Member
Joined 2003
Wow, just fired up that program and I was not aware that it is solely for spectrum analysis ! Will give this a try, let you know my experience with it.
When measuring distortion vs frequency in Steps, you can pick thd and individual harmonics, 2nd - D2, 3rd - D3, etc..
That is certainly true, but the problem is in order to measure the peak amplitude, you have to position that cursor EXACTLY on top of the harmonics otherwise it gives a way too far figure.
Thanks bogdan, yes I already noticed that possibility but compared to the spectrum analysis given by Arta I am still prefering Arta, sorry.
Question about Full Dual Channel Measurement with Arta.
The external sound card I am using is a Motu M4.
Is there any reason I could split the speaker out from the amp going to the speaker and feed it back into the interface and use that for dual channel measurement without using an Arta Box?
The line level in can be attenuated with a knob on the interface, and as long as I monitored the level in and keep the amp below a few watts, would this work and be safe?
The external sound card I am using is a Motu M4.
Is there any reason I could split the speaker out from the amp going to the speaker and feed it back into the interface and use that for dual channel measurement without using an Arta Box?
The line level in can be attenuated with a knob on the interface, and as long as I monitored the level in and keep the amp below a few watts, would this work and be safe?
Member
Joined 2003
The ARTA jig app note includes zener diodes for "safety" but I omit them, not necessary unless you're prone to doing silly things. Is it less safe? Sure, but several years in all my gear is still in perfectly good working order.
You really don't need a fancy jig in a box, it's just a couple resistors and some scrap wire really. 950 Ohm and 5500 Ohm provides about 17dB of attenuation before hitting the audio device input, provided your input impedance is 10-20kOhm. That's not insignificant, so if you want to rely solely on the input gain control, you may find it very sensitive and easy to overload, and higher potential for damage of a nice audio interface, that could have been prevented with 50 cents of parts. The choice is yours, wonders of DIY
You really don't need a fancy jig in a box, it's just a couple resistors and some scrap wire really. 950 Ohm and 5500 Ohm provides about 17dB of attenuation before hitting the audio device input, provided your input impedance is 10-20kOhm. That's not insignificant, so if you want to rely solely on the input gain control, you may find it very sensitive and easy to overload, and higher potential for damage of a nice audio interface, that could have been prevented with 50 cents of parts. The choice is yours, wonders of DIY
I am just getting started with ARTA, and I have a few questions. I have tried to search this thread, but I was not successful in finding what I was looking for.
I want to properly make an impulse response measurement with swept-sine signal. The purpose of this is to measure far field gated frequency response and phase of loudspeaker drivers. I am using the "semi-dual channel setup" as shown in Fig 1.8 of the ARTA manual.
The sampling frequency of my audio interface 48k, so in the "Impulse Response Measurement" dialog box I set the sampling rate to 48000.
1) I understand that the FFT block size needs to be 2x the swept sine length. Does ARTA automatically make them compatible? Or do I need to set FFT block size to be 2X swept sine length?
2) What is considered a "good value" or a "best practice" value for FFT block size and swept sine length when making far field gated driver measurements?
3) My ARTA installation defaults to IR Window = "Uniform" and Signal Window = "Hanning". Are these settings appropriate?
Thanks !
Jim
I want to properly make an impulse response measurement with swept-sine signal. The purpose of this is to measure far field gated frequency response and phase of loudspeaker drivers. I am using the "semi-dual channel setup" as shown in Fig 1.8 of the ARTA manual.
The sampling frequency of my audio interface 48k, so in the "Impulse Response Measurement" dialog box I set the sampling rate to 48000.
1) I understand that the FFT block size needs to be 2x the swept sine length. Does ARTA automatically make them compatible? Or do I need to set FFT block size to be 2X swept sine length?
2) What is considered a "good value" or a "best practice" value for FFT block size and swept sine length when making far field gated driver measurements?
3) My ARTA installation defaults to IR Window = "Uniform" and Signal Window = "Hanning". Are these settings appropriate?
Thanks !
Jim
Duration of IR measurement with swept sine excitation
Hi Jim,
When you measure IR with swept sine you have to choose the length in samples of resulting IR. The choice of length is not critical as long as time duration of IR is larger than measurement room reverberation time.
During the measurement ARTA treats swept sine signal as nonperiodic signal, and record possible signal decay after the excitation stops. Practically in analysis ARTA uses FFT size that is double of proposed IR length.
Best,
Ivo
Hi Jim,
When you measure IR with swept sine you have to choose the length in samples of resulting IR. The choice of length is not critical as long as time duration of IR is larger than measurement room reverberation time.
During the measurement ARTA treats swept sine signal as nonperiodic signal, and record possible signal decay after the excitation stops. Practically in analysis ARTA uses FFT size that is double of proposed IR length.
Best,
Ivo
Always use an Hann(ing) window. A good start is 25%.
Uniform (boxed/rectangular) windows are totally useless when it comes down to measurements with reflection. Those type are only usable with very strong impulses without any reflections, or reflections extremely small compared to the impulse. Or they can be used with continuous signals.
The window size depends on the reflections you're measuring.
FFT block size is not very critical in ARTA
Thanks, but I'm not looking for minimum phase.
What I want is a correct phase response, like the one I got with my old Clio system.
In measurements with Arta, I always get this sawtooth visualization of the phase, as if there is always some group delay. I cannot understand where my mistake is.
What I want is a correct phase response, like the one I got with my old Clio system.
In measurements with Arta, I always get this sawtooth visualization of the phase, as if there is always some group delay. I cannot understand where my mistake is.