Is this impulse response correct??

[Ozric]

Dear Colleagues,

Please, see this impulse response - in most cases using such zoom it looks like solid and in bellow case I noticed strange gaps - may that be caused by a time constant from wrong configuration? Or hardware malfunction? I will be grateful for any support!

 

[John Mulcahy]

It just shows some ringing at low levels, at around 60 Hz or so at a guess. In a log plot like that lines which go to the bottom of the plot are where the signal passes through zero (log of zero is minus infinity), so that area shows some regions where the signal spends a little time above or below zero. It doesn't indicate any kind of fault or error, there is nothing abnormal about it.

 

[Ozric]

Thank you very much for such quick replay!! I didn't expect...

That shows how poor my knowledge about impulse response is...

How do you calculate that it is around 60 Hz if I may ask?

Anyway - waterfall or spectrogram does not show ringing at 60 Hz but there is quite deep hole at 58 Hz - it is like it would be notched and I consider it as result of interference due to loudspeakers positioning.

 

[John Mulcahy]

A deep notch can be from mic or speaker being placed at a node for a modal resonance at that frequency, the mode can sometimes be seen ringing at a low level later in the waterfall in those cases. The 60 Hz guess is from the time between the peaks.

 

[Ozric]

1000 ms / 17 ms = 58 Hz => it is so easy that I'm in deep shame that I missed it! Thank you to illuminate me Thank you very much!

Would you say similar about this graph:

It has been measured in the same room, but before implementation of some Helmholtz resonators tuned for 40 Hz (change between the graphs proves that I succeed in HR implementation because there is much less ringing - do I think correctly...?

PS: I'm asking, because one man having small company who makes acoustic adaptations told me, that it is probably some time constant in my measurement system and measurement is probably wrong... I will be happy if that's not true

 

[John Mulcahy]

There's not necessarily a lot to be learnt from looking at an impulse response alone, particularly in regions that are approaching the noise floor - that's why there are waterfall and spectrogram plots. If you want comments about measurements it would be a lot more useful to post the mdat files rather than images, particularly ones which omit the graph axes - why not use the graph capture button? I have no idea what the comment about time constant is meant to mean. The main time constants in measurement systems relate to the high pass frequencies of the various components in the chain, placing a limit on the lowest frequencies, but those limits are usually well below 10 Hz.

 

[Ozric]

Thank you very much for all the answers!

Generally I used to use mentioned waterfall and spectrogram as well as rt60 and other graphs. I use impulse response mainly to see reflections, but - as I wrote - one man suggested some hardware problems but using the same equipment in one place I have it, in another not - that is why I asked about this particular thing.

Now I think I know more - having questions about measurements itself I will first search forum before I'll ask.

Thank you one more time!