VituixCAD v2

[kimmosto]

Shift+mouse wheel is fine tuning, available also for delay. Smaller step is one sample.

[kimmosto]

Shift+mouse wheel is fine tuning, available also for delay. Smaller step is one sample.

Oops. Memory chip failure. Delay step of Buffer block is automatically sample interval without pressing Shift-key while wheeling. So for example with 48k sample rate the smallest possible step is 21 us. That is the most common approach in DSP gears. Exceptions exist. For example FourAudio uses steps of two samples so designer needs to know which delays are possible.

[mecedo]

I miss Shift+mouse wheel for delay fine tuning
Another improvement suggestion: 1.2 x wavelength in auxiliary wave length calculator.

[kimmosto]

I miss Shift+mouse wheel for delay fine tuning
Another improvement suggestion: 1.2 x wavelength in auxiliary wave length calculator.

Both done in rev 2.0.90.2. Delay step is 1 us with Shift + mouse wheel if DSP system setting is any digital. Without Shift key step is 1/samplerate which is usually only legal option in real life.

[mecedo]

Both done in rev 2.0.90.2.

Great. Thanks

[tktran]

Has anyone successfully modelled and designed a dipole in VituixCAD2?

[kimmosto]

At least I have designed also dipoles while VCAD has been available, but dipoles do not include any modeling with recommended design procedure.

[kimmosto]

For example this is one prototype with 2x15" woofers and 12" PA coax for MT. Directivity data at LF ca. below 120 Hz was generated with Diffraction tool but otherwise it's normal measurement data combined in XO simulation.

[Reet]

Has anyone successfully modelled and designed a dipole in VituixCAD2?

I have an open baffle “hybrid” where the bass is monopole, but midrange and tweeter are dipole. Followed same measurement process as any other speaker.

[kimmosto]

How about multi-platform VituixCAD 3.0 programmed with Java? Licensing might change because conversion could take a year.

[mecedo]

I have bad experience with java software. For me java programs are logy and open/save dialog windows are underdeveloped

[kimmosto]

What is logy? I think features of Open/Save dialog boxes have nothing to do with Java though multi-platform support might bring some limitations.

[mecedo]

They work slow. Response after clicking option is longer then in normal windows application. Opened windows loads with some delay etc.

[kimmosto]

Probably slower without multi threading, but proper CPU helps. For example this is quite okay 😉

[kimmosto]

2.0.91.0 (2022-10-08)

Main

• Added Bandpass and Bandstop shapes to generic Active Peak/Notch filter block.

Bandpass:

Bandstop i.e. notch:

Common dsp gear may not support these two without possibility to enter parameters as biquad coefficients a0...b2.

[Reet]

I have 2 simple suggestions.

• With recent releases of REW, I believe since 5.20.10 REW now includes a proper full dual channel measurement capability. Within the measurement timing dropdown, select "use loopback as cal and timing reference" for full dual channel, or "use loopback as timing reference only" for semi-dual channel. I wonder if it's worth updating the measurement instruction for REW to include some detail of the full dual channel measurement set-up, making it more consistent with the instruction for ARTA as far as the physical setup goes.
• Within the calculator tool there is an "average of A responses" function. I think it may be useful to have some ability here to apply weighting to each response in this averaging function.

Thanks!

[kimmosto]

..."average of A responses" function. I think it may be useful to have some ability here to apply weighting to each response in this averaging function.

Weight of each response can be adjusted with Scale dB value so this is already possible. "Problem" is that total is divided by response count. Not by sum of scaling factors. I could replace count with sum of scaling factors, but side effect is that level of result can't be scaled with common scaling text box above Scale dB column.

[Reet]

Hmm... SPL scaling and weighting of average is not the same thing. Adjusting SPL scale is still the same weight, but the SPL of the average overall is changed which is not the desired result. I suppose I could load in the same response multiple times to increase it's weight in the average as a sort of band-aid solution as it is.

I'm not sure how to best incorporate the weighting factor into the UI, my thought would be an additional weight column would be visible when the averaging function radio button is selected. I'm guessing this is more easily said than done.

[kimmosto]

Calculator calculates with absolute values (not in dB) and Scaling dB values are converted to multipliers with 10^(scale/20) so result is weighted average. Complexity at the moment is that sum of Scale dB values should equal to number of loaded responses in order to get correct original scaling for the result.
Example, weighting factors are: 2, 1, 0.5 and 0.5. Sum = 4 = number of responses.

[mecedo]

Kimmo, I have one little suggestion for software improvement. In "Convert IR to FR" tool if I select impulse response for load then destination folder is automatically set to the same folder. And when I change destination folder for output results then folder for impulse response loading is automatically set to the same folder as for output. Is it possible to unlink these directories? I think storing input data and output data in different folders is more common and practical.

[Reet]

Calculator calculates with absolute values (not in dB) and Scaling dB values are converted to multipliers with 10^(scale/20) so result is weighted average. Complexity at the moment is that sum of Scale dB values should equal to number of loaded responses in order to get correct original scaling for the result.
Example, weighting factors are: 2, 1, 0.5 and 0.5. Sum = 4 = number of responses.

Ah, I got it now. I can work with that, thanks for the tip.

[kimmosto]

Is it possible to unlink these directories? I think storing input data and output data in different folders is more common and practical.

Initial directories are separated in rev 2.0.91.1.
I don't know what is the most common, but I export frequency response (txt/frd) files to the same directory with impulse response (pir/mls) files. Both are the same thing, just in different domain. That also keeps directory structure more simple.

[tktran]

Hi Kimmo,

What's your current code in? C/C++

It will be a whole of pain moving to Java. What will you gain in the process?

[kimmosto]

What's your current code in? C/C++

It will be a whole of pain moving to Java. What will you gain in the process?

C#. Not much different than Java. Quite much work, and I don't need multi-platform so conversion is very improbable. Mostly just studying.

[Reet]

I’m sure everyone using non-Windows platform such as Linux or the popular M1 macs will be excited to be able to use VituixCAD without needing to fuss around with Wine or a VM. Myself included on that list but I’m sure it’s a huge undertaking.

[kimmosto]

I’m sure it’s a huge undertaking.

I could live with this work if every user would pay 2 €/month. Close to nothing, but "too much" for almost every DIY user. Linux users may not be much poorer than Windows and macOS users, but my impression is that less than five Linux users have donated in last five years. Linux culture is that software including operating system is free (and possibly open source) or included in device such as Android phone. Why should I work for a year that hundreds...thousands of users would get a bit more native application without paying a dime. Small extra inconvenience is that Java is worse than C#, and IDEs such as Netbeans or Eclipse are worse or more difficult than Visual Studio. Some Linux distros can be installed without problems, and some can not.
Therefore one possible target for conversion from .NET Framework to Java VM is to make Linux users pay for my work and update service. Quite obvious consequence is that users would continue with the last free .NET version and run it on VM with some old "free" Win, Wine etc. as long as possible. Just like users who were using Win XP until the first VCAD version which required Vista sp2 or Win 7 (due to .NET 4.5.2 requirement).

[Reet]

Well, there's at least 1 Linux user who has donated in the past 5 years. But, being that it's donation-ware at the moment, the income from donations should be viewed as a bonus rather than expectation. 2 €/mo works out to about the same USD$25/yr for software updates I was paying for SoundEasy, so you could say that cost is in line with another competing software. I often see the support for the lesser OS as a bit of a catch 22, developers won't support because there's little users, and there's little users because of lacking developer support, so it continues to be a free hobby OS for the average desktop computer, of course servers and data centres are a different story.

But, if you were truly to support Linux, there will be complaints of the closed source nature of the software . Anyway, I don't know how much specific effort is needed to support any specific OS with Java, perhaps John Mulcahy can provide some feedback there. I would hope that developing in Java makes the application multi-platform by nature without extra effort, but perhaps that's not always the case.

Mac support is probably a much larger market, I've met a few users of M1 macs running VituixCAD via Wine or parallels. I know a lot of recording studios favour macs as well, and with some recent youtube videos from PresentDayProduction perhaps there will be some increase in DIY speaker development for recording studios.

[Quza]

You probably don't like the idea, but as far as I'm concerned you could ask for donations more actively. People (read: me) have to be reminded that you put in a lot of effort into VituixCAD. Show a banner every 10th start or make shift+mouse-wheel fine-scrolling of parameters a 'premium' feature for 1€/month. You deserve a lot more credit (badumtss) for VituixCAD's influence on loudspeakers. The latest speakers of Thomann's own brand don't sound like shit mostly because VituixCAD makes it easy to work on them.


Now, to the reason I came to HTGuide today before I got sidetracked to donate:
Mabat (of diyaudio / at-horns.eu) normalises waveguide simulations to 10° off-axis as it makes it 'easier' to see relevant information. Could this be a useful feature for VituixCAD, too, or does mabat's reasoning only appy to waveguides/single speakers? My guess is high smoothness of ER, reference angle and generally off-axis responses relative to LW would make a good speaker, so it might be nice to normalise against LW.

[JonMarsh]

You probably don't like the idea, but as far as I'm concerned you could ask for donations more actively. People (read: me) have to be reminded that you put in a lot of effort into VituixCAD. Show a banner every 10th start or make shift+mouse-wheel fine-scrolling of parameters a 'premium' feature for 1€/month. You deserve a lot more credit (badumtss) for VituixCAD's influence on loudspeakers. The latest speakers of Thomann's own brand don't sound like shit mostly because VituixCAD makes it easy to work on them.

This must be where I come in and say, "Ain't that the truth!" OTOH, I would also say that anyone who has enough intelligence to do DIY development work would know and understand it intrinsically- but sometimes empathy requires an active effort and habit which is lacking in a transactional world view.

Now, while it would be a convenience to have VituixCAD running on M1 Macs for me, (Apple is my primary computing platform, and I have spent close to two years working on a special project in Apple Cupertino labs for the company I was employed by for 35 years), nonetheless, the realities of the world mean that computing is a mulit-platform situation and almost certainly always will be. SO, I have a 17" LG Gram that is used as a controller for my Audio Precision APx555, and also for running VituixCAD and pinch hitting on Altium when necessary. And I have a 2015 Mac Pro that is pretty much dedicated to running Windows 10 which is my primary Vituixcad, Altium, TINA, SIMETRIX and SIMPLIS platform. The windows systems are NOT used for online work except to download primary software and updates. And yes, all the nag messages on my LG Gram to upgrade to Windows 11 and to setup all the new little pieces of Windows infrastructure are annoying, and sometimes not easily dismissed, but can be dealt with using some persistence.

As to Kimmo and his work, his efforts have done no less than revolutionize the expectations and capabilities of speaker modeling and development. Hat's off to him again, along with a monthly donation. I had a previous habit of making a donation after every time I finished a piece of project work using VituixCAD, but that's relatively irregular income for Kimmo, in spite of how busy I am these days (four projects concurrently ongoing).

It's pretty redundant for me to say, "Keep up the great work, Kimmo!", but it needs to be said.

~Jon

[Reet]

Well, there's at least 1 Linux user who has donated in the past 5 years. But, being that it's donation-ware at the moment, the income from donations should be viewed as a bonus rather than expectation. 2 €/mo works out to about the same USD$25/yr for software updates I was paying for SoundEasy, so you could say that cost is in line with another competing software. I often see the support for the lesser OS as a bit of a catch 22, developers won't support because there's little users, and there's little users because of lacking developer support, so it continues to be a free hobby OS for the average desktop computer, of course servers and data centres are a different story.

But, if you were truly to support Linux, there will be complaints of the closed source nature of the software . Anyway, I don't know how much specific effort is needed to support any specific OS with Java, perhaps John Mulcahy can provide some feedback there. I would hope that developing in Java makes the application multi-platform by nature without extra effort, but perhaps that's not always the case.

Mac support is probably a much larger market, I've met a few users of M1 macs running VituixCAD via Wine or parallels. I know a lot of recording studios favour macs as well, and with some recent youtube videos from PresentDayProduction perhaps there will be some increase in DIY speaker development for recording studios.

[Reet]

You probably don't like the idea, but as far as I'm concerned you could ask for donations more actively. People (read: me) have to be reminded that you put in a lot of effort into VituixCAD. Show a banner every 10th start or make shift+mouse-wheel fine-scrolling of parameters a 'premium' feature for 1€/month. You deserve a lot more credit (badumtss) for VituixCAD's influence on loudspeakers. The latest speakers of Thomann's own brand don't sound like shit mostly because VituixCAD makes it easy to work on them.

Please don't put a banner asking for donations baked into the software, to me that's the digital equivalent of pan handling. If anything, go the "shareware" route like ARTA and provide everything except the ability to save project files for free, with a one time license cost for people who want to use the software seriously. You could also limit support to only licensed users to avoid the newbie with a USB mic crowd

[mecedo]

Kimmo. I'm making measurements in 10 deg angle step, but additionally I measure 5, 15 and 25 deg for better listening Window calculations. In options I have "Interpolate" checked and 5 deg for "Angle step" selected. But with such configuration artificial measurements are displayed on directivity chart and readability suffers. I've tried to improve readability by using "User's off-axis angles" but then lines on chart are bolded and big chart legend is put on chart. Can you try improve readability? I think it's possible in two ways:
- "Show only measured angles" option on directivity chart.
- Unbolding lines if "User's off-axis angles" option is selected and maybe option to hiding chart legend.

Edited:

I think I found the solution. Setting "Angle step" to 0 resolves issue an then on directivity chart only measured angles are displayed.

[kimmosto]

If anything, go the "shareware" route like ARTA and provide everything except the ability to save project files for free, with a one time license cost for people who want to use the software seriously.

One "problem" is that I don't need more money. Payment of at least 50 € is just one possible way to show appreciation, and single payment max 10 € is good for insulting because equals to work of 20 minutes or less. Copy protection with individual machine-dependent license keys requires quite much manual work and communication. Low price such as 85 € covers time loss for sales work, but not development of the product. Automated web shop is required to save time and guide money to product development.

Anyway, VituixCAD is primarily programming challenge, self-education and meets personal needs related to speaker designing. First two items motivate to convert to Java, but the last motivator is missing because Linux and macOS don't play any role in my life. So freeware with possible donations is the easiest and most flexible way. If story of freeware ends, commercial customers will continue with machine-dependent key files or some other protection technology or private free updates.

[Reet]

I don't need more money.

I could live with this work if every user would pay 2 €/month

These 2 statements appear to be in conflict. Regardless of how you choose to proceed with development, I would welcome the change to Java if it meant that Linux could run the application natively, in the same sense that I can use REW natively within Linux. It's not a big deal though either way, since I can run a VM within Linux equally as well currently

[kimmosto]

These 2 statements appear to be in conflict.

Idea behind those two statements was that I could quit my daily job and focus to VituixCAD full time if every user would pay 2 €/month. Close to full time is required in order produce JVM version within a year.

Fundamental conflict is that linux platform is community project where some members make applications for other members for free. Almost like a parasite ecosystem financed indirectly/unwillingly by companies paying salaries etc. I am not a member of that community because it can't offer me anything valuable/usable at the moment. And taking 2 €/month directly from users is not the "linux way".

[tktran]

what about subscription?

free for x months; €y/month for 12 months (or €z/annually) thereafter for software updates?

[Reet]

Software subscriptions are the worst.

[kimmosto]

The latest build 2.0.91.2 (2022-10-25) obfuscated with the latest Eziriz Intellilock v.3.0.0.0. Looks working fine, but please send feedback if any problems. Still one alert left with VirusTotal though the last version should decrease misunderstandings due to obfuscation.

[tktran]

Well, I donated €50 after a few months of use, and then €30 the following year around Christmas time. I plan to donate regularly, as long as I’m using it, and I don’t forget to donate..

how much is a coffee or other beverage in one’s local currency?
if you’re using it once a month, why not buy Kimmo a coffee?

The truth of humans is if it’s free the humans would rather have it for free. So yeah you’re gonna lose some users.

But now that you’ve got an established user base, and had your fun with a coding challenge, it’s time to monetise. Not to make money, but to give it your undivided attention( full-time job). Think of why you made in the first place there was nothing else that was very good. Do you want it to be better?
we do! Loudspeaker CAD is still in the Bronze Age IMHO. Think of sustainability. Think of succession planning. Do you have a five year plan, or two year plan?

what more could you do or add?
what features do your users want or keep asking for?

I want multi-monitor support.
I want full acoustic integrated measurements support.
I want full automated Spinorama support with 3rd party turntables.

Don’t forget that the Justin Finkel, the maker of REAPER, is able to dedicate his full time to it, because he already made his wealth from WinAMP and can work for free for the rest of his life. Not everyone can do that. The 99% of us have to earn a living. How are you going to make something that can compete with Apple Logic if you’re doing a part time.


€300 one off payment for a professional license is too cheap in my honest opinion, particularly if it has a such a small user base. Even if it cost it should be 10 times as much, would it be sustainable for the next 5 years?

[kimmosto]

I want multi-monitor support.
I want full acoustic integrated measurements support.
I want full automated Spinorama support with 3rd party turntables.

Operating system has multi-monitor support.
World is full of measurement programs. Interface between measurement program and simulator will be frequency response txt/frd file so it really does not matter who makes the measurement program. I'm in co-operation with Audiomatica so zero interest to make free or very cheap competitor for CLIO products.
Measurement programs have turntable support so I don't have to reinvent that.

Don’t forget that the Justin Finkel, the maker of REAPER...

Sorry, I don't know him.

€300 one off payment for a professional license is too cheap

The reason is tax policy. I can't exceed VAT limit. That would be fatal if/when sales drops. It has already dropped because significant manufacturers already have VCAD licenses.

[tktran]

On the multi-monitor support - i must have missed it; is there a way to have the 6 pack view on a separate monitor from monitor showing the crossover screen?

sorry, Justin Frankel.
Creator of WinAMP (early MP3 player), Gnutella (early peer to peer file sharing and REAPER (current Digital Audio Workstation for PCs)

[kimmosto]

On the multi-monitor support - i must have missed it; is there a way to have the 6 pack view on a separate monitor from monitor showing the crossover screen?

No because they are in the same window. That is very original feature (which will not be changed I'm quite sure).

[Reet]

It is possible, to extend the window across 2 monitors, drag the dividing line between crossover and graphs to the monitor separation. Not ideal but it works. I prefer ultra-wide monitor 3440x1440 is more than enough space for VituixCAD on one screen.

[kimmosto]

I have 17" laptop with 1920x1080 resolution setting (though it's native 4K). That serves everything required with single screen. Speaker projects are really simple so requirements are not high.

[tktran]

I have an ultraportable 14" laptop 1920x1080 and it's too small.
My other monitor is external- 27" 4K.

That explains why I think it would be nice to have the 6pack on the bigger screen and crossover on smaller screen.

Either that I need to get my eyes checked...