Question on measurements

[Leo11]

I would like to maximize the quality of my measurements by average over multiple acquisitions, in order to minimize noise interference etc.
Is there a way to do this with Audiolense?

In addition, can microphone phase calibration be used in Audiolense or is it left out?

 

[JM1]

Yes, mic phase is used if in the cal files.
What interface are you using? If using a USB mic, switching to an interface that provides precise time alignment between sweep and microphone capture might be something you would want to implement. If your delays are not rock solid with repeat measurements, that needs to be corrected. If you are using a USB mic, plug it directly into a port on the computer, hubs can cause timing problems.
Have you tried looking at your SNR? REW is an easy way to do that. I find that with a 20 second sweep at about 85 dB, noise is no issue in my environment.
It's good that you are looking into this, good measurements are mandatory for good filters.

 

[juicehifi]

No it is not possible to average several measurements like that. You will achieve the same if you increase the length of the sweep. But my experienc is that the upside of this is only theoretical. I never use more than 10 s sweeps, and some users prefer 3 s.

Microphones are minimum phase devices, so the phase info is not used, but calculated from the magnitude info.

 

[Leo11]

Thanks for your answers.

Considering the calibration uncertainty of the microphones (usually bigger than deviation), as well as the unknown of the calibration field (diffuse, free, etc.), does it make sense to estimate the phase as a minimum phase system?

Making longer sweeps is something that I have considered, however longer duration means greater chance of catching unwanted noises in my environment (city).
If I used the multiseat correction option by loading multiple measurements of the same point, would I get the same result?
Out of curiosity, does the multiseat correction also take into account the phase of the loaded measurements or only magnitude?

 

[juicehifi]

Yes it makes sense to calculate the minphase response of the microphones, since they are minphase transducers. And it takes out any nonminimum phase artifacts that might otherwise be introduced.

The multiseat accounts only for frequency responses for the other measurements.

 

[Leo11]

Thanks a lot for the answer.

In order to optimize the measurement I have a further question:
AL automatically suggests the band to measure for each driver based on the crossover set. Now, for alignment purposes it should be useful to measure over the entire available band in order to exploit the greater clarity of the IR at high frequencies, unless the alignment is done by estimating the minimum phase response of the system and considering the respective excess phase of the measurement.
I can't find any information about the AL alignment method, so I wonder if measuring the drivers over the entire useful band compared to the one actually used could be more or less useful.

Another thing pertinent to the IR: the two FDWs that we set in the procedure are referred to the IR peak of each driver or to the IR beginning?

Finally I take this opportunity to ask a last question about the size of the FDW for TTD: theoretically under Schroeder it should be useful to include the first reflections with respect to each axis of the room so that by canceling them all the tails are automatically reduced.
Obviously it is not possible to analytically determine the size of the FDW to include only the first reflections.
So I was wondering, where I set a larger FDW so that it also includes second/third reflections, if the filter has already done so to cancel the first reflection some of them will not recur, so if the filter tries to correct these later ones it is actually introducing distortion.
Is this deduction correct?