Reverberation Time Measurement

[NVHNovice]

Hello,
I am trying to analyse the reverberation time of a partially constructed hemi anechoic room. I have taken SPL measurements at various positions in the room using a speaker and one microphone and moving them around in the room at various positions(3 positions for the speaker and 4 for the mic along with 5 measurements per position for repeatibility check). I have gathered throughput data using Testlab system.My question now is how do i best calculate the reverberation time(RT60) using these measurements. I would like to fit a straight line to the OA level SPL curve which decays when the speaker is shut off and i would like to do this is Matlab. ANy help is most appreciated

 

[GregLocock]

Off the top of my head using RT to characterise a SAC seems a pretty odd idea. You might find some hints in a book by Beranek, or an architectural acoustics book.

So what data did you gather exactly? White noise through the speaker and then cut it off?

RT varies with frequency, so first you need to decide on what frequency band(s) to analyse over, then just plot the decay in dB in that band against time. Then best fit a straight line to that.

A long time ago I was involved in the design, construction and testing of a couple of SACs. Rather than using RT I measured the sound field around the centre of the SAC (our noise source was pretty much fixed in location), and the drop off with distance. I then compared that with the free field performance of the same sound source. Then we knew what we were dealing with.


Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[GregLocock]

Incidentally, your RT60 is likely to be around 0.3-0.5s, and you probably only have 20 dB of usable data, so think VERY carefully about frequency resolution and smearing and so on. You might be better off using a bandpass filter in the time domain rather than an FFT.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[NVHNovice]

Hi Greg,
Thanks for the help. Ya i know using RT to characterise a SAC is not very common but this is an exercise we are dealing for academic interest. Ya i have done exactly as u said taken the overall levels against time and tried to fit a straight line to that. This room is not completely constructed so i am getting a slightly higher time, but i guess its still pretty ok.
I have gathered the data as u rightly guessed. Started wth the speaker random noise and cut it off and measured microphone response at various positions.
Thanks again for the help.