In the design of insulating wall and sound-reproducing system, the scoustic power of sound source and the required power for audio device are fundamental data to determine the specification. To obtain the reference data fitting to this purpose, acoustic power of orchestra music was investigated, since it is typical and generates the largest peak power among common renditions in auditorium. There are many theoretical sound-field equations proposed for the measurement of the sound source power with its relation to the sound pressure at a distance from the source, however, their applications are limited to the case of a nondirectional sound souce. To simulate wide arrangement of orchestra instruments and then to approximate sound pressure with the measurement of an orchestra, a model of orchestra was composed of fine nondirectional sound sources as shown in Fig. 2. Using this model, powers of orchestras will be estimated later. In this paper, firstly, the experimental verification of the theoretical sound field given by Eq. (6) for a single nondirectional sound source is shown in Figs. 3, 4, 5. The sound field Eq. (6) was derived from the view point of geometrical acoustics: The summation was carried out for all incident sounds to the receiving point prescribed by the law of geometrical reflection up to the upper bound term k, which is indirectly given by ub, the ratio of partial sum for k-1 terms to that for k terms in Eq. (6). From other consideration, it was found that ub=0. 99 is practically large enongh to obtain precise sound pressure by Eq. (6). In the region of high frequency, absorption of sound in air was taken into account and theoretical sound pressure was calculated with the absorption coefficient m=10^&lt-3&gt to get good agreement with measured sound pressure level shown by dots in Fig. 5. Secondly, the total sound pressure produced by spatially distributed sound sources that simulate orchestra instruments in the assembly was examined by comparing theoretical results with measured ones. Both computed and measured results in Fig. 6 were obtained under the condition that each sound source radiated equal power noise consisting of three half-octave-band noises whose outputs followed mean orchestra spectrum. Since another good conformity between these two sound fields is seen in Fig. 6, Eq. (6) is believed sintable for the evaluation of the power of distributed sound sources as well. Thirdly, to prove the validity of these distributed sound sources as an orchestral model, the measured rms sound pressures of an orchestra music were compared with the sound field by Eq. (6) and a good agreement between them was obtained as shown in Fig. 7. It should be noted that to attain the same uni-directivity condition when the orchestra music was picked up, the correction procedure from non to uni-directivity depicted in Fig. 7 was applied in advance to the theoretical sound field that is computed under the condition of nondirectivity. Lastly, the follwing acoustic power of an orchestra is obtained by applying Eq. (6) to the orchestra model. P. I. TCHAIKOVSKY'S SYMPHONY NO. 6 (69-piece orchestra) MEAN 16mW PEAK 13W. If acoustic and geometrical conditions of hall and relative location of orchestra to microphone are assumed the same, acoustic powers of other orchestras will be estimated in the similar way referring to the sound pressures given in the literature below. For example, M. RAVEL'S BOLERO (107-piece orchestra) MEAN 83mW PEAK 52W. S. Ehara: "Instantaneous Pressure Distributions of Orchestra Sounds, " J. A. S. J. , Vol. 22, No. 5 (Sept. 1966) pp. 276-289.