Experiment to measure HRIR, HRTD using FocusRite 2i4 instead of NI DAQ?

[merc5361]

This might be a weird question on this forum, but I'm a literally newbie embarking on a acoustic project for my masters thesis. Have no previous knowledge on acoustics apart from theoretical relating to psychoacoustics and Sound Pressure Levels. Totally new to all the hardware and setting up the experiment for data acquisition for subsequent signal processing.

So my lab has a FocusRite Scarlett 2i4 hanging around and my supervisor mentioned I can use this audio interface for data acquisition. However, I am stumped on how to proceed with this piece of hardware. I was hoping the project budget would permit me buying the National Instruments audio DAQ hardware which would basically be plug and play to MATLAB.

So just wondering if I should somehow try to convince my supervisor to buy the (expensive) NI hardware or if I can make do with the FocusRite Scarlett 2i4?

Basically, what I intend to do, is produce a 30s Signal through MATLAB, pass it through the DAQ which transmits the audio to the speaker and also records the impulse response at the microphones located at the dummy head. I would then utilize this data later on for subsequent analysis.

Thanks so much, and my apologies if I have not conveyed enough information. Please feel free to ask me for more info if something's not clear on the above. Thank you again and hope someone can help me out here!

 

[AnalogSteph]

How many measurements are you going to need? Is there any automation involved? The latter is what tends to make things $$$, generally adding one or two zeroes to the price tag.

If it's not too many and you have to do everything by hand anyway, the trusty 2i4 ought to be quite sufficient. (BTW, 1st or 2nd gen?) Matlab can import and export WAV files, then you just need software to playback / record. The specifics then depend on the host computer at hand and personal preferences. On a Windows machine, I might get everything together to compile Audacity with ASIO support and then use software playthrough. (WASAPI ought to work OK given that a number of issues with the WDM drivers have been ironed out over the years, but if in doubt ASIO ought to be more robust with these. I kind of doubt you even need 16-bit performance with a sensible input gain setting, but anyway.)

Also keep in mind that you'd have to interface the microphones to the fancy NI hardware first. I assume the mics in the dummy head are the phantom powered variety matching the 2i4, not IEPE.

 

[NTK]

Welcome to ASR!

If you are using MATLAB, I don't know if there is any significant advantage to using NI DAQ hardware. I think it is just as easy using the Focusrite Scarlett with MATLAB as it is with any NI DAQ. I'd only consider NI hardware if I am using LabVIEW.

I don't use MATLAB (or Octave) that much but I think this is what I'd do.

• Connect as shown (note: diagram shows only 1 ear). If you are recording both "ears", you'll need 3 total input channels, 1 more than what's available in the Scarlett 2i4. You may be able to get by without recording the amplifier output, but that's not ideal IMHO.
• Make sure the amplifier output voltage does not overload the input of the Scarlett 2i4, which is max 22 dBu = 9.75 Vrms, or 13 dBu = 3.46 Vrms, if you use the "instrument input" mode which band pass filters the signal to 22 - 22k Hz. Otherwise, you will need an attenuator to attenuate the signal.
• Generate a chirp (linear should be fine) that covers your frequency range of interest.
• Play the chirp, and record the outputs (see link). Extract the signal data arrays.
• The transfer functions from the amplifier output to the mic inputs can be calculated by taking the FFT of dummy head mic readings, divided (element-to-element division) by the FFT of the amplifier output readings.
• Since these transfer functions will include that of the loudspeaker, you will need a separate measurement of the loudspeaker only (i.e. without the dummy head) transfer function to compensate for it.