A Bookshelf Multi-Way Point-Source Horn

Disabled Account
Joined 2012
After many years of various speakers and systems... up to tri-amp'ed with dual transmission line 18 inch bass and QUAD ESL mid and RTR ESL treble. [Tweeter panel was directly driven from push-pull tube plates thru coupling cap etal.] Super tweek. --------- I still found music some what compressed sounding... lacking in dynamic range. Compared to Live sound. Until I got a pair of JBL M2 loudspeakers with matching dsp x-over and iT5000HD power amps. Wow! Now there is NO compression effects from the speakers. Guess compression was in the drivers and not so much in the CD/music production. Now, I listen at 75-85dB at the listening position and peaks are up there. Easily 95-100dB spl.

Lack of dynamic range and compression is very common and should get a lot more attention in choice of speakers/amps.


DSC02434.JPG







THx-RNMarsh
 
Last edited:
How would one simulate different back chamber volumes for the midrange and woofers using hornresp - if that is possible at all?
- or is it assumed that they share the same back chamber volume?

Maybe I need to play with it a bit.

Yes you do need to play with it. If you do, you will see one hornresp record for the driver at the apex of the horn, (Nd record) with its own chamber, and separate hornresp records for the mids (ME1 record) and yet another for the woofers, (ME2) if any, each with their own chambers and ports through the horn walls. You can activate/de-activate mid and woofer "records" while simulating the driver at the apex (Nd record) to see how they combine. You get to define XO and PEQ filters and delays....
 
How would one simulate different back chamber volumes for the midrange and woofers using hornresp - if that is possible at all?
- or is it assumed that they share the same back chamber volume?

Maybe I need to play with it a bit.
I just shown on example result !
Hornresp can do it using multiple saved drivers. I my exemple, i got :
_12" in closed 30l as back chamber, 12L front chamber, 20cm long 50cm² front port
_8" with 10L closed back, 0,4L front chamber, 5cm port of 18cm²
_a 4" fullrange driver with 0,6L back chamber, and a 50cm²-1600cm² horn of 35cm lenght.
The 8" entry is at 16,5cm from the 4".
The 12" is put at horn end.
Got two 1rd electric high pass, and 2 1rd low pass.
Got -2,1ms delay for the 12".
All of this is computed id hornresp, and sims as shown.
:)
 
Here some of my input screens of only one sim. Can only upload 10 image at time :)

Look at right panel parameters input, and all the stuff that can be shown using selectors (excursion1 2 3/impedance1 2 3/power 1 2 3 total.....)
 

Attachments

  • 10.PNG
    10.PNG
    24.3 KB · Views: 115
  • 9.PNG
    9.PNG
    24.1 KB · Views: 109
  • 8.PNG
    8.PNG
    22.3 KB · Views: 111
  • 7.PNG
    7.PNG
    22.9 KB · Views: 120
  • 6.PNG
    6.PNG
    24.8 KB · Views: 168
  • 5.PNG
    5.PNG
    16.9 KB · Views: 471
  • 4.png
    4.png
    21.2 KB · Views: 467
  • 3.PNG
    3.PNG
    19.2 KB · Views: 455
  • 2.PNG
    2.PNG
    14.9 KB · Views: 468
  • 1.PNG
    1.PNG
    20.4 KB · Views: 475
Last edited:
And some others input screen the same synergy sim :p

That's a LOT of options, and a LOT of parameters, but still easier, faster, and lot more user-friendly than the same sim done in akabak code...
 

Attachments

  • 16.PNG
    16.PNG
    25.8 KB · Views: 109
  • 14.png
    14.png
    19.1 KB · Views: 102
  • 15.PNG
    15.PNG
    25.9 KB · Views: 105
  • 13.png
    13.png
    17.8 KB · Views: 106
  • 12.PNG
    12.PNG
    23.2 KB · Views: 109
  • 11.PNG
    11.PNG
    22.1 KB · Views: 118
After many years of various speakers and systems... up to tri-amp'ed with dual transmission line 18 inch bass and QUAD ESL mid and RTR ESL treble. [Tweeter panel was directly driven from push-pull tube plates thru coupling cap etal.] Super tweek. --------- I still found music some what compressed sounding... lacking in dynamic range. Compared to Live sound. Until I got a pair of JBL M2 loudspeakers with matching dsp x-over and iT5000HD power amps. Wow! Now there is NO compression effects from the speakers. Guess compression was in the drivers and not so much in the CD/music production. Now, I listen at 75-85dB at the listening position and peaks are up there. Easily 95-100dB spl.

Lack of dynamic range and compression is very common and should get a lot more attention in choice of speakers/amps.
THx-RNMarsh
Exactly why i simulate for 108-112db at xmax ^^.
That's one of the Bjorno awesome frequent advices for low freq SQ cab.

Just to add :
Even with all his new tool, time alignement tweaking it really hard with hornresp without seeing ringing to decide what spike of a way to align with wich one of another way. We're approching real life here, and far from standard crossovers theories. Here a example, using akabak (still the same project). And from this experience, with 3 way, seeing phase is not enough to align considering the crossover and inductance delaying effect. The best i managed to sum a the same time response of the 12" woofer, and response of the fullrange to the one of the 8" filler was making starting the 12" 2,1ms before the two others ways. So got at the same time good phase track in low between 12" and 8", and not too bad between the 8" and the fullrange in low-mid range.
Having played like this have given me lot of experience that will be really useful for real-life project.

EDIT : Sorry for the off-topic ! Now, 00:57 at clock here and wake up at 6:20 for child and work tomorrow :sleep:
 

Attachments

  • akabak.PNG
    akabak.PNG
    144.6 KB · Views: 103
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
Little humble Akabak VS hornresp sims with my 3ways multiple entry horn project (using huge fixed delay, and 1rd crossovers, with kind of filler driver woofer) :

1) with some watts added in hornresp sims to show curves differences easier (easy to see/evaluate in low freq) : power response (not acoustic/on axis sim) using fix delays + 1rd crossovers. Filler driver in green, and fullrange in red, total in black from akabak sim. Only show here total hornresp power resp in red.

2) 0° freq response (horn simplified to 1 segment in hornresp) at exact same watt power, 1st order high pass. In low range, non coherent black line is just hornresp sim without filtering. Don't pay attention to it.

My conclusions : more high with akabak/less mid and lot more difference in directivity => i remember having readen David explaning that the model he use in hornresp for directivity is better in high freq.
Real result measurement in few days...

Thanks for doing a comparison. It will be interesting to see how close they model the real thing. Akabak has been pretty good for modeling my speakers. I like the interactive ability on HR though.
 
Or a bit cheaper, simply find a step by step tutorial, download and install a oracle virtualbox machine (seems there's a compil for mac http://www.oracle.com/technetwork/server-storage/virtualbox/downloads/index.html), download a windows XP image, install it, install and open akabak and hornresp on the virtual winXP windows. 1 hour of work... Here I got a french tutorial link (with even a needed windows xp image link) : http://microzoom.fr/tuto-virtualbox-windows-xp
I did it on a windows 10 machine 64bits, to get akabak working. There's no reason hornresp would'nt work too on this virtual machine.

Before that i used akabak with wine virtual machine on a old linux computer. Same idea. Finding good tutorial is the harder thing ^^
 
Last edited:
I couldn't get the Mac version of LT Spice to work on my Macbook Pro. Instead I used Wineskin Keeper to install the Windows version of LT Spice as a native mac app (double-clickable). Wineskin Keeper is a WINE wrapper program that turns Windows executable files into double-clickable mac apps. It works on about 80% of the Windows versions of audio design programs I've tried. A small number of programs don't work, most notably the TubeCAD ones. The best part is that its free, doesn't take up a huge amount of disk space, and if it doesn't work, you just trash the files and move on.