ARTA

I'm sorry for noob question but I still not able to understand clearly how to set proeprly Gating window in ARTA in measured impulse reponse.
I try to measure 10" speaker with install height 950mm and on same height mic with a distance between them 1m. Vituixcad says that splitting frequency to connect with nearfiled measurement is about 3xx Hz and windows time until first reflection is ~3.3ms


But on every video\articles I see only one impulse while I see two of them. How so set start\stop gating window properly?
I use dual channel mode measurements. Pic of measurements is below

zNXmzSs.jpg

Everything is right. The second impulse is important to you.
 
It is reference position for dual channel measurements. It
is the position of zero delay. In ARTA, the reference position is sample index 300.

I may be wrong, but if the the first impulse (red) is the reference from the loopback connection, it looks a bit too noisy for my liking. The loopback reference should normally be a very clean impulse without the hash seen here. What do you guys think?
 
I may be wrong, but if the the first impulse (red) is the reference from the loopback connection, it looks a bit too noisy for my liking. The loopback reference should normally be a very clean impulse without the hash seen here. What do you guys think?

You are not wrong.
When there is small replica of distorted antialiasng filter response, it is in 99% cases result of high cross-talk between left and right input channels. Some converters are really bad in this sense, i.e. Behringer UC202, or motherboard systems.

In ARTA user manual (chapter 4.2) I have described how to test soundcard dynamic range and cross-talk.

If your soundcard has large cross-talk you should use single channel measurements, but the problem remains when you measure impedance.

Ivo

P.S. You may get some IR replicas when measuring with clipped swept sine.
 
Thanks for comments and it sounds dissapointing for me about quality of sound card.

Well I use E-MU 0202Usb and windows 7.

For measurements I use E-mu phones one channel output -> Emotiva A-300 power amplifier with two resistors voltage divider on soundcard right channel input and speaker measured -> Behringer ecm8000 mic to get

Well like here http://www.artalabs.hr/AppNotes/AP1_MeasuringBox-ver-2-Rev1Eng.pdf on page3 but simplified without switches.

I will try to measure card loopback like in chapter "4.2. Testing the soundcard" wthout amplifier and with amplifier then.

p.s. I used a-300 because i dont have another one and i tried car amplifier previously Revolt Audio SS 85.4 with 60Ah battery used for LIMP impedance measurements and RLC measurements of Resistors and capacitors and found that with A-300 measurments more close to specification of components so it is more accurate I suppose
 
Hello,

I did impulse response measurement using External Excitation. Any idea why Burst decay and some other tools won't work? Says "To use this analysis, record or import impulse response file!"

Thanks!
 

Attachments

  • burst_decay.png
    burst_decay.png
    122.4 KB · Views: 411
  • example.pir.txt
    65.7 KB · Views: 48
How do you do this?
I've tried their beta driver with Windows 7, 8, 10 over the years, NO success. Last time with Windows 10, it frozen the PC solid.
Just curious, the E-MU went to a closet, and I've been living happily with a Focusrite Scarlett.

I can't know if procedure will work for your brand but can say below method works for other brands that only support older driver packages for their hardware devices be it sound cards or whatever devices where driver came packed into one EXE file.

In general Win7/8/10 supports using a driver package down from WinVista and up so pick the latest one relative to your running system and run it in combability mode, example is say you have a WinVista driver package that should run on Win10 then don't run it out of the blue but rightclick EXE file pick properties and in compability tab set settings so that it run as WinVista mode and push Okay, now one can run the installation. Some packages install some permanent procces that run in background and to get those run in compability mode open taskmanager and look for procces that belong to hardware then rightclick pick properties and in compability tab set to WinVista mode and push Okay.
 
Last edited:
How do you do this?
I've tried their beta driver with Windows 7, 8, 10 over the years, NO success. Last time with Windows 10, it frozen the PC solid.
Just curious, the E-MU went to a closet, and I've been living happily with a Focusrite Scarlett.

Zung, some users reported problems about E-MU0202 USB and Win7 . I saw several discussions with at least 50% success of working cards. Most time people solved issues by flashing different versiosn of firmware even stock one.

I had lastest firmware and clean MSDN Windows 7 SP1 without updates as in my system due some issues with OS as updates could not be installed and I'm lazy to reinstall OS.

I used EMUU_PCAppDrv_US_1_40_00_BETA.exe as driver package. Other versions did not tried.

The only issue I have as sometimes sound may disappear in all applications while card in steady state(I am not alone with this). if it plays something - no issues. So nothing special here , only luck
 
I need to double check an assumption: if I'm measuring directivity of a horn, I would rotate around the mouth, right? That seems like the axis that wouldn't introduce any incorrect differences in distance but I'd like to check that before I waste time on invalid measurements.
 
I would rotate around the mouth, right?

Yes, that's the safest rotation center no matter are you measuring in single or dual channel connection and mode.

Reflections could disturb automatic detection of impulse peak in single channel measurement mode. Positioning error usually happens to off-axis angles >120 deg. That shouldn't be a problem if you measure 0-90 deg only. Dual channel mode locks timing so starting point of time window stays where you have set it up to 180 deg off-axis so no need to adjust time window and export response of all measurements separately.
 
Last edited:
Audio Interface with acceptable dynamic range and cross-talk

You are not wrong.
When there is small replica of distorted antialiasng filter response, it is in 99% cases result of high cross-talk between left and right input channels. Some converters are really bad in this sense, i.e. Behringer UC202, or motherboard systems.

In ARTA user manual (chapter 4.2) I have described how to test soundcard dynamic range and cross-talk.

If your soundcard has large cross-talk you should use single channel measurements, but the problem remains when you measure impedance.

Ivo

P.S. You may get some IR replicas when measuring with clipped swept sine.

My USB audio interface also has the replica pulse at -2 to -3 mS so I'm looking for a USB Audio Interface with acceptable dynamic range and cross-talk for use with ARTA.

The ARTA site gives an extensive list of interfaces but they range in price from very reasonable to expensive.

The specifications on some devices I've looked at rarely quote a figure for cross-talk and only state one value for dynamic range without specifying the frequency. The ARTA-Manual tests on my USB interface show some similar results to those in the ARTA-Manual for low cost sound card SC1 but quote in the specification a value for dynamic range of "107 db (A Weighted)" so what to believe?

Has anyone with a USB interface done the ARTA tests and found their interface has acceptable results like for those shown for the RME BabyFace Pro. The RME is some 5 times the price of some of the other units listed on the ARTA site like the inexpensive, popular, readily available devices such as Tascam US2x2, Roland Quad Capture, Scarlett 2i2, Steinberg UR 22 MkII to name some reasonably priced basic interfaces.
 
One can easily become obsessed with 100dB and more noise´n stuff. In reality you should define what you want to measure with ARTA, before discarding sound cards or build in sound chips. For many applications like loudspeaker measurement there are many links weaker than even a cheap netbook soundcard chip. If you don´t measure in a special location, any sourounding noise will be more present than the internal.
If you want to measure amps and DAC´s things get more complicated, but ask youself: "Can I do anything about the measured data". If you are an audio developer, there may be some more professionel stuff around, but you would know. If you just measure for fun, it is a hobby and you can spent as much as you like or not. There are surprisingly good soundcards around, at eBay sometime for something like 12$ or even less. Because of the multi chanel hype some stereo or low channel models have lost any value. If you use a desktop, consider Creative Soundblaster cards. They are as low noise as advertised and dead cheap. Maybe use external USB for mobile measurements and a desktop for high quality stuff. Or pay the price...
 
Thanks for highlighting these, ScottG.

I aimed for something that can trustworthy measure up to 30K to see if there are any artifacts appearing above 20K which could be a sign of caution. And high undistorted SPL tolerance for THD measurements of high output testing levels. So I'm still quietly thankful to someone who recommended it here years ago. But I know there should be less expensive alternatives lying around somewhere.

Regarding el-cheapo & time consuming DIY mic exercises - any experience with Linkwitz Panasonic capsule mod and preamp circuit? I've searched and read descriptions of some build attempts, and they were controversial.

..at less than half the price:

EMX-7150 Measurement microphone (bulk) - iSEMcon_US


-and you could probably get an even better calibration file for it (..maybe from Cross-Spectrum).



Speaking of Cross-Spectrum, they carry MicW ..if you really wanted to spend as *much as the Earthworks mic., you could ask Cross-Spectrum about getting a M215 from them with their additional higher quality calibration.

Cross·Spectrum - Calibrated MicW i436 Microphones for Sale

MicW


*actually it should still be less.
 
Regarding el-cheapo & time consuming DIY mic exercises - any experience with Linkwitz Panasonic capsule mod and preamp circuit? I've searched and read descriptions of some build attempts, and they were controversial.


I think all my really cheap mic.s have the Panasonic capsule.. or some version/copy of it.

I've not tried the Linkwitz mod.. :eek: Isn't it a voltage-to-current operation change? If so that usually results in a much lower noise result, but that's probably the only change unless there is an inductive rise from the capsule at just the right range (20-100 kHz): which I doubt there is. Who knows though, the change in amplifier might simply allow it to extend much higher in freq.. Even then though, you still have no idea of linearity. :eek:

Basically the least expensive Mic's I've seen that are of the 40 kHz -3db with calibration are from Earthworks.. and at that price I'd start looking hard at spending more for an ACO Pacific reference. My guess is that calibration (full bandwidth file) services that go beyond 20-25 kHz are expensive as well. :eek:


A calibrated MEMS capsule would be ideal for this (cheap + high freq.), don't know of any though.. Of course all MEMS are "calibrated" so you might be able to do an inversion response to linearize it assuming you had the original response (though as seen below, it's just a "typical" response):

https://www.zachpoff.com/site/wp-content/uploads/Knowles-Ultrasonic-Application-Note-AN17.pdf
 
Last edited:
Hi,
In attachment I send you simple impedance jig that enables impedance measurement without power amplifier. Just use headphone output (RME outputs 3/4) as generator with reference resistor 47 ohms.
Use RME inputs 3 and 4 as unbalanced one.

Best,
Ivo

Works very well. Ref resistor 47R, 38R value in Limp to get proper reading. Is calibration necessary?
 

Attachments

  • baby1.jpg
    baby1.jpg
    126.3 KB · Views: 353
  • baby2.png
    baby2.png
    51.9 KB · Views: 354