Volume 11, Issue 3, January 1940
Index of content:
11(1940); http://dx.doi.org/10.1121/1.1916032View Description Hide Description
11(1940); http://dx.doi.org/10.1121/1.1916033View Description Hide Description
For a number of years hearing impairment has been expressed by the audiogram, which is a measure of the hearing loss for sounds which are just audible to the deafened ear, with the implication that impairment for threshold sounds indicates the impairment for sounds above threshold in the range of interest to the deafened. It is shown that the audiogram is not an accurate indication of hearing impairment for above threshold sounds in cases involving nerve deafness, when the loss in loudness sensation is taken as the criterion; but when the capacity to interpret speechsounds, as indicated by articulation tests, is the criterion, the audiogram does give an approximately correct indication of the hearing impairment. The relation of the audiogram to experiences in interpreting the speech of everyday life and the relation of hearing loss to absolute ear sensitivity are given.
11(1940); http://dx.doi.org/10.1121/1.1916034View Description Hide Description
Using apparatus designed to collect a large number of data in a short time, the following measurements have been made: peak and r.m.s. pressures in one‐eighth‐second intervals, and in various bands of frequencies up to 12,000 cycles per second, from the voices of six men and five women; comparison of r.m.s. pressures in one‐eighth‐ and one‐fourth‐second intervals, from a single male voice; and distribution of the instantaneous pressures in whole speech, from a single voice. Derived from these data are peak factors in one‐eighth‐second intervals, and frequency distribution of speech energy in long intervals. Both the absolute value and the distribution of energy are found somewhat different from previously published results.
11(1940); http://dx.doi.org/10.1121/1.1916035View Description Hide Description
Curves have been calculated for the total radiation resistance and reactance of a single circular piston as a function of frequency when another circular piston is vibrating in the same infinite plane, the distance of separation of the two pistons being a parameter. The curves shown are for equal velocity amplitudes and for phase differences of 0° and 180°. These results enable one to calculate the impedance of a piston when there are present an arbitrary number of pistons in arbitrary phases.
11(1940); http://dx.doi.org/10.1121/1.1916036View Description Hide Description
Application of Piezoelectric Vibration Pick‐Ups to Measurement of Acceleration, Velocity and Displacement11(1940); http://dx.doi.org/10.1121/1.1916037View Description Hide Description
11(1940); http://dx.doi.org/10.1121/1.1916039View Description Hide Description
11(1940); http://dx.doi.org/10.1121/1.1916042View Description Hide Description
A solution which is suitable for numerical computation is obtained for the classic problem of random flights in two dimensions with flights of unequal length. The solution is obtained by a development in series of the exact integral solution. The solution is applied to one aspect of the problem of the fluctuations present in the intensity of sound as it decays in a room. It is assumed that the room has reached a steady state of vibration under the excitation of a single frequency, and that the normal frequencies of the room are randomly distributed. The period of the decay is then broken into two sub‐periods, in the first of which there is some degree of coherence in the phases of the normal vibrations. The treatment presented in this communication applies to the second part of the decay, comprising all but the first π/k seconds, for which the phases as well as the frequencies of the normal vibrations may be considered as distributed at random. It is found that the relative extent of the fluctuations which may be traced to beats between normal modes of vibration, does not depend appreciably on the number of characteristic vibrations excited unless the number is so small that statistical considerations are not helpful, nor does it depend significantly on the size or shape of the room.
11(1940); http://dx.doi.org/10.1121/1.1916047View Description Hide Description
- PROGRAM OF THE TWENTY‐SECOND MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
11(1940); http://dx.doi.org/10.1121/1.1902145View Description Hide Description
For more than a century, wide differences of opinion have been held as to the function of the vocal cords in speech production. In the present paper, this problem is approached by utilizing high speed motion picture photography to determine the vibrations of the cords during phonation and a method of Fourier analysis to determine the characteristics of the sound waves.
By taking pictures at a rate of 4000 per second and projecting them at the normal viewing rate of 16 per second, the motion is slowed down by a factor of 250 to 1. Thus, if the cords execute 250 vibrations in one second, they appear to make one vibration per second and the details of their movements may be seen. Pictures suitable for monocular viewing, and also stereoscopic pictures showing the movements of the cords in full relief, have been obtained.
For the lower pitches, the cords appear to be completely relaxed, and there appears to be a phase difference between the movements of the lower and upper surfaces. They do not move as units, but more as folds or lips, the motion of the upper surfaces lagging behind that of the underneath surfaces. The tension increases as the pitch is raised, the cords elongate and tend to move more nearly as units with somewhat smaller amplitudes than for low pitches. The movements appear to begin toward the front end of the cords and to extend progressively toward the back, so that the motion has a horizontal as well as vertical phase difference.
The analysis of the vocal cord vibrations indicates a vibration form composed of a fundamental having a frequency equal to the vibration rate of the cords and overtones, which diminish in amplitude with increasing order or number of the overtone. Such vibration forms do not possess the distinguishing characteristics of the different speech sounds. The latter characteristics are produced by the passage of the sound waves through the throat, nose and mouth cavities.
The paper will be illustrated by motion pictures showing the movements of the cords in ultra‐slow motion, and by variable area recordings of the sound waves and their corresponding analyses. Stereoscopic motion pictures will be available for limited viewing so that the cord movements may be seen in relief.
11(1940); http://dx.doi.org/10.1121/1.1902146View Description Hide Description
During recent years several studies of vocal pitch in speech of various types have been made at the University of Iowa by means of measurement of the fundamental sound wave frequency from phonograph records.
One investigation was concerned with the voices of well‐known actors and actresses, disclosing among other data median pitch levels of 140.7 and 233.2 c.p.s. for the respective groups, with total pitch ranges in the neighborhood of 12 tones.
In a second experiment five actors simulated different emotions successfully in five readings of the same material. Simulation of contempt was found to be characterized by a range of 10.5 tones about a median of 124.3 c.p.s.; measurements of simulations of anger revealed a mean range of 10.3 tones about a median of 228.8 c.p.s., and a more rapid rate of pitch change [25.6 tones/sec.] than the other emotions studied: the highest median pitch level [254.4 c.p.s.] and widest range [11.2 tones] were found in fear; in grief the mean range was 9.0 tones with a median of 135.9 c.p.s., the mean extent of inflections [1.7 tones] was the narrowest and the rate of pitch change [15.6 tones/sec.] the slowest of the five emotions; the narrowest range [7.8 tones] and the lowest median pitch level [108.3 c.p.s.] were measured in simulations of indifference. Most of the subjects employed total pitch ranges of over 18 tones in the simulation of all five emotions. Another study investigated the pitch usage of carefully selected superior male speakers reading a 55‐word test passage, and found a mean total range of 10.5 tones about a median level of 132.1 c.p.s. In the latter measure individual subjects were grouped closely about this value. The pitch usage of these subjects in speech was related to their total singing ranges including falsetto, these ranges averaging 20.4 tones. It was found that the pitches used in the measured samples of speech lay within the lower fractions of the singing ranges, and that the ratio of the extent in tones between the lowest sung tones and the median pitch levels in speech to the total singing ranges averaged 0.25.
In a fourth experiment the same trained subjects re‐read the test passage in response to instructions to read with more variability of pitch, with less variability of pitch, at a higher pitch level and at a lower pitch level. Extremes were avoided. Measurements revealed that the subjects followed instructions, and certain characteristic features of pitch variability were disclosed. Of interest was the finding that pitch level and pitch variability tended to vary concomitantly, i.e., increased variability was accompanied by a rising pitch level and vice versa.
Another investigation made a preliminary exploration of the pitch aspects of voice change in the male. Ten‐year‐old, 14‐year‐old, 18‐year‐old and adult subjects were studied. The two younger groups were found to have median pitch levels in the neighborhood of the values reported for adult females, while the 18‐year‐old subjects used levels approximately equal to those of the adult male group. Wave‐to‐wave measurement of so‐called “voice breaks” revealed several interesting features. (1) Almost as many breaks were found at 10 years as at 14 years. (2) Downward breaks, although outnumbered by upward breaks, were frequent. (3) Breaks averaged approximately one octave in extent. (4) All breaks measured were made up to and down from the median pitch levels of speech or their vicinity; none were found above these levels. However, in two atypical cases of voice change, not included in the above groups, the breaks departed from this typical location below the median.
These experiments are immediately suggestive of further researches, some of which are under way at present.
11(1940); http://dx.doi.org/10.1121/1.1902148View Description Hide Description
Several types of measurement are reported. R.m.s. pressures in one‐eighth‐second intervals, from the speech of a single voice, are compared with similar measurements in one‐fourth‐second intervals. The values obtained in the first case are only slightly more widely distributed than in the second, and the difference is least when attention is confined to high frequencies.
To earlier r.m.s. measurements, on the voices of six men and five women, is applied the relation found more recently between pressure at a point and total voice power. The resulting curves represent average spectra of voice power for the men and women.
Measurements of the distribution of instantaneous pressures in speech are also given.