Добавил:
kiopkiopkiop18@yandex.ru t.me/Prokururor I Вовсе не секретарь, но почту проверяю Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Ординатура / Офтальмология / Английские материалы / Auditory and Visual Sensations_Ando, Cariani_2009.pdf
Скачиваний:
0
Добавлен:
28.03.2026
Размер:
12.86 Mб
Скачать

Chapter 8

Applications (I) – Music and Concert Hall

Acoustics

This chapter offers some applications of the central auditory signal processing model for temporal and spatial sensations as well as for subjective preferences. The measurement of pitches of notes sounded by a piano are discussed in Section 8.1. Examples of adaptive acoustic design of a public concert hall using global listener preference data are presented in Section 8.2. A seat selection system for enhancing individual listening experiences in a concert hall is discussed in Section 8.3. The preferred temporal conditions for music performance by cellists are discussed as it relates to acoustic design of the stage in Section 8.4.

8.1 Pitches of Piano Notes

It is well known that the source signal of a piano is a complex tone with mostly low-frequency harmonics. According to the method described in Section 6.2, source signals of pianos were analyzed and compared with those calculated using ratios of

neighboring notes in an equally tempered chromatic scale, i.e., semitone steps of

21/12.

For this study, source signals were picked up by a single microphone placed at the center position above a grand piano with its top lid opened at the usual angle for performance (Inoue and Ando, unpublished). The piano was equipped with an automatic performance system, was tuned before measurement, and produced source signals that could be reliably reproduced. Examples of the ACF analyzed for notes A1 (55 Hz), A3 (220 Hz), and A6 (1760 Hz) are shown in Fig. 8.1. It is clear that the delay times of the first peak τ 1 extracted from the ACF correspond well to the measured fundamental frequencies of notes A1 (55.2 Hz), A3 (219.7 Hz), and A6 (1785.7 Hz). Its amplitude φ1 is large enough (more than 0.8) and in this condition, a clear pitch is perceived. Calculated and measured pitches for all of 88 notes are shown in Figs. 8.1 and 8.2, and these values are listed in Table 8.1.

Y. Ando, P. Cariani (Guest ed.), Auditory and Visual Sensations,

143

DOI 10.1007/b13253_8, C Springer Science+Business Media, LLC 2009

 

144

8 Applications (I) – Music and Concert Hall Acoustics

Fig. 8.1 Examples of the ACF analyzed for source signals from a piano. (a) Note A1 of the pitch of 55 Hz. (b) Note A3 of the pitch of 220 Hz. (c) Note A6 of the pitch of 1760 Hz. The missing fundamental phenomena may be observed by τ1 extracted from the ACF, which corresponds to 55 Hz and 220 Hz. However, the pitch of 1760 Hz is observed by τ1 at just the fundamental frequency

8.1 Pitches of Piano Notes

145

Fig. 8.2 Relationship between calculated and measured pitches obtained by the value of τ1 extracted from the ACF of source signals from a “tuned” piano

Most of the

measured

values are in good agreement with calculated ones.

It is interesting,

however,

that measured pitches below G3 (207.6 Hz) were a

little bit higher than calculated ones. Below this pitch, the amplitudes of the fundamental frequency component in the spectrum analyzed were small and not significant, and thus the low pitch that one hears from these piano notes is primarily a missing fundamental phenomenon. For pitches below 55 Hz, no appreciable energy at the fundamental frequency in the measured spectrum was observed.

It is worth noting that the upper frequency limit of the ACF model for the pitches of missing fundamentals is about 1,200 Hz (Section 6.2.3). Above the frequency of this pitch, discrepancies between calculated and measured pitches grew large, reaching about 2%. At G7 (3135.9 Hz), in particular, a large discrepancy was observed, where a measured value of 1612.9 Hz corresponded roughly to half the calculated one. The octave error could conceivably have been caused by a mistake in tuning adjustment or, perhaps more likely, in the autocorrelation analysis (picking the second major peak rather than the first one). In some neural autocorrelation models (Cariani, 2004), this problem is largely avoided by analyzing the ACF for regular patterns of major interspike interval peaks rather than choosing the highest peak. The method uses a dense set of interval sieves that quantify the pattern strengths of all possible periodicities, which allows the model to estimate the relative strengths of multiple, competing pitches that may be heard in a given note or chord.

146

8 Applications (I) – Music and Concert Hall Acoustics

Table 8.1 Calculated pitches and the values measured by the ACF (τ1) of the 88-note signals for a piano that was said to be “tuned”

Note

Calculated pitch (Hz)

Measured pitch (Hz)

Difference (Hz)

 

 

 

 

A0

27.5

29.9

2.4

B0

29.1

31.5

2.4

H0

30.8

33.5

2.7

C1

32.7

35.1

2.4

Cis1

34.6

34.8

0.2

D1

36.7

36.9

0.2

Es1

38.8

39.2

0.4

E1

41.2

41.4

0.2

F1

43.6

43.6

0.0

Fis1

46.2

46.5

0.3

G1

48.9

49.2

0.3

Gis1

51.9

52.3

0.4

A1

55.0

55.2

0.2

B1

58.2

58.8

0.6

H1

61.7

62.1

0.4

C2

65.4

65.7

0.3

Cis2

69.2

69.4

0.2

D2

73.4

74.0

0.6

Es2

77.7

78.1

0.4

E2

82.4

82.6

0.2

F2

87.3

87.7

0.4

Fis2

92.4

92.5

0.1

G2

97.9

98.0

0.1

Gis2

103.8

104.1

0.3

A2

110.0

111.1

1.1

B2

116.5

117.6

1.1

H2

123.4

124.2

0.8

C3

130.8

131.5

0.7

Cis3

138.5

138.8

0.3

D3

146.8

149.2

2.4

Es3

155.5

156.2

0.7

E3

164.8

165.2

0.4

F3

174.6

175.4

0.8

Fis3

184.9

185.1

0.2

G3

195.9

196.0

0.1

Gis3

207.6

208.3

0.7

A3

220.0

219.7

+0.3

B3

233.0

232.5

+0.5

H3

246.9

246.9

0.0

C4

261.6

263.1

1.5

Cis4

277.1

277.7

0.6

D4

293.6

294.1

0.5

Es4

311.1

312.5

1.4

E4

329.6

327.8

+1.8

8.1 Pitches of Piano Notes

147

 

Table 8.1

(continued)

 

 

 

 

 

Note

Calculated pitch (Hz)

Measured pitch (Hz)

Difference (Hz)

 

 

 

 

F4

349.2

350.8

1.6

Fis4

369.9

370.3

0.4

G4

391.9

393.7

1.8

Gis4

415.3

416.6

1.3

A4

440.0

444.4

4.4

B4

466.1

465.1

+1.0

H4

493.8

495.0

1.2

C5

523.2

526.3

3.1

Cis5

554.3

558.6

4.3

D5

587.3

591.7

4.4

Es5

622.2

625.0

2.8

E5

659.2

666.6

7.4

F5

698.4

704.2

5.8

Fis5

739.9

740.7

0.8

G5

783.9

787.4

3.5

Gis5

830.6

826.4

+4.2

A5

880.0

892.8

12.8

B5

932.3

943.3

11.0

H5

987.7

980.3

+7.4

C6

1046.5

1041.6

+4.9

Cis6

1108.7

1111.1

2.4

D6

1174.6

1176.4

1.8

Es6

1244.5

1265.8

21.3

E6

1318.5

1333.3

14.8

F6

1396.9

1408.4

11.5

Fis6

1479.9

1492.5

12.6

G6

1567.9

1538.4

+29.5

Gis6

1661.2

1666.6

5.4

A6

1760.0

1785.7

25.7

B6

1864.6

1851.8

+12.8

H6

1975.5

2000.0

24.5

C7

2093.0

2083.3

+9.7

Cis7

2217.4

2272.7

55.3

D7

2349.3

2380.9

31.6

Es7

2489.0

2500.0

11.0

E7

2637.0

2631.5

+5.5

F7

2793.8

2857.1

63.3

Fis7

2959.9

3030.3

70.4

G7

3135.9

1612.9

+1523.0

Gis7

3322.4

3448.2

125.8

A7

3520.0

3703.7

183.7

B7

3729.3

3703.7

+25.6

H7

3951.0

4000.0

49.0

C8

4186.0

4347.8

161.8

A little less than one octave mistuned.