VituixCAD v2

[Reet]

VituixCAD doesn’t have the ability to measure anything, so first it would need measurement functionality before any incorporation of automated turntable would occur. Of course, ARTA has included this functionality for ages, and you can certainly use measurement data directly from SoundEasy as well with VituixCAD, though I haven’t looked closely enough at the new automated process in SE v28 to see if export of the impulse response can be included as part of the automated process. Reading through the manual pages, it doesn’t look like it.

[kimmosto]

2.0.85.0 (2022-04-01)

Main
* Added 'Copy biquad coefficients' window. User can include/exclude block type, biquad numbering, buffers, a0, parameter/coefficient name, value as decimal, value as 8.24 hex and comma (end of line). Selected (partial) text can be copied. Direct export as text file. Variables in template text file (.vxt) can be replaced with biquad coefficients and buffer parameters.
* Added 'Platin SM 48k' to DSP system list in Options window.
* Polar frequency response export uses file naming selected in Options: Generic 2D, CLIO 3D, EASE 3D, VACS 3D, MF 2D or MF 3D.
* X=0mm is centered in front view of Driver layout window.

Enclosure
* Enclosure tool opens with default driver configuration: driver count=1, isobaric=no, extra mass=0 g, output resistance=0 Ohms, to avoid wrong result due to unwanted value from previous session.

Merger
* " mrg" is not anymore added to the end of filenames to support balloon (3D) naming formats. Program refuces to save merged files to the same directory with far field HF responses if file extension is not changed txt->frd or frd->txt.

Diffraction
* Directivity export uses file naming selected in Options: Generic 2D, CLIO 3D, EASE 3D, VACS 3D, MF 2D or MF 3D.

General
* Response file writer compresses .cal and .mic files by skipping frequency points where magnitude slope changes less than 0.12 dB/oct. At least one point per octave and first and last point are written though response is perfectly flat. Feature is easiest to use with Calculator tool.

[kimmosto]

how come your website no longer have all the old projects posted?

Projects were mostly experimental prototypes and our weather is almost continuously unsuitable for measurements outdoors so data was partly inaccurate.

By the way, any future plans to have VituixCAD interface with any automatic turntables?

As already told, automatic turntable control is most useful as integrated to measurement program. Turntable manufacturers have their own remote apps which can be really good, but user has to click two programs.
My motivation is quite close to zero because I have CLIO 12 QC, CLIO QCBox 5 and Outline ET50-3D turntable. VituixCAD reads mls files by Audiomatica so package is quite all inclusive.

[kimmosto]

Microsoft does not support and distribute .NET 4.0 anymore, and uninstalling VS 2019 will remove also development tools for .NET 4.0 from my PC.
Next VituixCAD revisions (2.0.85.1...) will require .NET 4.5.2 and Windows 7 sp1 or later. XP users should upgrade directly to Win10 imho.
This brings also some new possibilities such as project packing without external tools.

[kimmosto]

2.0.85.1 (2022-04-03)

General
* System requirements changed: Windows 7 sp1, Vista sp2, 8, 8.1 or 10.
* Microsoft .NET Framework 4.5.2.

Main, Copy biquad coefficients
* Byte order of buffer:unmute as 8.24 hex reversed to be compatible with SigmaStudio.
* Added generic variables $(filename) and $(description) to Replace command.

Enclosure
* Improved driver tracking when database filter settings or PR status is changed.
* Web search done with currently selected driver (active or passive) instead of project's active.

[sheeple]

Hello Kimmo,
ist it possible to have DI trace starting from lowest SPL value/bottom?

[kimmosto]

^It is for sure, but it's also possible that DI is negative at LF. For example side woofers and port in the back produce negative DI.

[Quza]

I noticed a big difference in amplitude response when using a Bessel crossover filter and changing the 'DSP system' in the options from Analog to Generic. The group delay seems better in the analog version, as that's what it was designed for, but in the generic version there's hardly any difference to a LR filter. Example:





Can you give me a hint where I should have a look to understand this difference?

[kimmosto]

^Unfortunately there's no global standard, specification or gentlemen's agreement that Bessel LP/HP must be implemented as either "phase-normalized" or "level-normalized" so every DSP manufacturer decides which specification they use. "Phase-normalized" Bessel LP/HP has phase shift of 45 degrees x filter order at nominal frequency. "Level-normalized" has -3 dB point at corner frequency. Both produce exactly the same magnitude and phase responses so only difference is how nominal frequency is specified relative to response corner. Both are as logical or illogical imo because Bessel LP+HP does not produce constant magnitude response. Maybe Friedrich Bessel (1784–1846) specified which one is correct/original but ca. half of the manufacturers don't respect that.

[Quza]

Thanks for your explanation! That's indeed unfortunate.
FWIW: Analog Devices seems to be fitting the Generic DSP system setting, which seems to be the same as the miniDSP ones as they are using AD dsps.

And a small 'bug report': In the help file at Suffixes of variables in equations _ER is missing. Though it is already explained in Power&DI imo it wouldn't hurt to add it there, too.

[kimmosto]

^I will add that.
P.S. >90% of bug reports in my e-mail are actually ignorance, assumptions, blindness or laziness of the user. Feedback without any judgements or classifications works much better in both directions.

[tktran]

When measuring drivers in the far field, one should leave all the controls the same. I get that.

I measured the tweeter first, on baffle, using 2.83V output (RMS AC Multimeter), thinking I could kill two birds with one stone and measure the actual sensitivity as well.

When I measured the mid-driver, I got a pop-up message that I may have clipped the mic. I was not prepared for how sensitive this pro midrange was!
Instead of restarting all measurements a lower level (since I had some a complete CTA2034 spin at 10 degree intervals for the tweeter) I decided to continue and take measurements of the midrange at -6dB.

I figured I could always manipulate the data later, and add 6 dB.

As it turned out, the FR captured at 1.415V + 6dB is not exactly the same as 2.83V.



But is it good enough for crossover work?



Has anyone else made this boneheaded mistake and continued?

If I have to re-do all measurements, what voltage (or SPL) do people typically aim for when measuring at 1m?

[JonMarsh]

Just my opinion, but I think you're agonizing over this a bit much... if I understand your problem and the scale on the measurements correctly, you're talking about small fractions of a decibel.


Try tearing down and re-setting up and duplicating your measurements under the original condition- I doubt you would do any better.

[Reet]

Wow, you're off by 0.1dB, throw those measurements away! :P

[Reet]

Just a simple suggestion for the "traces" windows. It would be great if it highlighted the trace on the graph when selected, similar behaviour to when you hover your cursor over a trace. This would make it easier to make sure you've selected the right trace when multiple overlays are loaded and you want to remove one, for example.

[kimmosto]

^That and something extra is done. Hopefully direct and faster access and preview of trace properties is an improvement.

2.0.85.2 (2022-04-22)

General
* Row of highlighted (bold) trace is selected when Traces window opens.
* Trace of currently selected row in Traces window is highlighted in chart if the trace is visible and selection is not locked.
* Apply and OK buttons are removed from Traces window. All changes are applied immediately without possibility to cancel/undo. Use View->Default colors to restore default properties if required.

Main
* Adjusting frequency min/max limits of charts does not interfere Y axis scaling.

Enclosure
* Zooming or adjusting frequency min/max limits of charts does not interfere Y axis scaling.

[Reet]

Beautiful!

For some reason horizontal response with "generic 2D" naming convention are loading as vertical angles, as a result Vituix is refusing to load the vertical angles as they are duplicate.

[Edit: VituixCAD seems to be grabbing the "ver" text from the beginning of the file name, even though "from the end" is selected in the number format search option.]

[kimmosto]

^I've tried to keep logic simple:
- filenames containing vertical keyword are vertical
- all others are horizontal.
This also allows ignoring plane keyword if data is for horizontal plane or on-axis only. This is one the most common RTFM cases and mistake in file naming.
Search from beginning/end applies to angle value only. 'Number format' group tries to tell that.

Measurements for CLIO has something for this, but it's not yet copied to other versions:
"Format for dual plane far field measurements is drivername and plane keyword separated with _ or space. For example, PTT65W04_hor is PTT65W04 in horizontal plane.
Note! drivername should not contain plane keyword - especially “ver”. For example “Driver1_hor” is illegal causing
association to vertical plane."

One cheap trick to allow "ver" text in horizontal measurements is to start using balloon file naming.

[kimmosto]

Of course I could add detection for filenames containing both keywords. The last one is taken if "Search from end" is selected. Accidents happen and this would help if manufacturer or model name or some extra data contains keyword of perpendicular plane.

[kimmosto]

^Done in the latest build of 2.0.85.2 (2022-04-23)

[Reet]

^Done in the latest build of 2.0.85.2 (2022-04-23)

Thank you. I am okay with the ver/hor search logic, as long as there is still some method to load the file anyway and manually enter the angle values. The problem is when the program refuses to load the file simply because the naming doesn't conform exactly to what the software is expecting. So instead of refusing to load the files, my suggestion would be to produce an error indicating duplicate angles, and highlight the second file in red until a manual correction is made.

[kimmosto]

Multiple files such as full balloon data imported from Klippel or EASE is quite impossible to control manually because all angles could be twice, for example 146 unnecessary lines to clean from each driver. Frequency response interpolation with four response is also quite demanding so there can't be any trash; extras or manually composed rows. Valid file naming makes loading fast and easy so it's just good opportunity.

[ergo]

Reet

I use the Total Commander for solving issues with bulk file naming/renaming. It works also without buying a license for occasional use.
This has helped me to solve many similar issues with little effort.

[Scottg]

Sonus Faber promo material:

https://www.soundstageultra.com/images/stories/opinion4/202109_opinion_5259.jpg

https://www.soundstageultra.com/index.php/features-menu/opinion-menu/1056-material-obsession-sonus-fabers-maxima-amator

Hmm, that looks a little familiar.. :W

[kimmosto]

Hmm, that looks a little familiar.. :W

Yes. There it is https://kimmosaunisto.net/Software/Software.html#Links

[tktran]

Well good on you Kimmo. Changing the course of loudspeakers for the better.

Maybe now Sonus Faber can design proper crossovers. They always made fancy cabinets to bring to market as a luxury foreign item, but there’s never been one Sonus Faber speaker that measured well or sounded neutral.

Hopefully B&W follow too…

[kimmosto]

Maybe now Sonus Faber can design proper crossovers. They always made fancy cabinets to bring to market as a luxury foreign item, but there’s never been one Sonus Faber speaker that measured well or sounded neutral.

Quite difficult to agree, but I've not heard much other models than Homage Stradivari which is/was one of the best speakers in the world.

Hopefully B&W follow too…

Sound United R&D in the USA uses VituixCAD, but I don't know is there any co-operation with U.K.

[Scottg]

They always made fancy cabinets to bring to market as a luxury foreign item, but there’s never been one Sonus Faber speaker that measured well or sounded neutral.

Despite the "character" of several SF models, I always thought they sounded pretty good ..except for dynamics.

What kills me is that "wood" is literally in the brand's name, but except for veneer or real-wood cladding these days - it's all MDF. :x (..so mostly glue and sawdust, most notably where the driver mounts to - which is just wretched and appalling in such an expensive brand.)

Still, it's good to see that manufacturers recognize the value of Vituixcad.

[JonMarsh]

Well good on you Kimmo. Changing the course of loudspeakers for the better.

Maybe now Sonus Faber can design proper crossovers. They always made fancy cabinets to bring to market as a luxury foreign item, but there’s never been one Sonus Faber speaker that measured well or sounded neutral.

Hopefully B&W follow too…

Gawd, there's so many comments I could make I just have to keep bottled up inside... after all, I am a moderator.

:rant:

Where's ET when you need him?

:B


:rofl:


OTOH, given my experiences of the last 35 years and some occasions upon which managment has been handed some remarkable opportunities due to breakthroughs in tools, and fumbled it, I wouldn't count on anything.

When I was a partner in an audio shop that carried B&W in the mid 70's, I was so disappointed in some of the characteristics of the DM6 that I bought a full driver set from warranty stock, designed new cabinets for them, and of course, a rather different crossover. Of course, they were bigger than the DM6... and more neutral and transparent and imaged far better. Other than that, same old same old drivers.

Kimmo has certainly set a benchmark for speaker design software- :yesnod: :T


back then, all I had was a calculator and a lot of decade switched component boxes I'd built for caps and inductors.

[kimmosto]

Almost every commercial speaker looks and sounds wrong and is too expensive when looking from DIY perspective. Unlimited possibilities, materials, driver and component budget, time, design, finishing etc. makes things much much easier than for average commercial designer. There is always someone in the company who is not happy with component costs, manufacturability, outlooks, ... High quality finishing is very difficult and expensive in the western world. Especially for small manufacturers who can't make or buy perfect mold(s) and order thousands of boxes for almost free. Price competitiveness is required to break in to market where big brands rule with their very cost-optimized products - regardless of price category. Sound quality is not usually impressive, but who cares as long as buyers prefer brand and resale price.
So I don't much respect criticism from DIY scene before designing and selling at least ten pairs to Finland for 5-40 k€/pair via audio store with taxes, decent profit for all etc.

Reviewing of products is another easy job. Just elevate yourself to a specialist, judge and arrange manufactures to lick your a**. Acting and/or writing skills are needed of course so one can't be totally brain dead.

[tktran]

I hear what you're saying. Like software, loudspeakers it's hard to make a product that's universally liked and people are willing to pay, such that someone is able to make a decent living wage.
Long hours, deadlines, good enough to release and debug/fix later. But we love VituixCAD and hope it gives you joy and energy rather than deflate or tire you.

As for for the Homage Stradivari I remember when this came out in Australia. Almost 20 years ago and it was going for AUD$100,000 which is approximately US$70K in 2005. Relative to inflation, that's the equivalent of
US$100K today.

Can we at least get the listening window correct?

Reference: Stereophile review- quasi anechoic measurements; not baffle step loss corrected, horizontal listening window +/- 30 degrees

[kimmosto]

Can we at least get the listening window correct?

Stereophile's measurements are totally inadequate for conclusions. In addition, Stradivari is not listened at tweeter's axis so listening window result is not valid in real life. So please do not judge without proper data and/or listening experience.

[Scottg]

As for for the Homage Stradivari

..not baffle step loss corrected, horizontal listening window +/- 30 degrees

Might want to check that with Vituixcad.. baffle is 650mm wide and the measurement condition was about 1.25 meters. :W

The narrowing horizontal directivity (at high freq.s past 5kHz) of the ring tweeter on a large baffle is fairly common. I hated that tweeter grouping (..though it did have good clarity), lots of people really like it though. :roll:

BTW, I believe Serblin's premium designs were all real wood, it wasn't until Fine Sounds purchased the Co. that they gutted the carpentry/faber for anything except aesthetic finishing.

[dwk]

Can we stop talking about Sonus Faber?

Before we moved ~18 months ago I was moving towards my own riff on the SF Elipsa which I find just stunning. I thought I had exorcised those thoughts, but now y'all have me looking at where my current speakers are positioned and thinking 'maybe they'd fit after all....'

[kimmosto]

BTW, I believe Serblin's premium designs were all real wood

Our companies have also abandoned real wood due to very high risk of shrinkage and cracking in our climate if all six separate walls are glued together. Otherwise boxes are made twice; first of wood, and after a year or two when cracked glue & dust, artificial stone etc. Of course wood is okay as a baffle or side wall frame or just decoration trim (with flexible joints, surface treated inside and out). Making boxes with close to perfect finishing for commercial speakers is ridiculously expensive no matter body material. It could be the most expensive component also in powered DSP speaker.

[JonMarsh]

Cocobolo holds up pretty well in dry climates... of course, I have a stash from years ago. In climate controlled storage!

:rofl:

[kimmosto]

Cocobolo is endangered/vulnerable. Our local woods such as alder, aspen, birch, pine, spruce and larch are the most responsible options here. Some of those crack before starting or in a week Carpentry and finishing is still too expensive so manufacturers order boxes from Poland or Baltics if higher than standard crap pro quality is required.

[Scottg]

Our companies have also abandoned real wood due to very high risk of shrinkage and cracking in our climate if all six separate walls are glued together. Otherwise boxes are made twice; first of wood, and after a year or two when cracked glue & dust, artificial stone etc. Of course wood is okay as a baffle or side wall frame or just decoration trim (with flexible joints, surface treated inside and out). Making boxes with close to perfect finishing for commercial speakers is ridiculously expensive no matter body material. It could be the most expensive component also in powered DSP speaker.

https://www.youtube.com/watch?v=iHp1rKVLOII

..well there are more financially disastrous materials like milled aluminum. :W

Still, I can see ditching hardwood in favor of veneer-clad void-free plywood (spot-braced). *MDF though.. on a 5-figure loudspeaker? :nonod: ..and I'd really like to see something more rigid and hard for the baffle - like that super-compressed resin-impregnated plywood that's available in Europe (or better, a synthetic stone). Good Bamboo ply is also very good.

*it suffers from
1. a lack of rigidity and hardness
2. durability/longevity (especially when exposed to water)
3. toxicity.

[JonMarsh]

One of my wood based favorites is piano sounding board maple ply.

Ironically, very happy with the performance on a recent project, completed, but the first usage case for me is still in storage, though that will change soon. Then that system will hopefully be completed this year.


But let's not forget, for any material, it's understanding it's strengths and weaknesses, and coming up with ways to optimize the usage principle.

This can work for MDF, which is pretty strong under compression, and can be used as a wall material shaped to necessary purposes.

On the other hand, while the results can perform quite well, they won't be inexpensive or simple to build.

[tktran]

When I started in this game 20 years ago there were all kinds of local and foreign (imported) timbers available, including solid timbers and pre-veneered MDF and plywood boards. All kinds of exotic timbers with beautiful figuring; like bird’s eye (Canadian) maple and or Tasmanian (island off south east Australia) myrtle and waterfall bubinga (African) available to purchase in Australia; imported from all around the world; felled from trees with little conservation or regulation, I’m sure.

These days all those timbers are hard to find. Thank goodness, too.

And after reading Brandon’s monster construction thread I’m ready to quit Russian or other Baltic birch plywood and use multi-ply MDF in a constrained layer form, for more reasons than one:

https://www.somasonus.net/box-construction-methods

Then I paint the MDF; in any colour I want, as long as it’s black.

[JonMarsh]

Well, you have a clear and relatively simple plan to follow! And I think ET would fully approve your color choice!


:rofl:


But, sometimes other choices have their attraction, too...

[kimmosto]

I'm updating Linux & Wine installation instruction due to .NET 4.5.2 requirement of 2.0.85.1 and later. It this good enough? VituixCAD_in_Linux_v0.2.pdf

[JonMarsh]

That looks pretty good to me, though I'm not a frequent user of these OS's- more a MacOS sort of guy.

[Quza]

[...] Is it valid to correct ground plane measurements by the difference of the diffraction simulations with normal baffle height and twice the baffle height? And if so, am I to apply that correction to the off-axis measurements as well? [...]

Ground plane measurements are far field measurements => goal is to measure so that mathematical corrections to response shape are not needed i.e. responses are directly valid for HF response list of Merger tool with level scaling by -6 dB only. One error source with ground plane is that (small) speaker on the floor is not exactly the same as speaker on final stand due to differences in baffle loss and diffraction. One workaround is to make small stand to cause diffraction and baffle loss also below bottom panel, and tilt speaker down so that DUT points to microphone while measuring 0 deg response. Rotation center should stay in the center point of DUT on baffle surface, and manual rotating table should be thin or extended to mic or some other invention to avoid diffraction from ground plane. This kind of tricks are easier with small speakers. Tall constructions need supports, turning upside down etc.

For reference, I did this with a python script and got this result:









This is an 8" speaker, nearfield merged to farfield, in a box WxHxD 27x47x29cm, so omnidirectional to ~150Hz seems a reasonable result - even though I did not expect to 'fix' this by applying the correction. Higher SPL behind than infront of the speaker at 150Hz looks very wrong. Measurements were done with "AP-Equiripple" window.
What I did was to subtract simulated air response by simulated ground response (height extended below ground (-470) by the height of the speaker), for every angle +-180 separately, and then apply the difference to my measurements.
I think with dB values I should divide instead of subtract, I'll try that with the next speaker.

[kimmosto]

Fine tuning. Some of those were already done to earlier builds of previous revision.

2.0.85.3 (2022-05-06)

Main
* Intensity on spherical surface, Intensity on cylinder surface, Include horizontal, Include vertical, Half space, Corner and Listening window DI settings in Options window saved to project file (vxp).

Impulse response
* Changed to top-level window.

Preference rating
* Minimize box disabled.

Convert IR to FR
* Right window set to 4.0 ms and IR max to Reference time + 5 ms and rounded up when Far 1 button is clicked.
* Filename added to overlay description with 'Save as overlay' command.