Manipulating data to find sound power for other frequencies?

[MechEng92]

I have a set of sound power for frequencies 125Hz, 250Hz, 500Hz, 1000Hz, 2000Hz, 4000Hz, 8000Hz, and 16000Hz however I need data for 63Hz, 125Hz, 250Hz, 500Hz, 1000Hz, 2000Hz, 4000Hz, and 8000Hz. Is there any way I can use the data that I have to create values for 63Hz?

 

[onatirec]

You are talking about measured data? I would think not.

Roughly speaking, you might have a loud, sharp single-tone at 63Hz or you might have a situation where the system is high-passed at 100Hz and there's virtually no lower frequency noise.

 

[MechEng92]

Thanks onatirec, the spreadsheet I have predicts overall sound levels by adding the sound levels from all sources in the system from 63Hz through to 8000Hz, is it possible for me to still use the data I have for this one component? Would I disregard the sound at 16000Hz and input 0dB(A) for 63Hz to get it to work?

 

[onatirec]

Oh, I see. you are interested in finding overall sound power, not necessarily extrapolating existing data to 63Hz.

You could of course sum what data you have. You can just put the 16kHz dB value in the 63Hz octave spot. They should be interchangeable; I'd try swapping a couple to make sure this is true.

Judging by the (A), it looks like your measurements are already A-weighted. They should still sum in the same way (as long as they are either all A-weighted or none are), assuming this spreadsheet is not also applying weighting to the individual bins before summing. If it is A-weighting them behind the scenes, then the above statement about swapping octave bin values is not going to hold.

 

[GregLocock]

Not too sure about swapping bins around. I suggest you look at your spectrum shape.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[MechEng92]

Thanks both. Greg, sorry to be naïve but what do you mean by spectrum shape and what can I deduce from the shape?

 

[GregLocock]

The shape of the curve on the screen. Personally I wouldn't try to guess what the 63 Hz is from the data you have, but if you are looking at high frequency or broadband problems I suppose you could give it a go.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Strong]

The simple answer is no. The best advice is to take the simple answer.

Walt