Acoustical Characterization

1
Acoustical Characterization
Yashas Honasoge Stephan Themis Wenjun Xu Xiao
Zhang
2
Reverberation Time Calculation
Reverberation Time is the time required for a
sound in a room to decay by 60dB (called RT60)

• T 55.2 V/Ac, where c 343 m/s
• T 0.161V/(Sa1Sa2San)
• Measurements of the room GFS 218

3
Reverberation Time Calculation

• Total Room Sound Absorption

• A S1a1 S2a2 .. Snan
• ? Siai 
• where
• A the absorption of the room
• Sn area of the actual surface (m2)
• an sound absorption coefficient

4
Reverberation Time Calculation

• Surface Area of Different Materials

5
(No Transcript)
6
Reverberation Time Calculation

• Sound Absorption Coefficient

• http//www.sengpielaudio.com/calculator-RT60Coeff.
  htm
• http//www.acousticalsurfaces.com/acoustic_IOI/101
  _13.htm
• http//online.physics.uiuc.edu/courses/phys199pom/
  Student_Reports/Fall02/Ala n_Truesdale/Alan_Truesd
  ale_Absorbtion_Coefficients.pdf

7
Reverberation Time Calculation

• Reverberation Time Calculation Results

8
Generated Impulse Response

• Generated impulse and recorded response at
• GFS 218 Middle of the room
• GFS 218 Corner of the room
• The measured response
• was sampled at a 44kHz
• with a sample size of 16 bits
• Recording was performed using the Goldwave
  software and a standard desk microphone

9
Measured Response
10
Noise Reduction

• Original concepts using Matlab
• Find the 60dB reduction point and corresponding
  index value
• Use the uncorrupted region of the impulse
  response to extrapolate a T60 value
• Design a filter in Matlab to remove the noise
• Final method using proprietary software Audacity
  to filter the noise

11
Original Recording (Listener at Corner)
12
FFT of Original Recording (Listener at Corner)
13
Bandpass Filter Design
14
Bandpass Filter Setting
15
Filtered Recording (Listener at Corner)
16
FFT of Filtered Recording (Listener at Corner)
17
Original Recording (Listener at Middle)
18
FFT of Original Recording (Listener at Middle)
19
Filtered Recording (Listener at Middle)
20
FFT of Filtered Recording (Listener at Middle)
21
Matlab Filter Design (FFT)
22
Matlab Filter Design (fdatool)
23
Filtered Response (FFT)
24
Filtered Response
25
Filtered Response (Audacity)
26
Calculating T60 Value

• Created a Matlab script to which we passed
• The location of the filtered response signal
• The threshold value (in our case 60 dB)
• The frequency at which the signal was sampled

function time60 reverb(fileloc, threshold,
frequency) loadwav wavread(fileloc)
thresh 10(-0.05threshold) MAX,MAX_IND
max(abs(loadwav)) thresh thresh MAX
bits,channels size(loadwav) found
0 for i bits-11
if((abs(loadwav(i) gt thresh)) (found0))
time60 (i - MAX_IND)/frequency
found 1 end end
subplot(2,1,1), plot(loadwav)
subplot(2,1,2), plot(20log10(abs(loadwav)/MAX))
27
Impulse Response Measurement

• Balloon Popping near the Blackboard

28
Predicted Signal (Environment Simulation)

• GFS 218 Listener Corner
• Recorded Signal vs. Calculated Signal from the
  impulse response

29
Predicted Signal (Environment Simulation)

• GFS 218 Listener Middle

Recorded Signal vs. Calculated Signal from the
impulse response
30
Predicted Signal (Environment Simulation)
original wavread('GFS218_Talk_Original.wav') or
iginal_fourier fft(original) impulse
wavread('GFS218_Pop_sourceBoard_listenerMiddle.wav
') impulse_fourier fft(impulse,278712) pred_fo
urier original_fourier.impulse_fourier pred
ifft(pred_fourier) subplot(2,2,1),
plot(original) subplot(2,2,2),
plot(abs(original_fourier)) subplot(2,2,3),
plot(pred) subplot(2,2,4), plot(abs(pred_fourier)
) soundsc(pred_time,44100) wavwrite(pred_time,441
00,'GFS218_Talk_Middle_Predicted.wav')
31
Longest Reverberation Time Measurement

• Longest Reverberation Time

Group 5s Secret Place for the Longest
Reverberation Time Recording NOW, HERE WE
GO
32
Longest Reverberation Time Measurement
VHE STAIRWAY!!
33
Longest Reverberation Time Measurement

• T60 Calculated using Matlab Code 2.0566sec

• Signal Waveform from Audacity