Hello everyone, I am trying to run simulations in Vituix, with the aim of creating a 3-way speaker with a cardioid pattern. I started by tracing the speakers from the manufacturers' websites, then applied the low-frequency loading and baffle diffraction using the various tools that Vituix offers. My goal is to replicate the speaker to see if I can achieve reliable simulations with what I have designed and measured.
I will post a couple of graphs to share the simulation results and perhaps get suggestions for improving the whole setup.
The speaker will be a stand-mounted design, with dimensions of approximately 40/45cm in height, 35cm in depth, and 20cm in width. The tweeter and midwoofer will be placed on the front panel axis, while the two subwoofers will be positioned on the two side panels.
The crossover will be active, and the goal of the project, in addition to evaluating the quality of the simulation, will be to attempt to simulate an active cardioid system using the two side subwoofers.
Many thanks
Stefano
I will post a couple of graphs to share the simulation results and perhaps get suggestions for improving the whole setup.
The speaker will be a stand-mounted design, with dimensions of approximately 40/45cm in height, 35cm in depth, and 20cm in width. The tweeter and midwoofer will be placed on the front panel axis, while the two subwoofers will be positioned on the two side panels.
The crossover will be active, and the goal of the project, in addition to evaluating the quality of the simulation, will be to attempt to simulate an active cardioid system using the two side subwoofers.
Many thanks
Stefano
Ciao Stefano,
you might get more responses if you clarify how you achieve cardioid response. Two side subwoofers will probably not be enough for an active cardioid setup.
Did you get polar responses from the manufacturers?
The directivity index looks exceptionally nice ... almost too good to be true
😊
you might get more responses if you clarify how you achieve cardioid response. Two side subwoofers will probably not be enough for an active cardioid setup.
Did you get polar responses from the manufacturers?
The directivity index looks exceptionally nice ... almost too good to be true
😊
I've done similar simulations, starting in Vituix with diffraction models and moving to ABEC for higher accuracy. I can tell you the approach has promise. After all, its what Kii does. I also use (simulate) two 8" side woofers out of phase with a front woofer to achieve cardioid but I use a 8" front woofer so I can extend the cardioid effect lower. I've been using FIR DSP in simulation, which makes it easy to achieve the desired response. Kudos for getting there with IIR. That is on my to do list...
Roundovers on the front baffle make a big difference in how much cardioid directivity you can achieve. The baffle diffraction apparently changes the pattern enough to have this effect.
If you show a line charts instead of polar maps at the bottom of the Vituix screen, its easy to see how deep the attenuation to the sides and rear is. From what you have shown, I don't believe you have optimized the delay and amplitude of the side woofers. You should be able to get the transition from red to yellow in the polar map inside the +/-90 degree lines. With active cardioid and DSP its possible to get 20 db attenuation to the rear. I would progressively dial in more (or less) delay on the sidewoofers while looking at the bottom lines on the horizontal line chart. This where FIR DSP really helps - allows you to tune the phase vs frequency which is what you need to do to get those last few dbs of attenuation to the rear. You may be able to do this with IIR all pass filters but its not so straightforward.
Roundovers on the front baffle make a big difference in how much cardioid directivity you can achieve. The baffle diffraction apparently changes the pattern enough to have this effect.
If you show a line charts instead of polar maps at the bottom of the Vituix screen, its easy to see how deep the attenuation to the sides and rear is. From what you have shown, I don't believe you have optimized the delay and amplitude of the side woofers. You should be able to get the transition from red to yellow in the polar map inside the +/-90 degree lines. With active cardioid and DSP its possible to get 20 db attenuation to the rear. I would progressively dial in more (or less) delay on the sidewoofers while looking at the bottom lines on the horizontal line chart. This where FIR DSP really helps - allows you to tune the phase vs frequency which is what you need to do to get those last few dbs of attenuation to the rear. You may be able to do this with IIR all pass filters but its not so straightforward.
Hello, my goal is to fully understand the capabilities of Vituix for moderately complex projects and, once I grasp how it works, try to apply everything to a real project.
I have no idea how to simulate FIR filters in Vituix, but I would still like to try, just as I would also like to learn ABEC.
In any case, I have tried to iterate the simulation following your valuable advice. Specifically, I had forgotten to adjust the delay of the two subs.
I'm posting some graphs again to show the progress.
Thank you very much for the advice!!!!!
I have no idea how to simulate FIR filters in Vituix, but I would still like to try, just as I would also like to learn ABEC.
In any case, I have tried to iterate the simulation following your valuable advice. Specifically, I had forgotten to adjust the delay of the two subs.
I'm posting some graphs again to show the progress.
Thank you very much for the advice!!!!!
there are (at least) two ways to do a DSP auto-eq in Vituix. One is to use the optimizer, which will configure a G(f) block to equalize a driver flat over a range with chosen high pass and low pass filter slopes at the ends of a range. i.e to a target crossover. This is documented in the Vituix help file. Another way to do is to put a G(f) block in your schematic flanked by high pass and low pass filter blocks but then, instead of using the optimizer, export the FR of that filter chain and then enter that FR into the Vituix calculator. In the calculator tool, select "mirror", adjust the offset to set the filter gain, and then press c"alculate and save". It then exports name.mir, where name is what you called the original FR export. Click on the G(f) block to configure it and select name.mir as its transfer function (bottom of main screen). when you are finally satisfied with the equalization results, go to the top left of the main screen and select "impulse response" from the view drop down menu. This will show you the impulse and step response and allow you to export the impulse response which then can be loaded into a DSP engine; but if you are just simulating you don't have to do that....
Select a PEQ block then click the "shape" options. You are using the default "Param EQ", you will see there is also a "phase EQ" choice. Phase EQs are useful for optimizing cardioid response over a narrow band. If you use one, you have defined a FIR and need DSP to implement it. Any filter chain can be exported as impulse chain as I described above for loading into a DSP engine.
Just noticed you haven't positioned your side woofers correctly. Left side should have x=-width/w and R=-90degrees. Right side should have x=+width/2 and R=+90.
Select a PEQ block then click the "shape" options. You are using the default "Param EQ", you will see there is also a "phase EQ" choice. Phase EQs are useful for optimizing cardioid response over a narrow band. If you use one, you have defined a FIR and need DSP to implement it. Any filter chain can be exported as impulse chain as I described above for loading into a DSP engine.
Just noticed you haven't positioned your side woofers correctly. Left side should have x=-width/w and R=-90degrees. Right side should have x=+width/2 and R=+90.
Actually, I hadn't included the translation along the X-axis, while I had added the rotations at + and -90° (even though Vituix doesn't display them in the crossover schematic).
I fixed the issue, but it didn't change much.
What I can't quite understand is how to limit the rise in the woofer's emission at ±180° around 400 Hz.
Many thanks about DSP auto-eq!!!!!!
I fixed the issue, but it didn't change much.
What I can't quite understand is how to limit the rise in the woofer's emission at ±180° around 400 Hz.
Many thanks about DSP auto-eq!!!!!!
those are serious side lobes around 400 Hz. are they still there if you mute the side woofers? I expect so. I see you have PEQs in that region in both front and side woofer filter chains. In your H line chart I see you only have about 8 db rejection to the rear there. So the phase in that range needs to be changed. I would stick a phase EQ block in the side woofer chain and try moving the phase in the direction that increases the rejection to the rear - the amplitude of the red traces in the H line chart. Play with both the phase in degrees and the Q.. Before that, make sure that your delay and amplitude on the side woofer amplifier block are optimal.