Software tools for building better DIY speakers

[Eetu]

@napilopez's excellent guide on taking full range quasi-anechoic speaker measurements using REW, Baffle Diffraction and Boundary Simulator and creating a spinorama in VituixCAD.

 

[Rick Sykora]

Hornresp? There's a definite learning curve with UI & workflow, so start with the wizards & docs! Once learned, this seems to be the tool used by the most experienced sub designers at DIYA, Data-bass, etc. Many programs sim basic sealed/BR/PR boxes. Hornresp goes the next few steps, adding support for varying expansion ratios, tapped horns...

Will check it out thanks. Really sparse website, but gather it is freeware.

 

[Wolf]

If you're working on passive xovers, Jeff's PCD/Blender/ResponseModeler/Diffraction and BoundarySimulator suite does quite a bit. I would also suggest Xsim by Bill Waslo.

 

[Rick Sykora]

If you're working on passive xovers, Jeff's PCD/Blender/ResponseModeler/Diffraction and BoundarySimulator suite does quite a bit. I would also suggest Xsim by Bill Waslo.

Thanks! Tried both PCD and WinPCD but not sure how they compare to VituixCAD. Have run into reliability issues in the PCD software and has kept me from giving them much of a test drive.

Have been using XSim and do like it for crossover mods and prototyping. if VituixCAD had some canned passive filters, not sure I would need XSim much though.

 

[HammerSandwich]

Have been using XSim and do like it for crossover mods and prototyping.

Xsim's one-axis only, right?

...if VituixCAD had some canned passive filters, not sure I would need XSim much though.

On crossover page, look above the driver icon to find "Lib." That's the library of canned filters, both active & passive.

 

[kimmosto]

Thanks! Tried both PCD and WinPCD but not sure how they compare to VituixCAD.

Features of VituixCAD are listed https://kimmosaunisto.net/Software/Software.html. I guess it's few times more than PCD and other Excel tools, WinPCD and XSim together. Main difference is simulation with quasi full space (dual plane) acoustical data vs. single axis which was dated decades ago.

 

[Phorize]

Will check it out thanks. Really sparse website, but gather it is freeware.

I believe that there’s quite a a lot of experience in using hornresp over at diyaudio.com and the developer still posts there:

https://www.diyaudio.com/forums/subwoofers/119854-hornresp.html

 

[Rick Sykora]

Features of VituixCAD are listed https://kimmosaunisto.net/Software/Software.html. I guess it's few times more than PCD and other Excel tools, WinPCD and XSim together. Main difference is simulation with quasi full space (dual plane) acoustical data vs. single axis which was dated decades ago.

Thanks for your efforts! It is an impressive and extensive tool...

The addition of a speaker database puts it beyond most other speaker design tools for me. Am preparing to direct Amir on how to bring the Klippel measurements for Directiva into VituixCAD. Is this simply using the import capability or would you suggest something else?

 

[kimmosto]

VituixCAD reads frequency responses exported from Klippel NFS in VACS balloon format. For example this is adequate dual plane response set: Phi0Theta10.txt
Phi0Theta100.txt
Phi0Theta110.txt
Phi0Theta120.txt
Phi0Theta130.txt
Phi0Theta140.txt
Phi0Theta150.txt
Phi0Theta160.txt
Phi0Theta170.txt
Phi0Theta190.txt
Phi0Theta20.txt
Phi0Theta200.txt
Phi0Theta210.txt
Phi0Theta220.txt
Phi0Theta230.txt
Phi0Theta240.txt
Phi0Theta250.txt
Phi0Theta260.txt
Phi0Theta270.txt
Phi0Theta280.txt
Phi0Theta290.txt
Phi0Theta30.txt
Phi0Theta300.txt
Phi0Theta310.txt
Phi0Theta320.txt
Phi0Theta330.txt
Phi0Theta340.txt
Phi0Theta350.txt
Phi0Theta40.txt
Phi0Theta50.txt
Phi0Theta60.txt
Phi0Theta70.txt
Phi0Theta80.txt
Phi0Theta90.txt
Phi90Theta0.txt
Phi90Theta10.txt
Phi90Theta100.txt
Phi90Theta110.txt
Phi90Theta120.txt
Phi90Theta130.txt
Phi90Theta140.txt
Phi90Theta150.txt
Phi90Theta160.txt
Phi90Theta170.txt
Phi90Theta180.txt
Phi90Theta190.txt
Phi90Theta20.txt
Phi90Theta200.txt
Phi90Theta210.txt
Phi90Theta220.txt
Phi90Theta230.txt
Phi90Theta240.txt
Phi90Theta250.txt
Phi90Theta260.txt
Phi90Theta270.txt
Phi90Theta280.txt
Phi90Theta290.txt
Phi90Theta30.txt
Phi90Theta300.txt
Phi90Theta310.txt
Phi90Theta320.txt
Phi90Theta330.txt
Phi90Theta340.txt
Phi90Theta350.txt
Phi90Theta40.txt
Phi90Theta50.txt
Phi90Theta60.txt
Phi90Theta70.txt
Phi90Theta80.txt
Phi90Theta90.txt

Vertical 180 deg and both 360 deg responses are already included in horizontal plane (Phi90 with NFS) so they will be skipped in the simulation anyway.

I don't have access to NFS so I cannot give any promises that timing i.e. phase responses are locked to certain (physical) position while measurements to different off-axis angles. That is required to get correct sum to off-axis angles while XO simulation.
Another possible problem is 1/48 oct. frequency axis which cannot cope e.g. group delay calculation to high off-axis angles due to excessive phase wrapping if sound flying time from DUT to mic varies too much. Different drivers should rotate around their own center point on baffle surface while off-axis sequence also with NFS if construction is large. Tiny 2-ways could have single rotation center for whole construction.

So measurements according my instructions could be better than NFS if previous conditions are not met. May need large space or much damping with sheep's wool to reach low frequencies as actual far field.

 

[kimmosto]

^I had to edit previous response list. NFS has quite unique planes: Phi=0/180 is vertical and Phi=90/270 is horizontal. User manual of VituixCAD says:
"Also 3D balloon data with CLIO, EASE and VACS file naming formats is supported. Program is able to read horizontal (phi=0) and vertical (phi=90 deg) planes from balloon data. Rotation of phi by +90 deg is needed if balloon data is captured with Klippel NFS having non-standard offset (-90 deg) for meridians. See Options for more information."

 

[Tom Danley]

Hi
I would add;
LSPcad for passive crossover design, ARTA for measurements, REW for it's EQ tool, REphase for FIR filters and AKABAK for complicated horns and cabinets. Fwiw, I developed the Tapped horns with the AKABAK program, powerful but something like PSpice and takes some getting used to.
AS mentioned in another post, a computer simulation programs real value is only as great as the ability to predict what you measure once you build the speaker. If your going to build more than one, it often pays to build a prototype first to see how close the measurement and predictions are.
Best
Tom Danley
Danley Sound Labs

 

[800]

@ctrl, was ABEC discontinued? I'm having trouble finding it online. Or is it part of AKABAK?

 

[ctrl]

was ABEC discontinued? I'm having trouble finding it online. Or is it part of AKABAK?

Yep, the successor software is AKABAK.
Unfortunately the successor is not 100% backward compatible (at least the last time I checked), so my parameterized scripts don't work with it. So I continue to use the old version.

The link to the previous version:
http://www.randteam.de/_Software/ABEC3_Pro/Download-ABEC-Pro.html

 

[800]

Yep, the successor software is AKABAK.
Unfortunately the successor is not 100% backward compatible (at least the last time I checked), so my parameterized scripts don't work with it. So I continue to use the old version.

The link to the previous version:
http://www.randteam.de/_Software/ABEC3_Pro/Download-ABEC-Pro.html

Thanks for the link. Guess I should have done a google site: search.

 

[alex-z]

Figured this thread would be a good place to mention a new REW feature.

V5.20 RC10 13th June 2021

Added an option on the Measure dialog to make a series of SPL measurements, e.g. when measuring polar responses. The selected delay is applied before each measurement.

Quite useful for anyone using a turntable setup to get spin data.

 

[Rick Sykora]

@ctrl could you give us short breakdown on what VituixCAD is capable of simulating and when you need to use AKABAK (or other comparable software instead?

Thanks!

 

[Chrispy]

LOL was more expecting table saw and router and jig recommendations

 

[BYRTT]

...Am preparing to direct Amir on how to bring the Klippel measurements for Directiva into VituixCAD. Is this simply using the import capability or would you suggest something else?

VituixCAD reads frequency responses exported from Klippel NFS in VACS balloon format...

From day one Amir don't export as "VACS balloon format" but share per each review into post-1 two "txt-files" named "SPL Horizontal.txt" & "SPL Vertical.txt" that each covers 36 directivities per 10º steps, in short to get those two txt-files VituixCAD friendly 1) remove comma's 2) split the 36 per 10º step column's to 36 seperate txt-files named in a way VituixCAD can read.

Myself have used MS Excel 2010 to do that task in manual way column per column and recorded process as a macro so as its a one click task run that MS Excel macro and spit out 36 times txt files per plane using Amir's two txt-files, at 07. of May year 2020 i shared that procedure and MS Excel macro file at below link, so should you be interested its all there in that link including @AnalogSteph kindly shared guide how to use free Libreoffice Calc instead of the need for a licensed MS Excel software feature ..

Link: https://www.audiosciencereview.com/...-vituixcad-using-amirs-shared-spindata.13136/

 

[ctrl]

@ctrl could you give us short breakdown on what VituixCAD is capable of simulating and when you need to use AKABAK (or other comparable software instead?

The following reflects only my own view.

I usually do not simulate the bass range (below 200Hz) with AKABAK/ABEC and use VCAD's enclosure tool or BassCADe. When building a bass horn or a transmission line you need other tools, of course, you can use BEM software (like AKABAK/ABEC) or a little less complex AJHorn or Hornresp.

When it comes to a concrete simulation of a loudspeaker above 200-400Hz, I always use BEM software (ABEC).
But of course this has its limitations. A huge silk midrange dome simulates badly, because it deviates very early from the ideal behavior (piston-like movement of the stiff membrane) - the influence of cabinet, waveguide,... is still simulated correctly, but the deviations of the final frequency responses can be considerable.

For the first rough appraisal of the simulation, I form, for simple constructions, mostly different example crossover in ABEC (via Lumped Elements).

For the final evaluation of a promising simulation, I export the simulated hor and ver frequency responses and import them into VCAD.
There, the potential of a possible loudspeaker can be assessed in detail and the effects of different crossovers can be simulated and analyzed in advance.

 

[Rick Sykora]

LOL was more expecting table saw and router and jig recommendations

Thanks. Fixed the thread title to clarify.