VituixCAD v2

[kimmosto]

2.0.108.0 (2023-08-19)

Enclosure

• Added Solve button to Vent Rear 1 group.
• Added bass reflex parameters Qp and End corr to radiator solver window. Window title indicates radiator type; Passive solver or Bass reflex solver.

---

[Reet]

Glad to see you are making progress

[kimmosto]

Java studies are still alive

[Reet]

Use the last option.

[DaveFred]

Hello everyone,

How would a box like this,



be modeled in the VCad diffraction simulator?

- Box size is 542mm high, 345mm wide,
- Taper back from baffle is 30º starting at 93mm wide at the top tapering down to 27mm at the bottom
- Driver is RS225-8, SD=213.8cm2, centred horizontally, 235mm centre driver to bottom edge

Here is the sim of the box with it just being a square box,



If we run with the premise that the bevel is the front edge of the box, using the bevel width at the middle of the driver to be the box width, this is the sim,



And lastly, if we call the bevel a big roundover, using the width of the bevel at the middle of the driver as the roundover on the box, this is the sim,


And lastly, here is all three overlaid,


Which is the best choice, or should I be doing something else?

Thank you,

David.

[Reet]

A bit off topic but I’ve always wondered why some manufacturers choose oddball sample rates for DSP. What’s wrong with 96kHz for example? I just hope there is good sample rate conversion process implemented as I’ve seen both good and awful examples of that.

[kimmosto]

2.0.107.0 (2023-07-04)

Main

• Sample rate with Hypex DSP changed to 93750 Hz to be compatible with FA and HFD 5.0.

Impulse

• Added 93750 Hz to Sample rate list box for FIR filter with Hypex FA.
• Taps limited between 1500 and 4500 with Hypex FA and 93750 Hz sample rate.
• Minimum FFT length limited to 16384 with Hypex FA and 93750 Hz sample rate to support 4500 input taps.

---

[Reet]

Thanks Kimmo, that all makes sense. The problem appears to have resolved itself after messing with all the settings. I am less concerned about the specific image dimensions, more so that the chart aspect is as selected in the options. It all appears to be doing what I expect now, sorry to have bothered you with it.

[kimmosto]

Text in user manual is not entirely all-inclusive so some details are not explained.

Image height can be calculated from image width and Aspect ratio. Aspect ratio options are Full, 4:3, 3:2, 16:9 and 10-50 dB/dec. Options in dB/dec use also Frequency axis and SPL/Directivity span settings. Default value of aspect ratio is 25 dB/dec, especially for CTA-2034-A. Default values of W x H can be set by double-clicking the label in front of text box. Settings apply to zoomed chart of six-pack, and copied and exported charts with the following rules:

1. Dimensions of zoomed, copied and exported image is calculated from aspect ratio and actual Y axis span and frequency range limits if aspect ratio is dB/dec and axis units are dB and Hz. Size is limited up to 4000x3000 px.
2. Zoomed chart is 16:9 if aspect ratio is dB/dec, but axis units are not dB and Hz.
3. Zoomed chart fills whole frame if aspect ratio is Full.
4. Zoomed chart follows aspect ratio settings 4:3, 3:2 and 16:9.
   Note! Copy and Export image use W x H size settings in cases 2-4.

1. Image height setting in Options is irrelevant (ignored) in Copy/Export when Aspect ratio setting is any dB/dec and Y unit is dB and X unit is Hz.
Adjusting image height in Options does not modify image width in order to keep user's width setting as is for charts which don't have dB and Hz units or Aspect ratio is not dB/dec. For example GD & phase, Impedance, Power, Excursion, Velocity and Impulse/ETC/Step.

So all this is actually quite complex due to different units and aspect ratio settings. Aspect ratio categories (Full, X:Y and dB/dec) do not act equally in all cases. Easiest is to adjust preferred image width(s) in px and Aspect ratio in dB/dec and exit with OK.

[Reet]

Above comment from investigation on native Windows 11 installation.

Running the same version of VituixCAD in Windows 10 VM, it is producing the 720 x 473 size with 25dB/dec selection and bad export aspect shown, where win11 used 720 x 385 and exported with the correct aspect. Very odd.

[Reet]

I can confirm 100% scale is used on my end. Tested again today, I think the problem lies in the image export settings. 25dB/dec image size should have been 720 x 385, not 720 x 473 = 20dB/dec. I tested and found changing aspect ratio or image width will automatically adjust the height accordingly, however adjusting the height value does not automatically update the width to conform to the selected aspect ratio. It seems there are unpredictable results if the image size does not match the selected aspect ratio. In my case, the image export kept the image dimensions and produced 20dB/dec aspect, while in your case the exported image is actually 818px wide to keep the selected aspect.

In any case, when image size matches the selected aspect, image export is what it should be.

[kimmosto]

This is not probable reason but screen scaling might have possibility to screw up some individual graphics operations though scaling should work on Win 10 and Win 11 with .NET 4.5-4.8. I always run with 100% scaling to avoid problems.
This setting is primarily for Visual Studio IDE, but it's okay also for VCAD with 4K display:

[kimmosto]

^Somehow I could not repeat that. Both 'Copy image' and 'Export image' commands seem to work here. Legend should not affect to aspect ratio because it is merged as a separate image to the right side of the chart, and total width is sum of separate parts.

[Reet]

Hi kimmosto, individual image export appears to be squishing the chart horizontally a bit from the expected 25dB/dec output. See example below, due to the legend the chart appears to be condensed horizontally, so only about 19dB/dec scale.

[kimmosto]

2.0.106.2 (2023-06-07)

Main

• All-pass responses with t.racks 96k DSP option adapted for FIR DSP 408.

Simulated phase response of 2nd order all-pass matches with measurement within Q=0.5-2. Small difference between t.racks filter designer and actual response exists so VituixCAD shows more realistic result.

[kimmosto]

I experimented. I had many more.

100 rounds would not change anything.

[Nil L]

after 10 rounds

I experimented. I had many more.

[kimmosto]

When the optimizer or user changes the value of the inductor back and forth for a long time, it eventually turns out that the resistance of the inductor has become unrealistic, for example, 1.2 mH has become 9 ohms

I also tested this claim because it sounds really wrong. Values in the beginning: L=1.0 mH, DCR=0.500 Ohms, diameter 1.0 mm.
DCR is 0.499 Ohms after 10 rounds (with mouse wheel or up/down arrow keys) down to 0.1 mH, up to 10 mH and back to 1.0 mH. So drifting is really insignificant if inductance and resistance values stay within practical limits where number of decimals is adequate i.e. controlled by the program. Automatic rounding can use max 7 decimals so in some point quite close to 0.0 mH DCR estimation will start to drift. As already said, optimizer does not touch DCR if not selected for optimization.
So wrong conclusion must be some mistake or accidental change in gauge or adjusting too close to zero mH where 7 decimals for DCR is not adequate anymore.

[JonMarsh]

Majorly useful addition.

[JonMarsh]

I like it as-is, but I'm just one guy... I set it up initially for the inductor series and winding style I plan to use- it tracks OK most of the time. But of course, it gets tweaked and locked when the final part is chosen.

[Nil L]

Remove the dependence of the resistance of the inductor on the value of the inductance? The resistance of the inductor too much influences the fitting of the elements to obtain the best frequency response and phase. This would be a backward movement. I didn't use the option to set max and min for DCR, but this is a ok solution. Thank you.

[kimmosto]

When the optimizer or user changes the value of the inductor back and forth for a long time, it eventually turns out that the resistance of the inductor has become unrealistic

Optimizer does not change DCR by default. Optimizer can change it, but not before you have checked Opt cell manually.
You can lock DCR between Min and Max to avoid exceeding possible range. This is one cure also to drifting due to rounding. Unfortunately calculation has to be relative because different coil cores have totally different DCR range with the same L and wire gauge. Calculation remains as is, but I will remove whole feature if too much complains. Personally I don't like or need it.

[Nil L]

That gives smaller error with target given by user, and target does include level. Or max gain constraint is checked while target includes requirement to extend bass range. Several possible reasons.
Try to limit number of optimized parameters to increase your own control to optimizing process and awareness what the consequences could could be if optimized parameters affect outsize target's range.

I played around with 'seek level' and 'weight' and got a decent result. Thanks for the advice.

When the optimizer or user changes the value of the inductor back and forth for a long time, it eventually turns out that the resistance of the inductor has become unrealistic, for example, 1.2 mH has become 9 ohms. If this is due to the accumulation of rounding errors, is it possible to add a check to ensure that the obtained value of the inductance and wire thickness correspond to the standard resistance value for them?

[kimmosto]

Purifi Audio drivers added to online database

PTT10.0X04-NAB-02
PTT4.0M04-NAC-03
PTT4.0M04-NAC-04
PTT4.0M04-NFC-01
PTT4.0M04-NLC-02
PTT4.0M08-NAC-03
PTT4.0M08-NAC-04
PTT4.0M08-NFC-01
PTT4.0M08-NLC-02
PTT4.0PR-NA2-02
PTT4.0PR-NF2-01
PTT4.0PR-NL2-01
PTT4.0X04-NAC-03
PTT4.0X04-NAC-04
PTT4.0X04-NFC-01
PTT4.0X04-NLC-02
PTT4.0X08-NAC-03
PTT4.0X08-NAC-04
PTT4.0X08-NFC-01
PTT4.0X08-NLC-02
PTT5.25M04-NFA-02
PTT5.25PR-NF2-01
PTT5.25X04-NFA-01
PTT5.25X04-NFA-02
PTT5.25X08-NFA-01
PTT5.25X08-NFA-02
PTT6.5M04-NFA-01
PTT6.5M04-NLA-02
PTT6.5M08-NAA-07
PTT6.5M08-NAA-08
PTT6.5M08-NFA-01
PTT6.5M08-NLA-02
PTT6.5PR-NA2-03
PTT6.5PR-NF1-01
PTT6.5PR-NL1-01
PTT6.5W04-NFA-01
PTT6.5W04-NLA-10
PTT6.5W08-NAA-07
PTT6.5W08-NAA-08
PTT6.5W08-NFA-01
PTT6.5W08-NLA-10
PTT6.5X04-NAA-07
PTT6.5X04-NAA-08
PTT6.5X04-NFA-01
PTT6.5X04-NLA-10
PTT6.5X04-NLB-02
PTT6.5X08-NAA-07
PTT6.5X08-NAA-08
PTT6.5X08-NFA-01
PTT6.5X08-NLA-10
PTT6.5X08-NLB-02
PTT8.0PR-NA2-01
PTT8.0X04-NAB-02
PTT8.0X08-NAB-02

[kimmosto]

2.0.106.1 (2023-05-18)

Main

• Blue color in rainbow palette changed to lighter (RGB=80,80,255) to be more visible in dark mode as line chart.
• Curve legends of exported/copied Directivity charts are not sorted aphabetically to maintain correct angle order.

This more official release. 80,80 does not change much compared to 76,76. It's just average of 64 and 96, and divisible by 16. Nerd numbering...

[Reet]

It's a bit better, thank you. At least for symmetrical cabinets, I can always select negative angles to use red and green for the more interesting traces for even greater visibility.

[kimmosto]

2.0.106.1 has blue=76,76,255. I can tune it more if needed.

[kimmosto]

I'm trying to find setting which would be the same and okay for both modes and all chart types. For example Blue=76,76,255 doesn't look too lame as waterfall

[Reet]

64 or 96 options are my preference, probably 96,96,255 is best. All are an improvement to visibility.

[kimmosto]

This is Blue=96,96,255 with 1px line width

[kimmosto]

What is the best?

Blue=0,0,255:

Blue=64,64,255:

Blue=96,96,255:

Blue=128,128,255:


Lighter blue might work also in normal mode.
Sorting of legends does not look good with Directivity Line charts I haven't noticed that before because I don't look directivity as line chart.

[Reet]

Is there any chance of lightening a few of the lines for the directivity line chart in dark mode? The dark blue lines on black background are quite hard to make out.

[Reet]

Not really my taste of music, but I enjoyed the artwork. It's too bad they didn't use and X in his name to make him cooler

[kimmosto]

Polite translation of previous. Replace "the Goofy Guy" with "F*ck up Man" to get more honest version.
https://lyricstranslate.com/en/vitui...goofy-guy.html

[kimmosto]

https://open.spotify.com/album/1YAvzSeSWxrqs5NszHfoga?si=9I7rO2o1TxmBYFHm-FvuZA&context=spotify%3Aalbum%3A1YAvzSeSWxrqs5NszHfoga

[kimmosto]

Sometimes the VituixCad optimizer decides to lower the overall level. Sometimes very strong, by 20-30 dB. What reasons can lead to such behavior?

That gives smaller error with target given by user, and target does include level. Or max gain constraint is checked while target includes requirement to extend bass range. Several possible reasons.
Try to limit number of optimized parameters to increase your own control to optimizing process and awareness what the consequences could could be if optimized parameters affect outsize target's range.

[kimmosto]

2.0.106.0 (2023-05-14)

Main

• Rotation R and Tilt T parameters added to Driver offset group in Room tab.
• Planes rotation range in Driver offset group extended back to ±90 deg.
• Eg, Tg and Rg changed back to global variables to allow changes in Driver offset group without jumping back to default values. Bug since 2.0.104.0 (2023-05-02).

[Nil L]

Sometimes the VituixCad optimizer decides to lower the overall level. Sometimes very strong, by 20-30 dB. What reasons can lead to such behavior?

[kimmosto]

Also, send me a message here if you want another hint regarding the DAC you asked about.

DAC deal went to local dealer. Antti will ask directly from sales dep. are they willing to sell two or just one for the lowest price. I guess he could pay a bit more if the lowest offer was primarily personal. Thanks anyway!

[Quza]

2.0.105.0 (2023-05-04)Main

• Added t.amp 48k to DSP system list in Options. Plays t.amp Quadro 500 DSP at the moment.

I'll test the 4x4 later today and the 204/206/408 series tomorrow. Also, send me a message here if you want another hint regarding the DAC you asked about.

[kimmosto]

2.0.105.1 (2023-05-08)

Enclosure

• Driver database can also be XML file with either .vxd or .xml extension.

[kimmosto]

2.0.105.0 (2023-05-04)

Main

• Added t.amp 48k to DSP system list in Options. Plays t.amp Quadro 500 DSP at the moment.

[JonMarsh]

This was an interesting change for which I enjoy the added control.

[kimmosto]

2.0.104.1 (2023-05-03)
* Impulse response window did not export anything due to modified filename extension filters. Bug since 2.0.101.5 (2023-04-21)

[kimmosto]

2.0.104.0 (2023-05-02)

Main

• Added t.racks 48k and t.racks 96k to DSP system list in Options. t.racks 48k should be okay for t.amp Quadro 500 DSP too.
• Added Tg[us] parameter to Generator component for time shifting of all acoustical responses.
• Smoothing of all drivers can be adjusted at once by pressing Alt-key while selecting from the list box.

Merger

• Minimum transition frequency dropped from 40 to 20 Hz.