impulse Response too short

Victor Reijs

Posted on 16/10/2015, 19:22

I get sometimes this errors message 'The impulse response is too short (<80ms) for spatial parameter...'. In the attached picture one can see that the selected part in Audacity is 465 msec and still I get this error message. Does someone know the reason?

Tanks.

All the best,

Victor

angelofarina

Posted on 17/10/2015, 08:19

Junior Member

Group:: Member

Posts:: 94

Status:

There was no picture visibile.. You must first upload it to Postimage for inserting it on a forum!
Apart that, the control about IR length is only triggered when you attempt to process multi-microphone impulse responses for which the calculation of spatial parameters makes sense, for example in the case of binaural measurements. For measurements performed with normal omni mics, this control is not done.
So, what kind of multi-channel measurement are you doing?

In any case, nothing forbids you to select a longer portion of signal, the "silence" (actually, background noise) before and after the impulse does not affect significantly the calculation of acoustical parameters.

Victor Reijs

Posted on 17/10/2015, 10:54

Both the two omnidirectional mic. or binaural dummy head choice I get this error (or is it a warning?). I have two microphones in a binaural setting by the way. I need to made a selection of at least 770 msec to get rid of the error. The IACC is still being calculated.

All the best,

Victor

QUOTE (angelofarina @ 17/10/2015, 09:19)

There was no picture visibile.. You must first upload it to Postimage for inserting it on a forum!

Here is the picture, using binaural dummy head

Edited by Victor Reijs - 21/10/2015, 20:50

angelofarina

Posted on 30/10/2015, 10:35

Junior Member

Group:: Member

Posts:: 94

Status:

Your measurements looks horrible, a very poor signal to noise ratio
Hence, probably, the module cannot detect the direct sound peak, as the peak to noise ratio is smaller than the detection threshold specified.
Of course, you can try to adjust the direct sound detection threshold, but the correct thing to do is to investigate why your measurements are so bad, and fix the problem in your measurement method...

JmCt

Posted on 3/11/2022, 14:35

Dear Angelo,

First of all, thank you for such versatile yet powerful tools you have made and shared openly.

I used (Audacity) AURORA to measure some IRs, and experimented with different lengths of the ESS signal: I used 15 seconds and 1 second. When convolving the signals, I get the curves for the IR’s RMS and I see that for the shorter ESS, the more this curve looks like a comb filter:

15-seconds ESS:
ESS-15s-long

1-second ESS:
ESS-1s-long

It may seem “obvious” that this is due to the length of the measurement signal but… It is not obvious for me WHY. To be more precise:
1) What does this plot tell about the IR? (Is it the energetic contribution of each instant of time? Or is it the total energy of the IR over time?)
2) Why shorter lengths may cause this behaviour? And why exactly the comb-filter-like shape?
3) And linked to question 1): how does this shape of the RMS affect the IR?

I hope my question is clear and thank you in advance for your help.

4 replies since 16/10/2015, 19:22 213 views