VituixCAD v2

[kimmosto]

Modifications around spinorama are not far from parody anymore. Available written references have been 'Sound reproduction' by Toole and CTA-2034-A final. Both are quite simplified and indefinite so at least Klippel, Pierre of spinorama.org and I have made own inquiries, interpretations and comparisons.

One of the first and biggest problems was ER Rear calculation due to wrong interpretation what "Hor 180 deg, +/-90 deg" means. It's whole rear sector in hor plane, but some of us interpreted it wrong. For example, mistake in Klippel affected years to ER Horizontal, ER Total, PIR and ERDI results published on ASR and everywhere else. Most of the reviews still have wrong results visible. The latest reviews look okay, but I have not verified very carefully.

Another PITA is power response / sound power calculation. CTA decided to provide their own weighting factors which deviate from textbook values and calculation. Table for 10 deg angle steps is okay but 5 deg is totally screwed up. Anyway, CTA specifies:
"The sound power is the weighted rms average of all 70 measurements, with individual measurements weighted according to the portion of the spherical surface that they represent. Calculation of the sound power curve begins with a conversion from SPL to pressure, a scalar magnitude. The individual measures of sound pressure are then weighted according to the values shown in Appendix C and an energy average (rms) is calculated using the weighted values. The final average is converted to SPL."
BUT Klippel, spinorama.org and VituixCAD with CTA-2034-A power weights checked do not use that order (as I read it). All three calculate: pressure squared -> weighting -> summing -> square root -> to SPL. VituixCAD does not force to use CTA's factors and formulas so I'm the worst skeptic allowing traditional textbook calculation (by unchecking CTA-2034-A power weights). Difference is not huge though.

The latest disappointment was to notice that VituixCAD is the only app. without squaring LW, ER and SP before weighting for PIR calculation. Squaring is not specified in CTA. PIR is not "official" curve in spinorama so maybe it's okay to skip. Toole has written that spec. to his book. I found it after comparison to Klippel and spinorama.org so there was small error for quite many years.

[kimmosto]

2.0.118.3 (2024-10-18)

Main

• CTA-2034-A button in Options does not change SPL/Directivity span, Aspect ratio of exported images and Angle step if 5 or 10 deg. Other angle steps are changed to 10 deg.
• Default width of exported graphs changed to 640 px.

Sorry that I continue to undo old modifications requested by some users. Graphical recommendations such as 50 dB span of spinorama and aspect ratio of 25 dB/dec in SPL charts presumably support publishing of CTA-2034 and other SPL graphs, but they don't support everyday design work. That is priority #1 so they won't be defaults anymore.
Klippel owners calculate CEA 2034 with 10 deg angle steps, but for example I usually run Outline ET with 5 deg steps and use the same resolution for calculations. So CTA-2034-A button in Options won't manipulate settings which are fully acceptable for designing and probably give more accurate result also in spinorama.

[Reet]

I understand a need for accuracy and consistency among multiple systems that are claiming to produce the same charts. To understand the context of this change of calculation, are we talking about something significant, that would impact design decisions, or just some "hair splitting" of small errors / differences.

[Reet]

Great, another unapproved post, for no reason this time. Love this forum...

[kimmosto]

I understand a need for accuracy and consistency among multiple systems that are claiming to produce the same charts. To understand the context of this change of calculation, are we talking about something significant, that would impact design decisions, or just some "hair splitting" of small errors / differences.

Small differences do not probably have any effect to project which is already designed mechanically and acoustically and designer is already tuning XO.

In practice, slope of directivity responses is as important as smoothness - especially for casual listening. Correct slope results enable one consistent feedback mechanism to explain and learn why certain speakers sound bad and some others good in the same environment and locations.
This looks weak point in Toole's and Olive's studies because listeners were probably nailed to the spot. Narrow band deviations and smoothness got high significance and slopes none. This probably reflects also to results and conclusions by reviewers and preference rating enthusiasts who rely (too) much on narrow scoped studies. Boomy or/and muffled speakers might be praised due to biased interpretation of spinorama.

So all this is kinda part of long term plan to turn twisted or narrow scope of objective evaluation, and encourage to select acoustical concepts which have possibilities to produce both shallow and linear enough directivity slopes.

[kimmosto]

2.0.118.4 (2024-10-25)

Enclosure

• Align Closed gives warning "Increase Qa to normal range 2-120 to avoid too small volume." if Qa < 1.5.
• Minimum volume with Align Closed decreased from 0.1 to minimum value of Vb text box (0.0001).
• Align Closed calculation of 2.0.109.2 (2023-10-05) restored to get Qtc detected from impedance response closer to selected target.

61 Celestion drivers added to online database. Thanks to Simon R/Celestion for entering data.

[kimmosto]

all this is kinda part of long term plan to turn twisted or narrow scope of objective evaluation, and encourage to select acoustical concepts which have possibilities to produce both shallow and linear enough directivity slopes.

Some suggestions on this topic: https://kimmosaunisto.net/Misc/speak...w_feedback.pdf
I may not have enough courage to publish or send that as PM on ASR or EAC. Wall of Toole's and Olive's science bubble is so strong on those medias.
Any thoughts?

[tktran]

Re: 61 Celestion drivers added to online database.

Can't see it. What am I doing wrong?

[kimmosto]

Can't see it.

1234

[tktran]

Some suggestions on this topic: https://kimmosaunisto.net/Misc/speak...w_feedback.pdf
I may not have enough courage to publish or send that as PM on ASR or EAC. Wall of Toole's and Olive's science bubble is so strong on those medias.
Any thoughts?

Yes I agree with many of your points.

I came to similar thoughts about usefulness of globe charts and normalized polar maps and harmonic distortion.

In general, much of the science in Toole's book is about 20 years ago and may be represent the research at the time on the speakers of the time.

In my field, by the time something's published in a book it's already outdated. Not completely wrong but has been further refined.

So for audio, following Toole and Olive may be easier for the consumer to understand. And perhaps that is the point of it- to help consumers make a purchasing decision. But it leaves a lot to be desired/understood in the current era. Active, DSP driven speakers, small (consumer) and large (PA) that might not rate highly on the Preference scale, as seen by Spinorama.org.. perhaps that's why they've changed to 'tonality'

With science we are willing to refine/change our thinking.

Changing process is more difficult eg. ASR methodology.

[Quza]

You can just send them an email. Dr. Sean Olive answered my questions on head and torso simulators within a few minutes and was highlighting the cons of each option, even though we never spoke before. Erin Hardjson is also open to suggestions on his reviews.
And fwiw I agree and am curious on their answers.

[Quza]

The default SPL trace in the Enclosure tool is 2pi, 4pi can be shown when a diffraction response is added. I'm wondering what the result of the Merger tool is when combining NF response from cone and port with diffraction response, also 4pi? If so, can the 2pi response be calculated?

[tktran]

Merger tool can also use “Diffraction response (5-30m)”
If you omit it and select “No baffle loss” it will output 2pi

[Quza]

Oh, ok. I think I misunderstood the whole thing. 'No baffle loss' means the speaker is used with the front being an infinite plane, right?
I was thinking about using a subwoofer in free air as 4pi and using a subwoofer on the ground as 2pi, but that would actually require adding 6dB and increasing the height of the baffle by the mirror image.

[Reet]

"no baffle loss" simply means to use the low frequency data you provide as-is without altering it. It's up to you to provide meaningful data. If that data comes from the enclosure tool, it will be 2pi.

[kimmosto]

Publishing of charts on different forums has been quite random. Especially taking screenshots from zoomed charts or main window is quite disturbing. So, few missing paragraphs added to online manual:

https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_20.html#Copy_image

https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_20.html#Export_image

https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_20.html#Copy_or_Export_with_higher_pixel_count

[Reet]

No amount of words in the help file will change the way people post screenshots / images. In 2024 people are taking photos of the screen with their phone, completely absurd.

FWIW I use screenshot more often than image export simple because I usually run VituixCAD in a virtual machine, with an odd OS screen resolution, the aspect of exported images is not correct. It works okay as long as I set a "normal" resolution of the VM before opening VituixCAD, so my problem is primarily one of laziness and convenience. At least, I try to keep constant chart aspect when taking screenshots.

VituixCAD also includes a well functioning project archive button, but I've yet to see anyone use it when sending me their files.

[kimmosto]

No amount of words in the help file will change the way people post screenshots / images. In 2024 people are taking photos of the screen with their phone, completely absurd.

😁 Now we have some text to link to forums or e-mails. Few weeks ago I received a question how to add trace legends to copied/exported image. Help was missing so my bad. I was quite shocked that users don't open context menu of chart and test what happens with 'Copy image' and 'Export image' commands. They usually take screenshots from zoomed window or whole program window - or photos. Photo via WhatsApp is very fast and easy for some purposes, but PC running VCAD is usually connected to internet.

[JonMarsh]

We each have our personal working styles and environments... it does seem that a lot of folks are "lazy" about reading things like help files. That's on them.

I use VituixCAD primarily on a relatively high end configuration 2013 Mac Pro, which is setup with boot camp to boot into WIN10. Was used for a lot of company work before I retired, due to issues with the Admin special laptop I was issued. Monitor is a relatively ancient LG 43" 4K monitor, fine for technical work but not with high color fidelity. UI scaling is set to 150 or 200%. Occasionally I try higher UI scaling, but VituixCAD is scaling up font bitmaps, and they don't render well at 200%.

I use the image export function for the schematic page, but usually work with only a single results screen displayed at a time, and use the snipping tool for them. I do that mostly because of my preferences for text scaling. I must confess to being amused how often I get photos of screen displays...

this project was built in an apartment in 2019, wood working done on the patio, and very limited space for measurements...




Typical working screen setup, though scaled by the forum software.

An exported SPL image using the internal function.

[kimmosto]

Schematic can be exported as SVG file too so the same method is available for publishing schematic for HD/2K, 4K and 8K as bitmap or PNG.

[kimmosto]

2.0.119.0 (2024-11-09)

Main

• Power & DI chart renamed to CTA-2034.
  Reference angle trace renamed to On-axis/ref angle (also in SPL chart).
  In-room response renamed to Predicted In-room.
  Power response renamed to Sound Power.
  Directivity index renamed to Sound Power DI.
  Early reflection traces renamed to Front Wall Bounce, Side Wall Bounce, Rear Wall Bounce, Floor Bounce, Ceiling Bounce, Early Reflections, Early Reflections DI, Early Reflect Hor, Early Reflect Hor DI, Early Reflect Ver and Early Reflect Ver DI.
• Show Listening window, Show ER Horizontal, Show ER Vertical and Show ER Total commands removed from context menu of CTA-2034 chart.
• Added 'Show Basic' command to context menu of CTA-2034 chart.
  Shows On-axis/ref angle, Listening Window, Predicted In-room, Early Reflections, Sound Power, Sound Power DI and Early Reflections DI.
  Hides Wall, Floor and Ceiling Bounces, Early Reflect Hor, Early Reflect Hor DI, Early Reflect Ver and Early Reflect Ver DI.
• Added 'Show Early Reflections' command to context menu of CTA-2034 chart.
  Shows Front Wall Bounce, Side Wall Bounce, Rear Wall Bounce, Floor Bounce, Ceiling Bounce and Early Reflections.
  Hides On-axis/ref angle, Listening Window, Predicted In-room, Sound Power and Sound Power DI.
  Shows temporary 'Early Reflections' text in chart title.

Optimizer

• Axial text in radio buttons renamed to On-axis.
• Power & DI group renamed to Power & In-room.
• Power response radio button renamed to Sound Power.
• In-room response radio button renamed to Predicted In-room.
• Updated chart and trace names in tooltips.

So this is just cosmetics for spinorama nerds preferring names more compatible with CTA-2034 and Klippel

Result looks like this, after Show Basic command


after Show Early Reflections command


Full set of traces can be enabled via Traces... window

[Reet]

2.0.119.0 (2024-11-09)

Main[LIST][*]Power & DI chart renamed to CTA-2034.[*]Axial text in radio buttons renamed to On-axis.

I am in favour of naming changes for Predicted In-Room, etc. as this matches terminology used in preference rating system so it should reduce any confusion there. However, I'm not sure on renaming Power & DI chart to CTA-2034, since it's easy to present the chart in a manner that doesn't conform to CTA-2034, for example I use 15 deg measurement increments instead of 10 deg, or it is possible to enter <2m listening distance. Just food for thought.

I don't fully agree with renaming of axial response to on-axis in the optimizer winder either, since it is not truly an on-axis optimization. It optimizes based on SPL chart that results from reference angle chosen and driver coordinates, which is only truly on-axis if x and y coordinates are zero, and reference angle is zero as well. Axial response was a more correct term here in my opinion.

[Reet]

Another unapproved post...

[kimmosto]

However, I'm not sure on renaming Power & DI chart to CTA-2034, since it's easy to present the chart in a manner that doesn't conform to CTA-2034, for example I use 15 deg measurement increments instead of 10 deg, or it is possible to enter <2m listening distance.

I agree that previous name "Power & DI" was more generic. Listening Window, Predicted In-room, Early Reflections and it's individual reflections are power averages (RMS). Sound Power DI plays actual DI, but Early Reflections DI is just ratio of two power averages. But they all (except On-axis) are kinda power and DI responses.

CTA-2034 has sound power weighting tables for 5 and 10 deg angle steps (though 5 deg table is screwed up), but VituixCAD 1.1 and 2.0 have not limited angle step or required constant step. That freedom has been available also for Wall, Floor and Ceiling reflections, and Listening Window. All six averages are calculated to specified sector with selected angle step instead of fixed 10 deg. That serves actual target of CTA-2034 Directivity Response without being literally bureaucratic and limiting.

My plan was to help publishing in case designer wants to publish Directivity Response per CTA-2034-A. "Directivity Response" is official name in the standard, but "Directivity" chart name is already in use. Klippel has CEA2034. So both apps are still wrong though look almost the same 😀

I don't fully agree with renaming of axial response to on-axis in the optimizer winder either, since it is not truly an on-axis optimization. It optimizes based on SPL chart that results from reference angle chosen and driver coordinates

Both SPL and CTA-2034 charts show selected Ref angle as On-axis/ref angle. Most of the cases users can leave Ref angle = 0 deg hor because optimising can be done with Listening Window which is very close to 15 deg hor in practice. Of course individual 5, 10, 15 or 20 deg hor could be slightly better axial direction for optimising in some cases, but rotation of Reference angle can be "just" nice method to verify individual off-axis frequency responses using SPL chart instead of Directivity as Line chart, Waterfall or Surface chart.

Also here "On-axis" -part of the name helps to publish charts with the most common terminology. Users are probably/hopefully aware that program works just like earlier.

Optimizer window could have different names than "On-axis" because it's not related to publishing, but I'm not sure that Axial or Ref angle would be better for common citizen. Maybe...

[kimmosto]

P.S. There are still some differences in naming between CTA-2034-A and VituixCAD. Floor Bounce should be Floor Reflection, Ceiling Bounce should be Ceiling Reflection, Early reflect Hor should be Horizontal Reflections and Early reflect Ver should be Vertical Reflections. So one day I may continue to mix this soup. Making the best decisions at once has always been difficult for me.

[Reet]

All good, I can deal with whatever naming conventions, I'm just being pedantic. As long as we don't start calling it a "Mike" instead of "mic"...

[Reet]

Every post I post is requiring approval, that must be annoying for someone.

[kimmosto]

One question before making any fast and hasty changes. What would be better...the best name for Reference angle in the top of main window, Optimizer, frequency response exports, SPL chart etc? "On-axis/ref angle" is quite long and possibly indistinct for Optimizer window and user manual. One possibility is "Design axis" which is quite common and shorter than "Reference angle". Perpendicular axis starting at X,Y,Z=0,0,0 on baffle is "on-axis" so we probably wouldn't mix that to off-axis in horizontal plane called "design axis". Or do we have some other as good and common term in English.

[kimmosto]

Every post I post is requiring approval, that must be annoying for someone.

I have accepted all your recent posts. No problem, but you should have moderator rights to avoid that. JonMarsh could you arrange that?

[JonMarsh]

I'll bring it to Sven's attention- I don't have that possibility myself.

[Reet]

Hmm, I think "design axis" is more commonly used term in online discussions, I don't know that it's a more correct term, but more people may be familiar with it. Listening axis would make it consistent with listening distance option, or perhaps both should change to design axis and design distance.

[kimmosto]

Name of ref angle trace could be variable
0 deg -> On-axis


hor 10 deg -> Design axis hor 10 deg


Optimizer with Design axis

[kimmosto]

CTA-2034 uses "Reference Axis", but that case is the same as "On-axis" while measuring responses to 0 deg. "Ref axis" would be easy and short abbreviation.

[Reet]

Reference axis or Reference angle is a good and generic term. Somewhat related, in the merger tool, the high frequency portion has the angle column labelled as "axial", perhaps this should change to "angle" to be consistent with the calculator tool

[kimmosto]

2.0.119.1 (2024-11-12)

Main

• Name of On-axis response in SPL and CTA-2034 charts is 'On-axis' when Ref angle=0 deg hor. Otherwise the name is 'Ref axis X deg hor'.
• Early Reflect Hor traces renamed to Horizontal Reflections and Early Reflect Ver traces renamed to Vertical Reflections.
• Added 'Show Reflection totals' command to context menu of CTA-2034 chart.
  Shows Vertical Reflections, Vertical Reflections DI, Horizontal Reflections and Horizontal Reflections DI. Hides nothing.

Optimizer

• On-axis text in radio buttons renamed to Ref axis.

Merger

• Axial column in HF list renamed to On-axis.

[kimmosto]

2.0.119.2 (2024-11-14)

• Size of texts changes when size of graph changes. This enables HD/2K, 4K and 8K bitmap copies, screenshots and exports directly without SVG files.

[JonMarsh]

Very excellent for me!

[kimmosto]

Tiny adjustments in the latest build 2.0.119.3 are related to the latest revision of this document: https://kimmosaunisto.net/Misc/speak...w_feedback.pdf
Refresh to see the latest 2024-11-17.

[kimmosto]

2.0.120.0 (2024-11-23)

Main

• Added 'Show Data labels' command to context menu of CTA-2034 chart.
  Shows smoothness (SM) and slope in dB/oct of ON, LW, ER, PIR, SP, SPDI and ERDI @100-16000 Hz.

Main, Enclosure, Calculate T/S

• Scale of impedance phase axis changed to +120...-120 deg.

Power dissipation

• Added 'Traces...' command to context menu of chart.
• Added 'Save as overlay' command to context menu of chart.
• Added 'Open overlay...' command to context menu of chart. Drag&Drop overlay file is also supported.

Data labels of CTA-2034 chart

[kimmosto]

2.0.120.1 (2024-12-06)

Main

• Added 'Normalized' command to context menu of CTA-2034 chart.
  Shows flat frequency response to reference angle.
• Chart areas are aligned left and right.

[tktran]

I like it!

[kimmosto]

2.0.121.0 (2024-12-22)

Main

• Added 'Show Slope target zones' command to context menu of CTA-2034 chart.
  Shows recommended slope target range for ON-LW, PIR, SP and SPDI @100-16000 Hz.

Enclosure

• Fixed reading of missing minimum limits in database filter.
  Maximum was moved to minimum if minimum cell was empty.

Slope target zones depend on slope of simulated/actual directivity index with some constraints. Few examples about adaptation and constraints.

[Reet]

That's a a nice feature, I think many new users to power & DI charts will find that helpful. I am also happy to see that my own designs are falling within the recommended slope zones well enough for my liking.

[kimmosto]

2.0.121.1 (2024-12-27)

Main

• Modified constraints of slope target zones in CTA-2034 chart.

Preference rating

• On-axis linearity calculated as difference to linear regression line to support others than conventional full space concept.

Slope target limits as C# code for spinorama.org and reviewers in case they decide to show the same targets:

[Mogens20]

Hey Guys, I'm continuing to learn how to use this amazing software. After reading all the docs I can find more than once and completing a few speaker designs, I've noticed something unexpected to me.

I'll use the Epe-3W_demo_V2.vxp reference design for simplicity and since we all have access to the data. Below is the horizontal directivity of the Vifa M10MD from the Vituix file. I muted the woofer and tweeter.

I believe the <100Hz trace data deformity has been addressed in the current revision of VituixCad2 thanks to the Force To Gradient function. Or the left side IR window was too small?

This post is about the apparent directivity of the midrange between approximately 200-600Hz. Of course this can , and in this case seems to be, caused by the acoustic center of the driver not aligned with the axis of rotation. But I see this in other designs including my own and I've very careful to do my best to ensure the DUT alignment is as close as possible. This is also visible in the directivity graphs of the other drivers in this design.




Below is the manufacturer's half space frequency response for the Vifa M10MD. While the off-axis is only to 60Deg, the driver appears to be behaving like a point source in this graph up to around 600Hz.



If the DUT is properly aligned and manufacturer (or other half space) driver measurements indicate point-source-like dispersion up to a certain frequency, is it correct to expect very little low frequency directivity graph divergence in Vituix?

If DUT is misaligned with axis of rotation, should I mitigate this by shifting SPL traces (in REW for example) such that they align at the lowest displayed frequency? Should I also adjust the IR alignment accordingly?

[Reet]

The low frequency inconsistencies can be caused by SNR of the measurement, at some low frequency the measurement may be dominated by the noise floor. It can also be related to a DC offset in the impulse response, in recent versions VituixCAD has improved the DC offset compensation in its IR-FR tool considerably, reducing this possibility.

Viewing the data un-normalized makes it clearer to identify the problem.


However, you may also notice that once a bandpass filter is applied to the driver, the error in response <100Hz is not impacting to the design results.


You are correct, force to gradient is one method of correction, but more of a Band-Aid solution, of course it's best to have good clean measurements to begin with. For the driver being discussed here, I would suspect it has not been merged with a nearfield response.

[ergo]

I believe that Kimmo's Epe-3W_demo data is collected from a real small speaker box. Thus it does have "baffle step" effect included. Also Kimmo probably set the rotation axis to a tweeter acoustic center and woofer in most cases being a bit further would have a slight effect on SPL due to that.

Vifa datasheet on other hand would show a response when speaker is mounted on a big IEC standard baffle and they move the microphone instead. This lack baffle step effect.

For my own designs I've considered that at any given angle - ie 0deg / 30deg / 45deg etc, it is important that the relative levels between tweeter and woofer are correct. They will be imho very much good enough if you measure them rotating around tweeter axis. It will show you issues with power response and directivity mismatch etc if xover is not optimal.

So onto building - measuring and listening

Good luck with the project.