I have a use-case where I need at least about 80dB.
When using a metal-cone midbass with really bad cone break-up modes, you want to design a steep lowpass so that the cone break-up modes are pushed down at least about 70dB, to ensure that they're not audible. (I first learned this from Jon Marsh over at htguide DIY forums, when he designed his MTM using Dayton RS metal cone drivers.) So, either a Cauer-elliptic or a high-order straight slope is needed.
With VituixCAD, since the max SPL range is 60dB, I don't know how far below the cone break-up modes are. I'd like to see where they're going. 80dB or 90dB would be nice.
VituixCAD is just stunningly good. My software stack earlier was:
In the interim, I've designed and built a 3-way active speaker, learned to use a new sound card (Focusrite Scarlett 2i2), learned how to get a custom-biquad digital active xo done on MiniDSP, learned Active Crossover Designer, got myself a calibrated EMM6 mic, and so on. And even after all this, I think VituixCAD will let me get my jobs done without having to look beyond, for at least a few more years.
I guess I'll just stick with this stack till they fall out of support and maintenance.
Some bugs and some wishes
There are too many things I love about VituixCAD, having used other programs:
- Unibox, and
- Speaker Workshop
- Unibox
- ARTA+LIMP
- VituixCAD 2
In the interim, I've designed and built a 3-way active speaker, learned to use a new sound card (Focusrite Scarlett 2i2), learned how to get a custom-biquad digital active xo done on MiniDSP, learned Active Crossover Designer, got myself a calibrated EMM6 mic, and so on. And even after all this, I think VituixCAD will let me get my jobs done without having to look beyond, for at least a few more years.
I guess I'll just stick with this stack till they fall out of support and maintenance.
Some bugs and some wishes
- Bug: When I renumber all the parts 1,2,3 in the "Parts Numbers" section of the menu, the new numbers don't show up. I have to click each component for its newly assigned part number to show up. Can these numbers be made to show up automatically on renumbering, please? This is almost certainly a bug.
- Bug: When I install VituixCAD on Windows 10 as an administrator, then logout and login as an ordinary user, I can use VituixCAD. But I can't see any of the library blocks -- I'm presented with an empty library. I logout, and login again as administrator, and run VituixCAD, and all the library blocks are there.
- Can I get an optional checkbox against each driver, called "Also show inverted SPL"? Basically, if I mark this checkbox on for a specific driver, then the SPL graph will plot two lines overlaid on each other, one with this driver connected straight, the other with the driver inverted. This will be superbly useful for manual tweaking of xo components, because I'll then be able to see at one stroke, the effect each change of value is having on the (i) normal summed SPL and (ii) one-driver-inverted SPL showing the reverse notch.
- Can the "save project" operation give me an option to save with all the driver data? This way, I'll get a big project file with all the FRD and ZMA data included, and I can easily email the (single, large) file to a friend for collaboration or analysis.
- It would be nice to have a "Hold" button on the graphs, so that the graph in the window where I click "Hold" will just freeze. Then when I go tweak xo component values, I'll see a freshly redrawn graph, and also the earlier held graph. (Idea borrowed from XSim.) Works for very effective before-after comparison.
- Can the inductor resistance change as per some formula when I change the inductance value? For instance, if I change the value of an inductor from 200uH to 1.5mH, the resistance of that coil too should change as per some formula, so that I can see the net effect of both new inductance and new resistance on the xo curves.
- It's great to have a "six-pack" image export. Can we have an additional image export option, where we get five different images in five different files, for the five graphs? Currently, it's either export-one-graph, or export-one-image-with-five-graphs.
- Can we toggle all the optimise-lock-unlock flags of all the components of one "way" (remember, the concept of "way" in VituixCAD v1?) with one operation? As far as I can see, if I'm optimising just the tweeter curve, I now have to manually go switch off optimisation for all the midbass components. Then, when I do the midbass curve optimisation, I go switch on each midbass component optimisation one by one, then go switch off all the tweeter components one by one. And with higher-order slopes, that's a lot.
- Can I be given the option to load an arbitrary FRD file for goal for optimisation?
- Can the xo schematic drawing editor be programmed to automatically discard any net which is free-floating and unconnected with any component? I don't see why such net fragments (entered by mistake or left over after component deletion) are allowed to remain.
- Bug: In the xo drawing editor, I can start entering a track and then sometimes, I can't exit even after hitting "Escape" or other mouse clicks. I am just locked down in the schematic editor, and each mouse click just adds one more segment to my net. I once had to actually abort the VituixCAD application itself to get out of this lock-down.
There are too many things I love about VituixCAD, having used other programs:
- I love the way the "Convert IR to FR" panel works. I load the PIR file and play with the gates, and watch the SPL curve change. As good as HolmImpulse. Better than ARTA. Awesome.
- I love the xo value tweaking ease. I just keep tweaking and watch all graphs change simultaneously.
- I love the fact that the optimiser supports higher order slopes, not just stopping at 4th order. (Why would anyone think that 4th order is adequate? Anywhere?)
- I love the GUI of the free-format xo schematic drawing window. Very nice aesthetics. Love the red dots of the floating terminals.
- Love the fact that the power response analysis and axial SPL response are both given equal visibility. I don't have to go to some separate section to just see the power response.
If any of these points are arising because I've not understood this (really great) tool enough, my apologies in advance.
Last edited:
There are too many things I love about VituixCAD, having used other programs:
Was going to edit the previous post and add this block, but I ran out of the 30-min edit window.
- I love the way the "Convert IR to FR" panel works. I load the PIR file and play with the gates, and watch the SPL curve change. As good as HolmImpulse. Better than ARTA. Awesome.
- I love the xo value tweaking ease. I just keep tweaking and watch all graphs change simultaneously.
- I love the fact that the optimiser supports higher order slopes, not just stopping at 4th order. (Why would anyone think that 4th order is adequate? Anywhere?)
- I love the GUI of the free-format xo schematic drawing window. Very nice aesthetics. Love the red dots of the floating terminals.
- Love the fact that the power response analysis and axial SPL response are both given equal visibility. I don't have to go to some separate section to just see the power response.
Was going to edit the previous post and add this block, but I ran out of the 30-min edit window.
[*]Can we toggle all the optimise-lock-unlock flags of all the components of one "way" (remember, the concept of "way" in VituixCAD v1?) with one operation? As far as I can see, if I'm optimising just the tweeter curve, I now have to manually go switch off optimisation for all the midbass components. Then, when I do the midbass curve optimisation, I go switch on each midbass component optimisation one by one, then go switch off all the tweeter components one by one. And with higher-order slopes, that's a lot.
[*]
^ +1
Renumber all 123 is also manual operation; click all components in desired order. You can keep Ctrl key pressed to see what components are already clicked. I don't like to do fully automatic numbering.
So this works as designed.
Can you run setup as administrator without logging in as administrator? This should be possible right after download by canceling automatic setup and finding and running setup.exe from user's VituixCAD\Download folder. That way user documents should be installed into current user's directory.
I have some bad habits; kimmo is administrator in my laptop so everything goes smoothly until...
I have noticed this few times but never in debug mode to catch the reason. Some status bit might cause non-recursive infinite loop. I'll try to remember this.
Rev. 2.0.14.1 (2019-04-22)
SPL scale expand and compress buttons are added close to Reference angle text box. Available spans are 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90 dB. Initial value is coming from Options window.
35 and 70 dB are really bad with secondary axis so these two will probably be deleted some day.
25 and 50 dB are a bit better, though 36, 72, 108, 144 deg are not so nice and readable as phase interval values.
45 and 90 dB aren't actually bad at all. Only problem is that 0 deg line does not exist. That would be possible to show as zoomed when more pixels are available for values. Scales has to work also with 200px image height so comparison to big single charts of some other programs is not relevant.
Another problem with secondary axis is Directivity Index curve in Power chart. Some span options require 3 dB major intervals for DI. Minor intervals are something else than full 1 or 2 dB.
Enclosure, Merger and Convert IR-FR have less span options: 20, 30, 40, 60, 80 dB.
SPL scale expand and compress buttons are added close to Reference angle text box. Available spans are 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90 dB. Initial value is coming from Options window.
35 and 70 dB are really bad with secondary axis so these two will probably be deleted some day.
25 and 50 dB are a bit better, though 36, 72, 108, 144 deg are not so nice and readable as phase interval values.
45 and 90 dB aren't actually bad at all. Only problem is that 0 deg line does not exist. That would be possible to show as zoomed when more pixels are available for values. Scales has to work also with 200px image height so comparison to big single charts of some other programs is not relevant.
Another problem with secondary axis is Directivity Index curve in Power chart. Some span options require 3 dB major intervals for DI. Minor intervals are something else than full 1 or 2 dB.
Enclosure, Merger and Convert IR-FR have less span options: 20, 30, 40, 60, 80 dB.
Great. I can see how the manual approach is more powerful. But for better UX, can you add code to pop up a message box immediately after I've selected the "Renumber 1,2,3..." menu option, telling me something like "Renumbering mode is now active. Click on the components one by one in the sequence in which you want them numbered. Have a nice day!"
I'll definitely try this. However, I wanted to let you know that I've actually installed it once as a non-privileged user and the install went through without asking me for admin privileges. This was on the same laptop where I'd installed it as a privileged user a few minutes back. And still I never got to see any library blocks.
Ya, I noticed that. But nothing prepared me to actually start clicking on the parts to make the renumbering happen. I discovered it accidentally by clicking on some part to see what its values were, and then I realised that it had just gotten renumbered. So I kept clicking. And at the end, I concluded it was a bug with screen refresh. I didn't realise even at that time that this was its intended behaviour.
Just checked this, but unfortunately running as administrator doesn't change anything because setup programs selects admin's Documents folder when privilidges are elevated. So user have to be added to admin group (at least for temporarily) in order to install user files to user\Documents\VituixCAD\Enclosure, Library and Projects folders.
I will try to find some workaround.
merger tool query
Hi all,
bit of a query about the merger tool.
When I do a nearfield sum of bass drivers and port,
in ARTA I get this:
12 Total summed bass via ARTA.PNG - Google Drive
Note 3 traces: combined bass driver trace, port trace and summed response
Whereas in VirtuixCAd I get this:
12 Total summed bass via Virtuix CAD.PNG - Google Drive
below 300hz the 2 diverge quite considerably and I cannot figure out for the life of me why. They are both build using the same .frd files. You can see on the traces as I've shown .
Which to trust?
any ideas greatly appreciated.
Ced
Hi all,
bit of a query about the merger tool.
When I do a nearfield sum of bass drivers and port,
in ARTA I get this:
12 Total summed bass via ARTA.PNG - Google Drive
Note 3 traces: combined bass driver trace, port trace and summed response
Whereas in VirtuixCAd I get this:
12 Total summed bass via Virtuix CAD.PNG - Google Drive
below 300hz the 2 diverge quite considerably and I cannot figure out for the life of me why. They are both build using the same .frd files. You can see on the traces as I've shown .
Which to trust?
any ideas greatly appreciated.
Ced
^Time window settings should be equal to all near field measurements to be summed. There is also small risk that IR actually starts just before (Left) time window opens if Reference time is set to IR peak automatically with Near button in Convert IR to FR tool. This could happen with large subwoofers or other drivers incapable to produce mid...high frequencies. One method is to set Reference time manually to sample #300 when IR measured with ARTA (and measurement channel signal doesn't include extra delays of DSP). Right time window to 600...1000 ms for all near field measurements. Then export to txt/frd. This way IR starts (for sure) after time window has opened and all measurements have common time reference i.e. summing with phase works.
Phase matching is monitored with GD & Phase graph.
In addition, speaker is not listened with reversed polarity of some individual driver, and phase mismatch can be used to get more balanced sound with smoother power response or smaller excess group delay / better step response. For example textbook But18 with MTM or Harch/LeCleach XO. So reverse polarity check is mostly just an old tradition to solve difference between acoustic centers when measurement data was "single channel" with random or normalized (min phase) timing.
Already considered this, but as zipped rather than included in xml due to size and performance. This feature would already be in the program but .NET 4.0 does not have direct support for multi-file zip. Single-file gzip only.
I have used 7zip for project archives and distribution. 7zip.bat, %1 = project directory below user\Documents\VituixCAD\Projects:
Code:
@ECHO OFF
IF .%1.==.. GOTO EXIT
"%programfiles%\7-Zip\7z.exe" a -r "%1.zip" "%1\*.vx*" "%1\*.txt" "%1\*.frd" "%1\*.zma" "%1\*.png"
:EXITSnapshot/overlay feature tries to play this role. Overlays are also loadable from files.
Wire diameter is available so few more assumptions about geometry and core material are needed to estimate DCR. But DCR could be one parameter to optimize independently without relation to inductance value. So this should be enabled per coil with some control key.
We could play also with custom library blocks with single inductor. That would enable DCR calculation with wire diameter, core length and core diameter. Single block for each core type.
Single chart export takes about 7 seconds. 5 x 7 = 35 s. How much you're willing to pay me for one hour work that you could save few times 35 s in each project?
But seriously, you all should estimate this kind of things while asking changes to freeware. Especially if asked features are cosmetic e.g. visual compatibilily with some relic/obsolete software, design time saving is only few seconds or feature does not directly affect to sound quality. I'm happy to say no to everything if it's obvious that my free work hours are used for business with zero donation.
This is available for single drivers. Not for total axial SPL or Power. Quite good reason to assume that those are flat horizontal or tilted.
Separate cleaning command is possible. Components with all terminals in red are easy. Groups of wires are more complex due to intermediate nodes.
Thanks for ideas!
Yes, all these approaches are probably good for various situations. I was hoping for an approximate estimate of DCR by using a linear formula linking the inductance and wire gauge, assuming air core.
This touched a raw nerve, so I've donated EUR 100 to you through Paypal just now. I know that for a person of your education and talents, this amount will probably not pay for even 30 minutes of your professional time, but it's a start. And if you feel that you need to be compensated more, it may be possible for you to suggest a group-donate activity like the group-buys this site constantly organises?Single chart export takes about 7 seconds. 5 x 7 = 35 s. How much you're willing to pay me for one hour work that you could save few times 35 s in each project?
But while the topic of compensation for your efforts is an important one, I feel it is best if it is not mixed up in a thread which is discussing feature requests and bugs. It's hard for users like me to interact with you if we are also asked to take responsibility for calculating how much the money value of developer time will be to implement suggested changes. I think that that calculation and those trade-offs are best left to you.
Yes, it does! I had missed it. Thanks for pointing it out. And I agree, it's needed only for axial response for a single driver, not for total axial SPL or power.
Last edited:
Now there is new built of 2.0.14.1 with Export all images command. It took 26 minutes from start to upload, but implementation is quick n' dirty script-like. I may improve that later.
Some simple estimation might be good enough. More accurate calculation (Wheeler or Maxwell) by mechanical dimensions could be too complex for arithmetic expressions of library blocks.
Do not mind too much about my donation comments. VCAD is freeware (not donationware) for many good reasons so especially non-productive cosmetic features are not for sale.
PS. Export all images function is now improved.
