The Spin-O-Rama VCAD feature is the best thing that has happened to me in my study's the last two years.
Its a lot to take in and learn and I am more than willing to do that, I just want to say thank you for opening up my eye's to a new segment of measurements.
Its going to make my life simpler once I have dialed in the techniques involved.
Its a lot to take in and learn and I am more than willing to do that, I just want to say thank you for opening up my eye's to a new segment of measurements.
Its going to make my life simpler once I have dialed in the techniques involved.
Rev. 2.0.33.2 (2019-12-04)
Convert IR to FR
* Added Z/U button to set default values for impedance, voltage and acceleration exports. Replaces automatic setting added to rev. 2.0.32.0 (2019-11-28). Manual operation is more convenient if 'Microphone Used On' was accidentally unchecked while acoustical measurements with ARTA.
Convert IR to FR
* Added Z/U button to set default values for impedance, voltage and acceleration exports. Replaces automatic setting added to rev. 2.0.32.0 (2019-11-28). Manual operation is more convenient if 'Microphone Used On' was accidentally unchecked while acoustical measurements with ARTA.
I've finally decided to try out VCAD. I've used other design programs, but the feature list for VCAD keeps growing. I read the "VCAD Measurement Preparations for REW" and I really only have one question:
The .pdf states "Rotation center is typically on surface of front baffle for the tweeter". My manual table is somewhat small so usually I center the speaker on the table. This results with the center of rotation being at the center of the speaker, not at the surface of the front baffle.
What is the recommended solution? I can think of a couple (e.g. bigger table), but just want to make I sure I'm understanding rotation correctly.
Thanks....
The .pdf states "Rotation center is typically on surface of front baffle for the tweeter". My manual table is somewhat small so usually I center the speaker on the table. This results with the center of rotation being at the center of the speaker, not at the surface of the front baffle.
What is the recommended solution? I can think of a couple (e.g. bigger table), but just want to make I sure I'm understanding rotation correctly.
Thanks....
If you rotate around speaker center then the distance from the speaker dome/cone to the microphone will not remain constant and that will be a problem the bigger this offset is. I solved this also by building a bigger turntable (got pretty big diameter lazy susan bearings from Alibaba with a reasonable price).
^^Bigger table works ... until you have much bigger speaker or need to measure vertical plane of rectangular ribbon/planar without possibility to just rotate the driver. Might need some temporary baffle for measurements only.
Rotating top plate does not have to be circular and axis hole not in the center. Square 80x80cm should be quite adequate. Axis hole e.g. 20cm from the edge to allow tilted or stepped front baffle or spikes in front of speaker etc.
Base plate could be circular, diameter close to 100cm and axis hole in the center.
Some kind of thin rod with four legs for distance measurement could be nice because that is the most difficult phase due to accuracy requirement. With measurement tape it's easiest to rotate table +90 deg, align mic with front baffle and adjust distance from mic to the center of DUT to 1000 mm. Then rotate to -90 and verify alignment and distance 1000 mm i.e. system rotates vertically without swinging.
Rotating top plate does not have to be circular and axis hole not in the center. Square 80x80cm should be quite adequate. Axis hole e.g. 20cm from the edge to allow tilted or stepped front baffle or spikes in front of speaker etc.
Base plate could be circular, diameter close to 100cm and axis hole in the center.
Some kind of thin rod with four legs for distance measurement could be nice because that is the most difficult phase due to accuracy requirement. With measurement tape it's easiest to rotate table +90 deg, align mic with front baffle and adjust distance from mic to the center of DUT to 1000 mm. Then rotate to -90 and verify alignment and distance 1000 mm i.e. system rotates vertically without swinging.
Last edited:
Hi Kimmo,
would like to ask whether your program can take FR from measured elements, let say a bass and a tweeter element for a 2 way speaker project, and then in the VCAD program adjust the filter component values in real time by dragging with mouse cursor by grabbing the FR lines to our liking, is it possible to adjust FR in this way with VCAD?
I am not very good in conveying the above in words from what I have in mind, but I have a short animation mp4 file which shows what I mean, since it's few months ago I found it unfortunately the memory escapes me what that SW name is since I can't find the thread here on DiyA.
If VCAD can not for the moment perform such operation, may there be perhaps any future plans for such a feature?
Thanks and Regards, Michael
would like to ask whether your program can take FR from measured elements, let say a bass and a tweeter element for a 2 way speaker project, and then in the VCAD program adjust the filter component values in real time by dragging with mouse cursor by grabbing the FR lines to our liking, is it possible to adjust FR in this way with VCAD?
I am not very good in conveying the above in words from what I have in mind, but I have a short animation mp4 file which shows what I mean, since it's few months ago I found it unfortunately the memory escapes me what that SW name is since I can't find the thread here on DiyA.
If VCAD can not for the moment perform such operation, may there be perhaps any future plans for such a feature?
Thanks and Regards, Michael
^VituixCAD has optimizer able to iterate component values to reach axial and power response target lines or driver's acoustical response or filter's electrical response target curves, but it's not so smart and fast than software in your video. Just simple multi-variable hill climbing algorithm which tests and selects values which improve the result i.e. decrease difference compared to targets.
I don't have resources to change algorithm much smarter. In addition, I design usually in manual mode without optimizer. That works as human learning too.
I don't have resources to change algorithm much smarter. In addition, I design usually in manual mode without optimizer. That works as human learning too.
Rev. 2.0.33.3 (2019-12-07)
Enclosure
* Added 'Open recent enclosure project' button (arrow down) to the right side of 'Open enclosure project' button. Max 20 recent vxe filenames are saved to the list.
* Buttons and checkbox on the right renamed and moved up to make more room for driver table.
Enclosure
* Added 'Open recent enclosure project' button (arrow down) to the right side of 'Open enclosure project' button. Max 20 recent vxe filenames are saved to the list.
* Buttons and checkbox on the right renamed and moved up to make more room for driver table.
^^ Thanks, as soon as I get a new computer with Windows (currently running Linux) I will be happy trying out what VCAD can offer.
If I may bring forth my objectives for improving VACD according to the aforementioned feature in previous post, it would be for instance beneficial for people with limited financial means as buying and trying out a lot of filter components for loudspeaker projects can quickly grow considerably on the financial burden, I also think it may help having an intuitively "visual" concept in form of adjustable frequency response translated into real component values, some people are very visually minded but "blind" when they have mathematical numbers before their eyes while other are of the opposite nature, ie. there's no right and wrong from a pedagogical point of view.
However, don't feel pushed, it's merely a wish, hopefully this now planted seed may sprout maybe some times later in the future, or perhaps already for Christmas, after all Santa Claus is from Finland!
If I may bring forth my objectives for improving VACD according to the aforementioned feature in previous post, it would be for instance beneficial for people with limited financial means as buying and trying out a lot of filter components for loudspeaker projects can quickly grow considerably on the financial burden, I also think it may help having an intuitively "visual" concept in form of adjustable frequency response translated into real component values, some people are very visually minded but "blind" when they have mathematical numbers before their eyes while other are of the opposite nature, ie. there's no right and wrong from a pedagogical point of view.
However, don't feel pushed, it's merely a wish, hopefully this now planted seed may sprout maybe some times later in the future, or perhaps already for Christmas, after all Santa Claus is from Finland!
Last edited:
Existing features already provides that. I have designed speakers ready at once without listening a second (already with version 1), and made totally different XO versions without having speakers at home anymore. That has nothing to do with algorithm (iteration/learning) optimizer uses and how it starts; by button or automatically when target is moved or value of some non-optimized component is modified. It's "just" combination of information in measurements, graphs and designer's understanding how they should be with selected acoustical concept and components. Certain UI for value setting does not provide that. Just compatibility and speed with UI style one could prefer.
For example XMachina has machine learning features, and it works on Linux too. That could provide some features you prefer, but something relevant could also be missing.
Last edited:
Main problem at the moment is how to motivate users to measure 0-90 deg or 0-180 deg off-axis, or limit simulated sector with program so that all simulated directions are covered by measurement data.
Now it is possible that the biggest off-axis angle of some drivers is dropped out of simulation if driver is not at the speaker's origin (X,Y <> 0,0). Program does not calculate response if driver is located so that exit angle from driver to virtual microphone exceeds the biggest measured off-axis angle, because response cannot be interpolated from measurement data. That causes error also to power response and DI.
For example if X or Y is -200 mm and listening distance is 2000 mm, we would need off-axis measurement to 49 deg to simulate 45 deg:
Only working method at the moment is to measure full 0-180 deg, or 0-90 deg and verify that drivers' Z >= 0 mm which ensures that virtual mic does not go to unmeasured shadow sector while simulating responses within angle limits of measured sector.
Now it is possible that the biggest off-axis angle of some drivers is dropped out of simulation if driver is not at the speaker's origin (X,Y <> 0,0). Program does not calculate response if driver is located so that exit angle from driver to virtual microphone exceeds the biggest measured off-axis angle, because response cannot be interpolated from measurement data. That causes error also to power response and DI.
For example if X or Y is -200 mm and listening distance is 2000 mm, we would need off-axis measurement to 49 deg to simulate 45 deg:
Only working method at the moment is to measure full 0-180 deg, or 0-90 deg and verify that drivers' Z >= 0 mm which ensures that virtual mic does not go to unmeasured shadow sector while simulating responses within angle limits of measured sector.
Hi, admit myself is on Windows and know not much about Linux, but i wonder if you noticed it looks VituixCAD run on Linux with Wine for many other members and there's a pdf file available on the website as seen in below visual highlighted with yellow mark.
Link: Software
Attachments
Then all loaded responses within range specified in Options window are included in simulation and visible in directivity graph as they should be.
Logic needed for interpolation is already complex, and adding some workarounds with assumptions would make it even more complex and impossible to estimate how much error is created.
So the latest build of rev. 2.0.33.3 (2019-12-09) is done like this:
* Off-axis direction is not included in simulation if measurement data of any driver does not cover angle required for response interpolation.
Responses visible in directivity graph are now okay, and all responses included in power, listening window, early reflections, DI and early DI contain all drivers in the crossover.
But few responses typically in vertical plane could be missing if measurement data is not full 0-180 deg or 0-90 deg when half space is selected. This is easy to test with typical 3-way by loading responses within 0-50 deg only to any driver, and selecting 5 deg angle step in Options. 1-2 responses to positive vertical angles are not visible anymore, but 1-2 extra (-55 deg) responses could be visible to negative angles.
Are you asking why frequency responses to simulated directions are interpolated from measurement data and not used as is?