Your measured results and modelling is quite similar in some respects to the original filter of the KS-138 where I recently modeled it, but obviously only based upon data sheet FMS and Z . See here:
I used your optimizer on the original filter and it appeared to improve things (theoretically) - see here:
But what I found really interesting was the theoretical results achieved from just single poles - see here:
Measurement:
Thanks for your solid advice, much appreciated. For an enthusiast (rather than hard core golden ears speaker designer) the OmniMic v2 appears as a complete kit including calibrated mic, software + (not sure if their test CDs add much?, presumably test tones/white/pink etc needed by the software?), where as the Scarlett is solely a twin Mic to USB, so I would need to figure out all the other requirements wouldn't I?
Cheers again
Lea
I used your optimizer on the original filter and it appeared to improve things (theoretically) - see here:
But what I found really interesting was the theoretical results achieved from just single poles - see here:
Measurement:
Thanks for your solid advice, much appreciated. For an enthusiast (rather than hard core golden ears speaker designer) the OmniMic v2 appears as a complete kit including calibrated mic, software + (not sure if their test CDs add much?, presumably test tones/white/pink etc needed by the software?), where as the Scarlett is solely a twin Mic to USB, so I would need to figure out all the other requirements wouldn't I?
Cheers again
Lea
Sorry, I can't edit my posts (for corrections) and they all have to be approved by moderators, getting a bit frustrated by this now!
Anyway...
So a little research on my behalf (with your pointers):
1 x Scarlett 2i2 2nd Gen £130 (not sure if this powers the mics' 48V ?)
2 x Sonarworks XREF20 £140 total
1 x ARTA? (Free)
Pretty good price so far, what have I missed?
Lea
Anyway...
So a little research on my behalf (with your pointers):
1 x Scarlett 2i2 2nd Gen £130 (not sure if this powers the mics' 48V ?)
2 x Sonarworks XREF20 £140 total
1 x ARTA? (Free)
Pretty good price so far, what have I missed?
Lea
With semi-dual connection reference channel comes from output of sound card. That has some advantages compared to full dual connection.
Measurement preparations for VituixCAD has some information about settings. I'll probably update that document for new Convert IR to FR tool. It's a bit more visual and flexible than native polar response export of ARTA 1.9.
An externally hosted image should be here but it was not working when we last tested it.
Measurement preparations for VituixCAD has some information about settings. I'll probably update that document for new Convert IR to FR tool. It's a bit more visual and flexible than native polar response export of ARTA 1.9.
I also tested few electrically 1st order options. This was one of those:
One test person was able to listen it about 30 seconds
Not so bad, but diva low treble and smoother power response through XO range was mission impossible for him.Not sure how I missed your document, perfect, albeit highlighting how complex this field is! Maybe I'll create a video for the DIY nuub once I've got my little brain around it?
Can you explain a little more about the reference channel, is it used as a one-off to 'calibrate' the sound card, or concurrently whilst running tests?
I'm wondering if REW can be used rather than ARTA as I'm struggling to find introductory YouTube videos for ARTA, but have found a few for REW?
30 seconds, ouch, perhaps I should simply ditch the drivers favouring something more modern and better matched? It's the sentimental Lea inside me wants to fix them!
Lea
Update: these documents were really useful in trying to understand how things fit together:
http://www.artalabs.hr/download/ARTA-user-manual.pdf
http://www.artalabs.hr/download/LIMP-user-manual.pdf
Can you explain a little more about the reference channel, is it used as a one-off to 'calibrate' the sound card, or concurrently whilst running tests?
I'm wondering if REW can be used rather than ARTA as I'm struggling to find introductory YouTube videos for ARTA, but have found a few for REW?
30 seconds, ouch, perhaps I should simply ditch the drivers favouring something more modern and better matched? It's the sentimental Lea inside me wants to fix them!
Lea
Update: these documents were really useful in trying to understand how things fit together:
http://www.artalabs.hr/download/ARTA-user-manual.pdf
http://www.artalabs.hr/download/LIMP-user-manual.pdf
Last edited:
I suppose you've already found some explanations for reference channel from ARTA's manual, but very shortly (because this thread is not for measurement programs):
1) Reference signal compensates linear distortion of sound card i.e. works as real time calibration of sound card, assuming that channels are close to identical. Note that all measurement programs or measurement modes don't use reference channel for response calibration.
2) Reference signal enables measurement of signal's travel time from output to mic input just by comparing left (mic) and right (ref). Total travel time includes processing time of possible dsp device etc. and sound flying time from speaker to mic. Latter one enables measurement of timing/phase differences between different drivers assuming that distance from mic to common reference point on baffle surface is kept constant.
Some proprietary measurement gear such as CLIO do not have external loop for reference. Reference signal is arranged internally so that measurement application is able to calculate "absolute" timing.
As mentioned in some other thread, semi-dual or full dual system is mandatory in practice if target is to simulate multi-way to 3D space.
Last edited:
REW has dual channel mode so it's valid for the job with generic 2-channel gear linked before. It does not have polar response measurement sequence and automatic file naming (by time) does not support off-axis measurements; naming by plane and angle. Complex with (too) many options/switches and slow to use, imo.
Program seem to be designed for other than speaker designing. Second evidence (in addition to name "Room eq..") for that is public/advertised support for Umik-1 which is bad gear for this hobby due to missing timing reference in the second channel.
Those drivers could be perfect for you so don't trash them before you have tried.
You can use rew for this job and achieve the same result as arta, it's not noticeably more work ime. All you have to do is name the measurement as you go (like H0, H15, H30) then export all as txt. There is also an option to apply a window to all measurements while keeping the ref time (aka the peak position) which is generally what you want for this job. What do you mean by "does not support off axis measurements"?
Thanks you Kimmosto
As you can probably guess, I'm struggling to locate novice instruction for ARTA (there is so much to take in) so quality of documentation is paramount and hence I have been looking at REW (and of course it's free - amazing! which is useful to get on board).
I notice REW does support semi-dual reference so that's also perfect for getting started, but allows just that little edge, timing of the OS response (due to my old lappy) could be important and snippets of information regarding this has really helped me build the BOM for this projects thank you.
I'm concerned that I'm hogging the thread a little and have gone slightly off topic so please do simply nudge me and I'll head over the the main forum, but again it's really appreciated, and being able to discuss with the actual designer is truly mindblowing.
For now the BOM is as you've recommended but maybe REW (in speaker testing mode), lots of quality documentation and much in novice format ;-)
I've learnt so so much, thanks.
Lea
As you can probably guess, I'm struggling to locate novice instruction for ARTA (there is so much to take in) so quality of documentation is paramount and hence I have been looking at REW (and of course it's free - amazing! which is useful to get on board).
I notice REW does support semi-dual reference so that's also perfect for getting started, but allows just that little edge, timing of the OS response (due to my old lappy) could be important and snippets of information regarding this has really helped me build the BOM for this projects thank you.
I'm concerned that I'm hogging the thread a little and have gone slightly off topic so please do simply nudge me and I'll head over the the main forum, but again it's really appreciated, and being able to discuss with the actual designer is truly mindblowing.
For now the BOM is as you've recommended but maybe REW (in speaker testing mode), lots of quality documentation and much in novice format ;-)
I've learnt so so much, thanks.
Lea
I'm aware of everything you wrote but wasting of time is evident if off-axis set is 0-180 deg with angle steps of 5 degrees, and speaker is multi-way. My typical measurement sequence is only 19 off-axis measurements per driver, but sometimes 37 directions is valuable.
I use CLIO which has root filename + automatic numbering and support for motorized turning table. ARTA has also off-axis measurement sequence with pause timer which is very handy with manual turning table.
So, I will never use REW or recommend to others as long as automatic naming with root filename + numbering is missing - despite of $0 price tag.
One thing not yet mentioned is 'Conversion IR to FR' in VituixCAD. It supports CLIO mls and ARTA pir files but not REW mdat. That might be possible to add, but will not happen without public binary file spec.
Use at your own risk
On that note, there is one key difference IME between REW and ARTA and that is the way the window is specified relative to the peak vs the arta approach of picking some absolute sample positions to apply to all measurements. If the position of the peak never moves then this is a non issue but if it does, and I think some measurement setups will be prone to this, then it is something to bear in mind (i.e. may need manual adjustment).
The other difference is that REW is prone to producing some rather large txt files which can things a bit slow to load/manipulate.
On that note, there is one key difference IME between REW and ARTA and that is the way the window is specified relative to the peak vs the arta approach of picking some absolute sample positions to apply to all measurements. If the position of the peak never moves then this is a non issue but if it does, and I think some measurement setups will be prone to this, then it is something to bear in mind (i.e. may need manual adjustment).
The other difference is that REW is prone to producing some rather large txt files which can things a bit slow to load/manipulate.
^Reference time moving (peak searching) for each measurement is not a problem with REW because we can easily apply common time window settings to all measurements before exporting to txt/frd. BUT that does not remove phase response wrapping which forces us to export txt/frd files with "resolution of measurement". That creates huge files with e.g. 128000 data rows.
The other option is to adjust "t=0 offset" for each measurement to enable e.g. 48 or 96 PPO txt/frd export without damaging phase data. That action would take some time if you have 3x37 off-axis measurements in the project.
This is one quite good reason why no one should use REW for speaker designing imo.
The other option is to adjust "t=0 offset" for each measurement to enable e.g. 48 or 96 PPO txt/frd export without damaging phase data. That action would take some time if you have 3x37 off-axis measurements in the project.
This is one quite good reason why no one should use REW for speaker designing imo.
And then I found this similar combo...
Compact-Studio-Monitor
I might model this xover design for mine just to see. More importantly, all measuring parts and cables ordered today, longest lead time seems to be the Mic, hopefully next weekend the fun begins 👍 Even found a 0.1% 100R metal film and crock clips to build the impedance measurement probe circuit...
Compact-Studio-Monitor
I might model this xover design for mine just to see. More importantly, all measuring parts and cables ordered today, longest lead time seems to be the Mic, hopefully next weekend the fun begins 👍 Even found a 0.1% 100R metal film and crock clips to build the impedance measurement probe circuit...
^I simulated also CSM in 2010 with axial measurement data. Higher XO point does not remove power response issue but filters spiky low treble smoother. It is one option to try.
PS My favorite was 1st order version.
An externally hosted image should be here but it was not working when we last tested it.
PS My favorite was 1st order version.
Do you mean this 1st order..?
Out of interest, why, simplicity/cost? I have a concern over the modelled 2R load impedance at 10kHz...
The Power DI curve of your axle response version isn't shown so I can't compare which is a shame as TBH I'm not sure what I'm looking for in Power & DI curve...I'll go read!
Cheers
Lea
Out of interest, why, simplicity/cost? I have a concern over the modelled 2R load impedance at 10kHz...
The Power DI curve of your axle response version isn't shown so I can't compare which is a shame as TBH I'm not sure what I'm looking for in Power & DI curve...I'll go read!
Cheers
Lea
Last edited: