Klippel Near Field Scanner on a Shoestring

Jump to Latest
b_force

b_force

Member
Joined 2004
oneworldconcepts.com/
ScottG said:
Audio Interface.

ex. the excellent id24 that fluid just purchased:
Click to expand...
Unless this is a really bad audio interface, but for the purpose of time reference basically any audio interface will be fine.

Distortion isn't important for referencing time btw.

There is an argument to be made for the mic preamp. That's why I was asking which one you were talking about.
But in most cases the low pass /high pass filters are far enough to not have any significant issues.

Edit: I just realized this last part isn't true.
As soon as the signal out of the amplifier is time synced, all other things are the same in a relative way.
Unless the mic preamp contains a ADC/DSP or so.

In a multi-mic solution, it does bring up the importance of close matching of the microphones.
Or at least within the frequency range we are interested in.
Otherwise there will be time/latency = phase differences between them.
Obviously this can be easily compensated for with a simple calibration process.
So I guess they don't have to be matched SO closely.
 
Last edited:
T

Tom Kamphuys

Member
Joined 2020
mbrennwa said:
Please try using the mataa_measure_signal_response(...) function for audio I/O instead of the mataa_measure_IR(...) so you can take a look at the data you get from the microphone and from the loopback channel. Check for things like clipping or cutoff at the beginning or end of the signals, or other undesired effects.
Click to expand...

Here it is:
Sweep.png

There is some kind of mean drift and a (little) spike at the end.

I also implemented a 'phase controlled exponential chirp' from https://www.uni-weimar.de/kunst-und...ation_of_ESS_impulse_response_measurement.pdf

phase-controlled-exponential-chirp.png
The peak at the end is gone, but the mean drift is still present and also the preringing (not shown).
 
  • Like
Reactions: 1 user
T

Tom Kamphuys

Member
Joined 2020
So, putting the samplerate at 192kHz iso 44.1kHz solved the problem. I don't know why, but will simply use 192kHz.

impulse.png

I measured my Elac Debut 2.0 from 0 to 180 degrees in 1 degree steps (still with my crappy test setup). I seriously doubt whether the 2D modal fitting and reevaluation is worth it once you have the measurement process automated. Get yourselves a cup of coffee and its done.

I measured with Octave, saved the impulse response to a wav file, read that into REW (see image above), applied a window too all* measurements and exported FRD data for use in VituixCad:

VituixCAD Directivity (hor).png


* I only could load ~120 measurements into REW before I ran out of memory.
 
  • Like
Reactions: 1 users
Kravchenko_Audio

Kravchenko_Audio

Member
Joined 2002
kravchenko-audio.com
Paid Member
Tom Kamphuys said:
* I only could load ~120 measurements into REW before I ran out of memory.
Click to expand...
It's easy to forget.

Thanks for the paper Tom. Number one problem of the fast signals is that there is little energy below 300 hertz. Rise time of the drivers is fast enough, but there is a delay in the time of the system. And there is an inherent lesser efficiency in the lower frequencies to test in the first place. Add in a port and you have another layer of trouble in getting a good signal to your microphone.

Mark
 
  • Like
Reactions: 1 user
mbrennwa

mbrennwa

Member
Joined 2005
audioroot.net
Paid Member
Tom Kamphuys said:
Here it is:
View attachment 1286667

There is some kind of mean drift and a (little) spike at the end.
Click to expand...
Both the drift and the spike at the end looks wrong. Do you have a filter somewhere? Also, the loudspeaker signal looks like it has not fully died off to zero when the recording stopped.

Tom Kamphuys said:
So, putting the samplerate at 192kHz iso 44.1kHz solved the problem. I don't know why, but will simply use 192kHz.
Click to expand...
Ok, that smells like some anti-aliasing somehow.

Tom Kamphuys said:
I measured my Elac Debut 2.0 from 0 to 180 degrees in 1 degree steps (still with my crappy test setup). I seriously doubt whether the 2D modal fitting and reevaluation is worth it once you have the measurement process automated. Get yourselves a cup of coffee and its done.

I measured with Octave, saved the impulse response to a wav file, read that into REW (see image above), applied a window too all* measurements and exported FRD data for use in VituixCad:

VituixCAD Directivity (hor).png


* I only could load ~120 measurements into REW before I ran out of memory.
Click to expand...
That's all pretty nice!

But I wonder:
  • Why do you export the data to REW to apply a time window and then export to VituixCad for plotting? You can do all of this right in Octave, fully automated. I attached an example I did in Octave a while back with the OSMC. Once you're done with the coffee, the plot shows up on your screen with not extra work.
  • Did you export the full impulse response data? 180 x 1.3 seconds at 192 kHz is quite a bit of data! Cropping the data before to the relevant 3 ms before exporting would reduce the size of the dataset by about 500x.
Keep up the good work!
 

Attachments

  • example.png
    example.png
    105.5 KB · Views: 18
  • Like
Reactions: 1 user
T

Tom Kamphuys

Member
Joined 2020
I think its time to shift the focus. I see 2 possibilities for the next steps.

1) Create a gui
2) Moving mic as a step towards 3D

Anyone any ideas about the GUI framework? What the GUI should look like? ...?

Anyone any ideas about the 'moving mic' hardware?


Both are not really my forte, so feel free to take the lead ;-)
 
  • Like
Reactions: 1 users
mbrennwa

mbrennwa

Member
Joined 2005
audioroot.net
Paid Member
As I understand it, we want the microphone moving in 3D to do what the Klippel system does. So we need to move the microphone in 3D.

I personally don't see why a GUI would be needed to scan the soundfield by moving the microphone around the speaker. I even think a GUI might interfere with the program. If at some point you absolutely want a GUI, I'd suggest you look at the cross-plaform GUI toolkits so it would work on most computer platforms. However, a warning is in order: it's quite difficult to make good GUIs!
 
G

Gulo_Gulo

Member
Joined 2023
As I understand the thread so far. 1 automated rotational axis, 1 automated linear axis for height and 1 manually adjusted axis for distance from rotational axis?

What would the needed accuracies be angular and height, and what time could be allotted for the system to stabilize after motion?

Would people be comfortable with permanent mounting holes or fixtures in their roof?
 
S

ScottG

Member
Joined 2003
Tom Kamphuys said:
I guess if you ask 10 people you get 10 answer (or more). I would like to have the radial distance motorized. The top and bottom of the cylinder also needs to be measured. I think I remember an accuracy of 5mm. For me, permanent mounting holes in my roof would not be my first choice.
Click to expand...
Reference again:



I could see 2 parallel beams above (height adjustable to the height of the room) with 2 vertical support strut’s on each side (with some cross-supports) of the testing area holding-up those beams. The beams would support a drop-down center spindle for the rotational axis and the vertical and inward/outward axis’s would hang down from the rotational axis’s base. It shouldn’t be to much bulk other than achieving stiffness considering that the motors aren’t that heavy.

ex. sort of with the dude hanging being the moving-parts (rotational axis base and the other 2 axis’s):
https://www.baseblocks.fit/pages/b-...MI8daBl7z_hAMVMnF_AB2vCwhxEAQYAyABEgLBCfD_BwE

..yes, the parallel beams would have to be slightly longer than twice the desired length of the inward/outward axis-arm; so if you want 4 feet as with the Klippel setup above you would need just over 8 feet in length.
 
Last edited:
  • Like
Reactions: 1 user
Kravchenko_Audio

Kravchenko_Audio

Member
Joined 2002
kravchenko-audio.com
Paid Member
Here's the best picture that I have been able to find.

https://www.klippel.de/fileadmin/user_upload/NFS_1.0.7_StudioMonitor.png

1710853557308.png


https://www.vevor.ca/drawer-slide-c...l-bearing-full-extension-500lb-p_010873950531

1710853600585.jpeg

60 inch drawer slides. Reasonably priced. You can us these for vertical and horizontal movements. 2 pair would be enough for horizontal and vertical movement of the microphones. And provide a fair degree of built in rigidity.

Slides Perspective.png

A little perspective on the size of these. They are more than adequate for the task at hand.

I'll try and put together a list of parts and places to order as I cobble together something for myself.

Kitchen is finally installed:
finger-brake-3d-3500-3500.jpg




Mark
 
Last edited:
You must log in or register to reply here.
Top Bottom