James Yu Mus422 Homework 4 1) My signal and spectrum:
I estimate the peak to be about 3 bins wide, which is approximately 140.6250 Hz wide.
To estimate the SPL of other peaks, we use the formula given in the text:
Which is derived using Parseval’s theorem. <w^2> in the case of a sine wave is ½. (b) I estimated the peaks and SPL: Peaks 8812.5 4406 890.6 656.25 562.5 421.8
SPL 115.222 121.68 125.66 134 133.4 134.6
The peaks are not exactly at the frequencies we made the sinusoids to be. This is because we are using a DFT, and we will not likely sample the DTFT at the exact spots where the delta functions are.
(c)
(d) I used the triangle spreading function defined as:
10 log10 (F(dz, L_m)) = (-27 + 0.37 * MAX(L_m – 40, 0) * u(dz)) * |dz| where dz = z(f_maskee – f_masker) and L_m is the masker’s SPL and u(z) is the unit step function The masking curve is then defined as: SPL_B = SPL_A – 15dB + 10 log10 (F(dz, L_m))
(e) None of our signals will be masked. Although, in the rest of the frequency spectrum, we see that the artifacts of our limited time DFT will be masked (the lobes of the imperfect delta functions of our sinusoids). The SMR for all the peaks are approximately as follows:
Frequency 440 550 660 880 4400 8800
SMR 20dB 18.54dB 19.8dB 21.8dB 31.4dB 35dB
Bits Needed 4 4 4 4 6 6
(f) For the hanning window, I get the following mask model:
We see that the rejection of the Hanning window is less. Therefore, more of the spectrum is potentially unmasked. However, in this particular case, only artifacts from the DFT are unmasked (which we know aren’t true signals). But for a real world signal, more of the signal will be unmasked. The SMR for all the peaks are approximately as follows:
Frequency 440 550 660 880 4400 8800
SMR 17.18dB 20dB 21dB 22dB 30.5dB 34.1dB
Bits Needed 4 4 4 4 6 6
(g) Here I graphed my masking model, along with the minimum masking in each critical band.
Band
Minimum Mask
Peaks
Peak SPL
SMR
Bits Needed
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
13.8705 27.4152 51.0559 59.9743 59.5854 55.7996 48.4321 36.105 25.0329 10.5878 1.0374 0.0922 -1.6546 -3.4206 -4.9077 -4.9758 20.2607 21.2243 6.0862 2.7544 9.3013 13.3684 17.9116 58.3591
17.5
3.6295
36.185
8.7698
43.7925
-7.2634
58.8877
-1.0866
440 Hz
83.4
23.8146
4
550 Hz
82.2
26.4004
5
660 Hz
82.8
34.3679
6
880 Hz
74.5
38.395
7
49.4105
24.3776
35.2918
24.704
29.9561
28.9187
23.6739
23.5817
19.0013
20.6559
15.2899
18.7105
12.2
17.1077
10.7942
15.77
7.9484
-12.3123
14.5982
-6.6261
70.5
64.4138
38
35.2456
18
8.6987
64
50.6316
4400 Hz
8800 Hz
18
0.0884
-13
-71.3591
11
9