Approved for Pu Release Dsstrib Unlimited

Monaural sone functions are obtained for a no noise condition and under five levels of masking noise using the method of fractionation This method precludes the use of both ears in obtaining such functions as has been the case with dichotic loudness balance and other related procedures The obtained curves are found to parallel previously found masked func tions in one case and in another to show a more rapid acceleration at low levels but identical slopes above one sone When the power function exponent of a 1000 Hz tone is plotted against overall SPL of a masking noise a power transformation which parallels that found for speech in noise is obtained Although no numerical calcula tions are presented it appears that above 60 dB of noise the exponent grows as approximately the 0 16 power of the noise MONAURAL LOUDNESS FUNCTIONS UNDER MASKING INTRODUCTION Several investigations (Hellman and Zwis locki 1964 Lochner and Burger 1961 Scharf and Stevens J C 1958 Steinberg and Gardner 1937 Stevens 1966 1967) have shown that the subjective loudness function as well as the power function exponent are influenced by the presence of a masking noise Lochner and Burger (1961) using a method of dichotic loudness balance established loud ness functions for a 1000 Hz tone at 15 35 and 55 dB sound pressure level (SPL) Their technique was to alternate a tone in noise with a pure tone alone (both of 1 3 sec dura tion) subject (S) was required to balance the loudness of the two by adjusting the knob of the attenuator of the tone alone Lochner and Burger concluded Masking noise does not only produce a shift in the threshold of the pure tone but it also affects its loudness at higher levels From the present work it appears as though masking reduces the loudness of a pure tone at all levels by a constant amount (p 1707) Although Lochner and Burger did not re fer to the earlier study the conclusion that masking produces a constant reduction in loudness was an hypothesis originally ad vanced by Steinberg and Gardner (1937) and by many th St nb g and Ga dn were primarily interested in the differences between loudness judgments made m each ear of persons with a unilateral hearing de feet They used a method of loudness balance in which S s good ear was masked by wide band thermal noise which raised threshold 40 dB S adjusted a pure tone in the un masked ear to equal loudness with a tone of the same frequency in the masked ear Their loudness functions from normal ears m the presence of noise were similar to functions obtained from abnormal ears in quiet The loudness functions for the normal ears under masking and for the abnormal ears both show a loss of hearing at low intensities with hearing becoming normal or near normal as the stimulus intensity increased Figure 1 is the function plotted at 1000 Hz by Steinberg and Gardner to show the loud ness functions obtained for the normal ear the nerve deafened ear and the conductive deafened ear Hearing loss due to nerve deaf ness is most prominent at the lower mtensi ties whereas the pattern for conductive deaf ness is more or less uniform throughout the intensity range When Steinberg and Gardner plotted func tions for the normal ear under masking the loss of sensitivity was greatest for the lesser intensities and the pattern resembled the function for the nerve deafened ears Stem berg and Gardner concluded Since the var lable type of deafness (nerve deafness) oc curs when there is a loss in the number of fibers normally active the hearing loss caused by a masking sound will be expected to be of the variable type (p 14) Using the method of magnitude estimation Scharf and J C Stevens (1958) presented white noise at var ous levels (35 65 and 95 dB SPL) simultaneously with a 1000 Hz tone Their data indicated that near the masked threshold the loudness function resembles the function obtained near the unmasked thresh old Their contention was later supported by Hellman and Zwislocki (1964) who used both the method of loudness balance and numen al 1 udn balan (a m thod mpos d f aspects of both magnitude production and magnitude estimation) of a 1000 Hz tone un der masking Hellman and Zwislocki conclud ed (p 1627) that At sound intensities near masked threshold the masked loudness curves tend to become parallel to themselves as well as to the unmasked loudness curve They also concluded (in line with the earlier work of Steinberg and Gardner (1937) and with interpretations of Stevens and Davis (1938) that The effect of masking on the loudness function is essentially the same as that of a sensormeural hearing loss (p 1627) Thus, it was assumed that a given masking noise reduced the loudness of a masked stim ulus by a constant loudness since it mcapaci tated some of the total number of neural ele ments normally contributing to the loudness of a tone. If this is the case, Stevens (1966, p. 725) has noted that a constant subjective value C^) subtracted from the psychophysical loudness function would be all that is necessary to account for the loudness function under masking. In symbolic form: * = k(0-0o)0 where (,}/) is the subjective value, 0 is the stimulus value, 0O is the effective threshold, and (j3) is an exponent which varies with sense modality. However, Stevens goes on to say that this simple constant subtraction from (v|/) is not an adequate explanation of the phenomenon. Rather, the presence of a masking noise produces a "power transformation" upon the characteristic of the auditory system. A power transform simply refers to a change in the slope and intercept of the straight line obtained when using logarithmic coordinates (log-log) to represent the subjective loudness function. The principal purpose of the present investigation was to examine the monaural loudness function under masking conditions using a purely monaural technique. This method was chosen because it precludes the operation of any irrelevant inter-aural effects. However, a second purpose has relevant clinical diagnostic applications. That is, an abnormal growth of loudness (recruitment) may be detected monaurally rather than by means of dichotic loudness matches. This is of extreme importance when one ear is not normal, for then it is of course impossible to use dichotic loudness matches. METHOD Subjects: Ten adults free from hearing defects, graduate assistants and laboratory personnel, were used. No attention was paid as to whether S was naive or experienced in judging loudness, since both Stevens and Poulton (1956) and J. C. Stevens and Tulving (1957) have shown that untrained observer» were able to make consistent qualitative judgments of loudness on their first attempts which differed in no significant manner from the judgments of more experienced observers. Apparatus: A 1000-Hz tone and a wideband white noise were used throughout. All stimuli were presented monaurally through a Permaflux PDR8 earphone mounted in an MX-41/AR cushion and calibrated in a standard 6-cc coupler. The response characteristics of the transducer were found to be essentially flat from 2001200 Hz. The unused ear was covered by a dummy wooden earphone also mounted in an MX-41/AR cushion. Procedure: The fractionation method (Reese, 1943, Hanes, 1949) was used. Ss were presented with a number of standard intensities; they adjusted a comparison stimulus to one-half the subjective magnitude of the standard. Standard intensities of 46, 51, 61, 71, 81, 91, and 101 dB SPL were used. However, since Hellman and Zwislocki (1961) pointed out that loudness is related to sensation level