Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Software Tools SPICE, PCB CAD, speaker design and measurement software, calculators

Software to Predict Directivity
Software to Predict Directivity
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 24th October 2017, 01:44 AM   #11
evilnui is offline evilnui
diyAudio Member
 
Join Date: May 2016
Another solution would be to do a ABEC direct sound simulation of the desired drivers at their distance and import the result in VituixCad for the actual crossover...

Problem would be that the former does not export the data in a form the latter understands. Luckily, it might be easy enough to write a script to convert the export into something importable, since its also relatively simple txt files.
  Reply With Quote
Old 24th October 2017, 10:01 PM   #12
evilnui is offline evilnui
diyAudio Member
 
Join Date: May 2016
Simulated polar responses of single 1,2,4,8,15 inch drivers. Exported those as txt files from VacsViewer. Converted txt files using python into several that VituixCAD was happy to read. I have no idea if the phase response was correctly exported/imported.
In VituixCAD you can then use the response for a pair of distanced drivers. I tried to enter the active filters as seen in a table and graph here: http://www.linkwitzlab.com/Horbach-K...t%202%20V4.pdf

Here is my result:

Click the image to open in full size.

Some of the values are only eyeballed, but it looks quite similar I think. You might notice that I failed to implement the correct filter for the tweeter. If I could use the sum of 2 paths of active filters I could easily do it correctly, but I dont think I can and I dont see another solution.

Seems to do what it is supposed to, which might also indicate that the phase wasn't too bad? One probably shouldn't read too much into the rear response, as that wasn't even simulated anywhere. Unlike their vertical version, a horizontal 5-way would be a bit ridiculous due to its size. Then again it only looses directivity control at around 100 Hz, so one could probably easily scale it down to less absurd dimensions without much of a loss?
On the upper frequency end, control is worse than a speaker with a waveguide. But as the authors say a tweeter with its own control could be used. I wonder if a tweeter with a waveguide would fit well between the next pair of drivers?

Patrick, what are your ideas for this anyway?
I could imagine a synergy horn as small as can be and woofers for low end response at the sides for power and control.

I have attached my folder with the VituixCAD project (which should load with all settings resulting in the above), a python script (I used a jupyter notebook so no cmdline parameters), and the abec project (just a reduced version from the one before)

Hope it helps
Attached Files
File Type: zip hornbachkeel5way.zip (503.2 KB, 7 views)
  Reply With Quote
Old 25th October 2017, 04:43 AM   #13
kimmosto is offline kimmosto  Finland
diyAudio Member
 
Join Date: Oct 2005
Location: Kuopio
^You should remove excess phase from the measurements. Adjust timing of measurements (below frequency responses in Drivers tab) to negative value giving close to minimum phase responses.

Another option is to use Calculator tool. You can convert all responses to Minimum phase or set constant negative time offset before loading to main program.
  Reply With Quote
Old 25th October 2017, 04:50 AM   #14
kimmosto is offline kimmosto  Finland
diyAudio Member
 
Join Date: Oct 2005
Location: Kuopio
Quote:
Originally Posted by Patrick Bateman View Post
4) VituixCad - I had really high hopes for this. As far as I can see, it can display the measurements that you make, but I don't see any way for it to simulate the performance of a driver based on it's diameter. IE, it can sim arrays as long as you actually build and measure the array first.
This is answered with video lessons in VituixCAD thread. Several options for directivity estimation is available, though radiator is calculated as ideal piston only. Calculator tool enables for example mixing with flat response if BesselJ1 function is too strong.
  Reply With Quote
Old 25th October 2017, 01:14 PM   #15
evilnui is offline evilnui
diyAudio Member
 
Join Date: May 2016
Quote:
Originally Posted by kimmosto View Post
^You should remove excess phase from the measurements. Adjust timing of measurements (below frequency responses in Drivers tab) to negative value giving close to minimum phase responses.
Oh now that looks a lot better and now it might be easier to dial in the other parameters better.
Click the image to open in full size.


Thank you very much. Also for the software
Since you appear to be reading this. Is it possible to construct the red filter in VituixCAD? The right part, where its just a sudden constant 0dB is what I fail to reproduce. With the sum of 2 active filter paths I see how I could do it.
Click the image to open in full size.
  Reply With Quote
Old 25th October 2017, 01:54 PM   #16
TNT is offline TNT  Sweden
diyAudio Member
 
Join Date: Apr 2003
Location: Sweden
If you use "Edge" and place the "mic" in a number of strategi locations, you will be able to get directivity information.

Tolvan Data


Oops - the was a dedicated app for this, XDIR:

Tolvan Data



//
__________________
More distortion to the people!
  Reply With Quote
Old 25th October 2017, 02:03 PM   #17
kimmosto is offline kimmosto  Finland
diyAudio Member
 
Join Date: Oct 2005
Location: Kuopio
Quote:
Originally Posted by evilnui View Post
Is it possible to construct the red filter in VituixCAD?
Yes with Calculator tool and Transfer function block, but x/o frequency can't be adjusted without creating new filter response.
Create low pass filter section with H-K MTM and brickwall blocks. Export Filter gain of driver to text file. Load exported response to B response in Calculator tool. Load flat 0dB response to A list. Select Subtract A-B, select destination directory and save the result with Calculate&Save button. Add Transfer function block for high pass and load calculation result. Not so nice but works for occasional testing.

Looks like this
Attached Images
File Type: png Horbach-Brick.PNG (50.8 KB, 60 views)
  Reply With Quote
Old 25th October 2017, 04:16 PM   #18
evilnui is offline evilnui
diyAudio Member
 
Join Date: May 2016
Ah I see. I dont think I understood you correctly. I just constructed the high- and lowpass section for the tweeter individually with brickwall filters for both, exported them, loaded them as A and B, summed with A+B.
This leaves a beauty mark though, as the resultant filter shows some interpolation between the brickwalls leaving a short slope. Your example has it, too.
While this solution perhaps has too many steps (ideally speaking), the ability to bypass crossover blocks and the fact that the main and calculator windows are independend of each other, let this solution remain pretty fast actually

Unfortunately linkwitzlab appears to be down, so I cant see the paper to try to hone in on the given parameters. Oh well.

Thank you very much
  Reply With Quote
Old 25th October 2017, 05:00 PM   #19
kimmosto is offline kimmosto  Finland
diyAudio Member
 
Join Date: Oct 2005
Location: Kuopio
^VituixCAD has 1/48 oct steps in frequency axis. Perfect brickwall is not possible to show (and convert from frequency to time domain).

I have both AES papers (Part I & II) of Horbach & Keele. Authors wrote below small graph you attached:
Quote:
The dashed lines represent one possible extension of the responses that serves to smooth the sudden transition that occurs at this frequency to simplify filter implementation.
Text does not point out that FIR needs much more taps for accurate IR if there is brickwall at fNx3. Therefore this filter is implemented in VituixCAD with proposed straight ramp from fNx3 to fNx6. Adjustable R parameter would be possible to add but I think that won't make it better.
  Reply With Quote
Old 25th October 2017, 05:37 PM   #20
evilnui is offline evilnui
diyAudio Member
 
Join Date: May 2016
I tried the existing implementation and it indeed smooths the response. The brickwall one really could only serve to preserve the nonfitting parts I see above. Thanks again
  Reply With Quote

Reply


Software to Predict DirectivityHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Predict this Toggle Circuits behavior imix500 Parts 58 5th September 2006 02:40 AM
What do you predict the Mackie SRM350 will be like? rintincop2 Multi-Way 2 25th September 2004 07:46 AM
How to predict PSRR ? Kilentra Solid State 6 9th July 2004 07:31 PM


New To Site? Need Help?

All times are GMT. The time now is 06:19 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki