and would have pretty good bass punch. Your MJK predictions look pretty smooth. There should be enough real measurements of MJK modeled projects to give a feel. I made a 15" transflex type and Hornresp was pretty accurate for that case--but it had no stuffing along the line.

[rjbond3rd]

Hmm, well to me, once you get the scales the same, they seem to be at least largely similar, but freddi, you know much more than I do.

1. Which MJK worksheet is that from? TL Sections or one of the BLH sheets?

2. MJK models room factors, not just walls / corners but also floor bounce, etc.

3. I can't see how HornResp would deal with baffle step as MJK is doing there, since it deals with areas/lengths (not widths as far as I can tell).

I am a newb so don't put any stock in what I've said. Like you, I'm highly interested and curious!

[rjbond3rd]

Hi markusG, oops, more thoughts:

4. What is HornResp using for Ang -- .5 (corner)?

5. Path length?

6. Is this a BVR or BLH? As I'm sure you know, BVR = big chamber like a BR, short path (maybe 10 liters chamber / <1m length) whereas BLH would be more like small chamber / long path (maybe 2 liters chamber / > 1.4m length).

7. You ask "how would this speaker sound" and I believe you can get another 10Hz, maybe 20Hz on the bottom.

[freddi]

maybe they agree more than I thought

[rjbond3rd]

Hi freddi,

Well I guess it's the differences that are most interesting You've measured lots and lots, whereas I'm just coming up to speed (if that).

markusG, in my extremely limited experience, when you model the horn by itself, you can achieve a "plateau" where there are no ripples. And then on the right side of that plateau, you get your first "bump" and then you get more ripple as the high frequency response drops.

The sides of the plateau are obviously related to the length. For example, if your length is about .7 meters (assuming a BVR), the right side of the plateau may be at ~120Hz (half wavelength) and the left side may be at ~60Hz (quarter wavelength).

[CLS]

Back engineer some commercial BLHs (or any one you can get the plan), and plug the numbers into hornrep, you'll find out many of them have similar (ragged) responses as markusG had on the above.

Too small of the mouth (terminus) area, improper proportion of the cross section areas, wrong expansions, unmatched back chambers.... etc. are all very easily getting ragged responses. Certain efforts (or experiences) are needed to get that nice looking 'plateau', even in the sims. When making a real thing, epecially a commercial product, all kinds of compromises cut things here and there....

Hi markusG, can we see the input screen (of hornresp)?

[markusA]

freddi> I think you are correct about hornresp, it's more limited. Chambersize, diffraction and stuffing isn't factored in afaik?

rjbond3rd> I'm using the BLH sheet with sections. The lengh of the horn is a little over 100".

CLS> Picture of the input screen comming right up...

My objectiv is to come up with a FR BLH that doesn't need bsc. I'm modelling a VERY narrow cabinet with a height of roughly 3'-4'. I'm aiming for 4' with the driver offset a little heightwise, maybe 3'2" off the ground or something along those lines?

Looking at "edge" (the software) you get a smooth baffel step but MJK uses a sinusoidal **. The peaks and dips in the graphs, are they to be ignored or will they make the speakers sound odd? Do I assume a rough mean value to be the right one or will the SPL fluctuate lke it does in the graphs? How does that sound?

[markusA]

Here's the input screen for Hornresp. I tried to make it roughly the same as what I did in MJK's worksheet.

[hm]

Hello,

may be my experience with Jx92 will help,
look the measurements, feedback (bottom)
and the plan,
i simulate with AJ horn and never match the real
measurements 1to 1 it is only approxmative,
which flare rate do you take in your simulation?

tubaFB
tubaplan