ARTA

Calibration of Soundcard Input Channels

Hi,

I'm new to ARTA, trying to see if it will meet my simple requirements for measuring amplifier frequency response and performing spectrum analysis. It could well be overkill.

I have my Juli@ soundcard setup in loopback mode; left and right outputs connected to left and right inputs. The card is setup as "ASIO 2.0 - ESI Juli@" with input channels 1/2 in Audio Devices Setup.

A test of the soundcard as per para 4.2 of the ARTA Handbook produces the result attached. Looks good and can be improved significantly by reducing the level.

I've completed para 1.5.1 Calibration of Soundcard Output Left Channel.

In para 1.5.2 Calibration of Soundcard Input Channels, the manual says: "Press the button 'Generate sine (400Hz)' and monitor the input level at bottom peak-meters".

The meters show no activity at all.

This is odd because the Juli@ control panel shows inputs and outputs at full scale.

What am I missing here?

The handbook talks about various measurement setups. I think I'd be looking to use the "Semi dual channel" setup which as I understand it eliminates soundcard artifacts from the measurements.

Is it required to tell ARTA which setup is being used? If it is, I haven't seen any mention of it.

Thanks.
loopback spectrum.png Juli control panel.png
 
Last edited:
I have the same sound card, are you attach an a.c voltmeter to sound card line out.
Are you looking at the right chanel?
Thanks for replying.

According to my True RMS DMM, with the output level set to -3dB, I'm getting 5.357V on both the left and right outputs.

I have the left and right outputs connected to the left and right inputs respectively.

I went through the instructions again. Nothing on the metres when I click Generate sinus (400Hz) but there is when I click Estimate Peak Input mv.

I've attached a screenshot.

Soundcard and Microphone Calibration.png

Another thing I've noticed is that in the Audio Devices Setup dialog, if I click the Control Panel button alongside "ASIO 2.0 - ESI Juli@", nothing happens.

I'll carry on through the manual.
 
Check the windows mixer settings.
Maybe something is in mute position.
Do the Windows Playback and Recording devices need to be set the same as the Soundcard Input and Output channels in ARTA?

I've been leaving the Windows Playback and Recording devices to Speakers and Microphone respectively.

Where the manual says "Set the left and the right line input volume to maximum" I assumed that was referring to the soundcard.
 
Standard error

Can anyone help me with the following issue I'm having with ARTA:

I'm trying to take gated measurements of a Woofer. I performed 4 measurements with the mic on axis with the woofer.

For all 4 measurements, no settings were changed: The mic location was kept the same, the speaker location was kept the same, same gating, same software settings....same everything.

However, my low-frequency response changes with every measurement. Anyone know why this would be the case?

Hello All,
Santa brought me some Test Tech for Christmas. I am looking to see if there is a standard input sweep voltage for measuring T/S parameters?

I do know about the issues you are talking about regarding testing woofers at low frequencies. You are windowing your sample time to exclude reflections. Using this time window the number of samples decreases as the frequency goes lower. With this decreasing sample size the statistical results will vary each time through. This is called Standard error. It is a sample size thing.

Thanks DT
 
ARTA spectrum analysis for dummies

Hi. I need an "ARTA spectrum analysis for dummies".

I'm familiar with sampling frequency and Nyquist but my University engineering days were many years ago.

My question is, what are the best settings of sampling rate, FFT size and Window for amplifier testing?

Fs of 44100 should suffice if I'm only interested in frequencies up to 20kHz.

Below are 2 spectra of a piece of equipment. One with Fs 44100, FFT 16384, Window Kaiser 5 and Linear averaging (which seems to give the best result) and the other with Fs 192kHz, FFT 131072, Window Kaiser 5 and Linear averaging.

Why are they different? In particular why do the higher Fs and FFT result in artifacts at odd harmonics that aren't present in the other one?
Loopback spectrum -30dB 230mV earthed mu metal shield.png Loopback spectrum -30dB 230mV FS 19200 FFT 131072.png
 
What sound card do you have ? And are you using an ASIO or directsound driver ? Are you sure the sound card natively supports 192Khz sample rate ?

If it doesn't, it will be sampling at a lower sample rate and then software upsampling it to 192Khz. This will cause a reduction in quality and can result in aliasing artefacts.

You need to confirm what sample rates are supported by the sound card in hardware, and don't go above the highest sample rate supported in hardware by the card as it's important to avoid any intermediate software resampling in the sound driver or windows.

Almost certainly the issue is the sound card or it's driver. I use 192Khz 24 bit all the time with ARTA and don't see artefacts like this.
 
What sound card do you have ? And are you using an ASIO or directsound driver ? Are you sure the sound card natively supports 192Khz sample rate ?
Thanks Simon. I'm using a Juli@ card that I no longer use in my music server. It does indeed natively support 192kHz sampling rate. From what I've read, the Juli@ seems to be widely used in this role.

I'm testing a sound card interface at the moment.

It occurred to me that I should redo the loopback test at 192kHz/131072 and as you can see below, the harmonics look pretty much the same, albeit with a slightly lower noise floor, so maybe they're contributed by the sound card.

From what you're saying, I should use the highest sampling rate my sound card supports.

Varying FFT size appears to produce different results too. Should I be using the maximum FFT size also?

I thought ARTA used the right channel as reference to cancel out sound card contributions?

Regards, Dave.
Loopback -30dB Fs 192kHz FFT 131072.png
 
I would imagine that dual channel mode can only cancel out linear distortions (frequency/phase response errors) not non-linear distortions, and it can only cancel linear distortions to a certain degree - see my earlier posts about problems with the low frequency response being attenuated too much in the reference channel.

Do those harmonics go away if you drop the input level slightly ? Some sound cards can not input a signal all the way up to 0dB FS without some distortion, sometimes you need to drop the level (in the analogue domain before the sound card) a couple of dB. I typically take my measurements at no more than -3dB FS.

I see that your test signal is around -20dB but that could be a result of the adjustment of the input level mixer in windows, which means the pre-amps may be still getting overloaded or being operated too close to 0dB FS. Does the distortion change if you adjust the line level input mixer in windows ?

Another possible cause is the pre-amp in the sound card, especially if it's a balanced input and you are trying to feed in an unbalanced signal.

Some sound cards support this using a servo balance input, but this inevitably results in significantly higher harmonic distortion.

This is the case with my Behringer UMC204HD - if I input a balanced/floating signal via the normal front panel XLR connectors everything is fine, however if I input an unbalanced ground referenced signal (with pin 2 earthed) then although this is a supported configuration, harmonic distortion jumps up quite a lot to the point where it's marginally acceptable for measurement purposes.

However I can work around it by bypassing the pre-amps completely and injecting the unbalanced line level signal directly into the "insert" input/outputs on the rear, which by their nature are unbalanced. And as the pre-amps are bypassed any distortion and noise introduced by them is avoided as well. This gives me the lowest possible noise and distortion levels when measuring unbalanced line level signals. (Although I lose the variable attenuator on the front and pre-amp gain) Maybe your sound card supports something similar ?

Regarding sample rate, ensure that windows is not trying to resample the audio. The best way to avoid this is to set the default sample rate of windows in the sound control panel (speaker properties, advanced) to the same sample rate and bit depth you have configured in ARTA, and make sure "Allow applications to take exclusive control of this device" is ticked.

If allow applications to take exclusive control of the device is not ticked AND the sample rate is not set to the same sample rate as ARTA, windows may software resample even if the hardware supports the desired sample rate. This can cause distortion and aliasing artefacts. It's absolutely critical that windows does not get in the way and start performing any resampling. A little bit of experimentation with ARTA sound device and windows sound control panel and/or vendor supplied sound card control panel software settings may be needed to ensure this.

If the sound card supports ASIO then use that as I think that avoids most of the potential problems of Directsound.
 
Last edited:
There are two types of licenses:

The personal license enables the single user personal use of the ARTA Software. The price for the personal license is 79 euros.
The commercial licence enables the institutional and multi-user use of the ARTA Software. The price for the commercial license is 149 euros.

Can anyone explain this n00b what exactly are the benefits of the commercial license compared to the personal? The term "institutional" sounds somewhat familiar (Shawshank Redemption!?) and I grasp the basic idea of "Multi-User" (like multiple users on one PC? Or multiple PC's running ARTA under the same License? Or cloud based platform to share measurements and ride the Unicorns?) but besides that I'm pretty clueless honestly. Thanks in advance!
 
Steps frequency response sweeps stop prematurely

Hi,

I'm working through the STEPS user manual. Setup and calibration is done.

I have the LR outputs of my Juli@ soundcard linked to the LR inputs.

Dual channel - Frequency response is selected. 20Hz to 20kHz in 1/12 octave increments. Sampling frequency is 44100 but I've also tried 96000.

After I click Start, some seemingly random number of steps is plotted and then the plotting stops. It actually got to 6092.6Hz once but it's usually much less than that.

Eventually I have to click Stop because the plot never completes.

I also notice that the generator level doesn't work the same was as it does in ARTA. The line that is plotted is always 0dB.

I confess that I do find the ARTA manuals challenging. I just want to do some simple amplifier measurements.

Any suggestions as to what I'm doing wrong please?
 
It would probably help if you posted a screen capture.
Dave.
Hi Dave,

Thanks for responding.

Here area a couple of screenshots. They don't add much.

As you can see, on this occasion the trace stopped at 761.9Hz. Generator level is set to -3dBFS but the response is at 0.03dB. 2nd and 3rd harmonic distortion aren't displayed.

I decided to check the calibration and realised that the input sensitivity wasn't calibrated. I fixed that and it put the output level at 20.19dB with the line off the scale at the top. The trace still wouldn't complete.
 

Attachments

  • STEPS calibration.png
    STEPS calibration.png
    36.9 KB · Views: 377
  • Soundcard loopback STEPS.png
    Soundcard loopback STEPS.png
    164.5 KB · Views: 375
Why do you have such a huge difference in your left/right calibration value? 20db?
Dave.
I suspect it's because I'd calibrated the left channel but not the right. I've fixed that now but the stopping and level problems still remain.

Soundcard loopback STEPS 2.png

Re calibration: where both the ARTA and STEPS manuals state "Press the button 'Generate sine (400Hz)' and monitor the input level at bottom peak-meters." I don't get any indication of levels on the meters. I only get meter readings when I press 'Estimate Peak Input mV'.

Anyway let's ignore the levels for now and focus on the premature stopping.

More information

  • When the traces stop the output from the soundcard ceases

  • I was able to get a nearly complete trace with my soundcard interface in loopback mode:
Soundcard interface loopback -3dB.png

  • I was able to get a nearly complete test of a tube amp:
amp 1 -15dB.png

  • One amp test produced a weird glitch
amp 1 -15dB weird glitch.png

  • Increasing Fs to 96kHz results in much shorter traces before they stop.
So, making progress.

Thanks again, Dave.