Cinematographic annual : 1931 (1931)

246 CINEMATOGRAPHIC ANNUAL Other phenomena must be considered. Noises due to external sources, audience, and ventilating and projecting equipment, ranging from 20 to 40 decibels above threshold according to our measurements, are usually present in an auditorium, sometimes necessitating an increase in loudness for good articulation. Coughing or nearby conversation occasionally raise the noise level above these figures. It has been found that ordinary speech is readily understood in the presence of what may be termed a usual amount of room noise.3 Everyday experience bears this out. The intelligibility of individual consonant sounds whose intensities are comparatively low will be reduced but with continuous conversation there is slight interference. Hence, as a rule, the level of reproduction in the theatre should be satisfactory when speech is heard at its usual levels. Such operation should have the advantages of being more restful than operation at unnaturally high intensities, besides being conducive to greater attention. In general we may neglect theatre noise as a factor in our problem, remembering that it is undesirable always and bothersome on occasion. In the case of music with its complexity of sounds there is not available such a complete analysis as has been made for English speech. The acoustic power flow at threshold for speech is of the order of 4x1 0~10 microwatts per square centimeter. This figure was obtained from the phonetic power, that is, maximum value of power averaged over an interval of one one-hundredth of a second for a given sound, of 40 microwatts measured by Sacia and Beck for the loudest vowel sounds when spoken in an ordinary manner.* It compares with the same value for the threshold of the pure tone at which the ear is most sensitive. Because of this it may seem that our selection is too low, but there seems no reason why the ear should not behave differently for complex sounds near threshold than it does for single frequencies. Accepting this value we then select an engineering limit of 80 decibels above it as the maximum mean power flow to be provided in theatres by electrical sound reproducing equipment. Such a selection should allow for variation in different types of speech as well as music. Our measurements on symphony orchestras in large theatres indicate that this figure is seldom exceeded. With speech the instantaneous peak powers may be 13 decibels above this as we shall see later. The conditions of noise and maximum power discussed here therefore limit the intensity range worked with to about 50 decibels. The maximum flux of acoustic power is thus 4X10-2 microwatts per square centimeter. Fig. 1 shows what this means in terms of sensation and loudness levels for single frequencies.3 The falling off at low frequencies is very marked in the case of the sensation curve. The loudness curve is apparently the really important one since pre 3Steinberg, J. C: "Effects of Distortion upon the Recognition of Speech Sounds," Journal Acoustical Society America, 1 (1929), No. 1, p. 121. Also Fletcher "Speech and Hearing," p. 266. 4Sacia, C. F. ; "Speech Power and Energy," Bell System Technical Journal, 4 (1925), No. 4, p. 627. Sacia, C. F. and Beck, C. J.: "The Power of Fundamental Speech Sounds." Bell System Technical Journal, 5 (1926), No. 3, p. 463. 5Fletcher: "Speech and Hearing," p. 230.