backward masking: detection versus recognition

Backward ' ß mask,ng. Detection versus recognition David E. Bland and David R. Perrott

Psychoacoustics Laboratory, California State University, Los Angeles, California 90032 (Received 23 June 1977; revised 12 December 1977)

The backward detection masking of 10-ms tonal targets by a 150-ms tonal mask was contrasted with the backward recognition masking of the same tones by the same mask. The target-mask interval required for 75% correct performance was about eight times as great in the recognition condition as in the detection condition. Furthermore, a generalized improvement in performance occurred over the initial course of training in both the detection and recognition conditions. In a second experiment it was found that a remote mask produced greater backward recognition masking but less backward detection masking. That these differences were observed with the same subjects, at the same level of training, and with identical stimuli, indicates that procedural differences alone cannot account for the differences between backward detection and recognition masking.

PACS numbers: 43.66.Mk, 43.66.Fe, 43.66.Lj

INTRODUCTION

In recent years a great deal of interest has surround- ed the question of whether backward recognition masking (BRM) represents similar processes as underly simultaneous masking and backward detection masking (BDM), Massaro (1970, 1971, 1972, 1973, 1975) has reported BRM data which differ sharply from those reported in the BDM literature (Elliott, 1962a, 1962b, 1967). Massaro finds that a trailing mask disrupts recognition performance with target-mask delay intervals as great as 250 ms irrespective of whether the mask is delivered ipsilaterally or contralaterally and irrespective of whether the mask spectrum is similar to or disparate from that of the target.

BDM effects, on the other hand, extend out to delays of only about 50 ms, are highly dependent on the spec- tral similarity of target and mask, and are greatly reduced if a contralateral backward mask is used.

Recently, several investigators (Cudahy and l.esho- witz, 1974; Leshowitz and Cudahy, 1973; Sparks, 1976; Watson, Kelly, and Wroton, 1976; Yost, Berg and Thomas, 1976) have reported resul[s which demonstrate that the BRM effects which Massaro has shown are

highly dependent upon such procedural variables as training and stimulus uncertainty. It thus becomes ex- ceedingly important that in making direct comparisons between BDM and BRM the psychophysical procedures used to measure each be as nearly the same as possible. If in a given psychophysical procedure the differences between BRM and BDM are maintained then one would

have a clear indication that the underlying processes are dissimilar.

Such comparisons are not now possible. BRM tasks are typically performed using stimulus parameters which do not allow for BDM. In the present study we develop a paradigm which allows for a direct comparison between BDM and BRM using identical stimuli.

I. GENERAL METHODS

Signals were led from three audio oscillators to separate electronic switches (Grason-Stadler. 1287).

After gating, the signals were led to the appropriate at-

tenuating, mixing, and impedance matching devices, and finally to a single earphone (Grason-Sadler, TDH- 49). All testing was conducted in a soundproof room (IAC 1200).

In both experiments 1 and 2 the subjects were tested in two experimental paradigms. In the first, the detection task, a target event was defined as a 10-ms sine-wave pulse presented at 50 dB SPL (re 0.0002 dyn/cmU). The tonal target was followed by a 150-ms tonal mask presented at 80 dB SPL. The rise-decay time was set at 1 ms for both target and mask. The presentation probability of the target was 0.5.

In the recognition task a tonal target occurred on every trial (one of two equally probable tonal pulses)

90-

50- RECOGNITION

• I i l[

TARGET-MASK DELAY INTERVAL

( MSEC )

FIG. 1. Percent correct plotted as a function of the target- mask delay interval in the detection (circles) and recognition (triangles) tasks. The abscissa is plotted logarithmically. Untilled points are from the first 750 trials of training; filled points are from the second 750 trials of training. The broken horizontal line represents detection performance with no mask and the solid horizontal line recognition nerformance with nn

mask.

1215 J. Acoust. Soc. Am. 63(4), Apr. 1978 0001-4966/78/6304-1215500.80 ¸ 1978 Acoustical Society of America 1215

Redistribution subject to ASA license or copyright; see http://acousticalsociety.org/content/terms. Download to IP: 141.209.144.122 On: Tue, 02 Dec 2014 17:57:36

1216 Letters to the Editor 1216

followed by the tonal mask. The subject was required to indicate which tone had been presented.

In both paradigms the target frequencies employed were 500 and 540 Hz with a 520-Hz mask. •

In all phases of the study a redundant feedback procedure was used. Immediately upon the subjects response feedback lights indicated whether the subject was correct or incorrect. The event presented on that trial s was repeated 800 ms later. The next trial occurred 1.65 s after the subjects last response.

Experiment 1. A direct comparison between backward detection masking and backward recognition masking.

Method. Two subjects were used. Both had demon- strated hearing within normal limits. Prior to the start of testing, the subjects were trained on the recognition task without the tonal mask until at least 90% correct performance was achieved.

A fixed test-block procedure was employed such that only one combination of task condition, target frequency (detection condition), and target-mask delay interval was used in each block of trials. Six such blocks of 50

trials were used for each combination. (Three hundred trials per combination of task condition, target frequency, and delay interval. )

Target-mask delay intervals of 13, 18, 23, 38, and 58 ms were used in the detection condition and 38, 58, 110, 160, and 310 ms in the recognition condition.

II. RESULTS--EXPERIMENT I

Figure 1 plots percent-correct detection and recognition performance as a function of the target-mask delay interval. In the detection task performance increased from about 64% at a delay of 13 ms to about 80% at a

90-

50

•0

DETECTION

RECOGN IT ION

I I I I i i I I0 18 28 38 85 160 310

TARGET-MASK DELAY INTERVAL

( MSEC )

FIG. 2. Details same as Fig. 1 except that untilled points represent trials with a 520-Hz mask and filled points trials with a 700-Hz mask.

delay of 58 ms. This is in contrast to the recognition performance which remained at chance at delays of 58 ms. l•ecognition performance did not exceed 80ø/0 until target-mask delays of 310 ms were tested.

Interestingly, the target-mask delay interval required for 75% correct performance 3 decreased from 29.2 ms over the first 750 trials (open circles) to 21.8 ms in the second 750 trials (a 25% change). The 75% correct criteria in the recognition condition decreased from 225 to 175 ms (a 220/0 change).

Experiment 2. Remote backward masking in the recognition and detection paradigms.

Method. Two subjects were used. One subject had participated in the previous study. The second received practice similar to that given the first subject.

The test procedures used in experiment 2 were the same as those used in experiment 1 except for the fol- lowing changes.

In addition to the 520-Hz tonal mask a 700-Hz tonal

mask was also used. As in experiment 1, a fixed test- block procedure was used. In the recognition condition, three blocks of 55 trials were given. In the detection condition six blocks of 30 trials were given. The first five trials of each block were treated as warm-up trials and were not scored.

In the recognition condition, target-mask delay intervals of 38, 85, 160, and 310 ms were used. In the detection condition delay intervals of 10, 18, 28, and 38 ms were used.

In a second phase of experiment 2 mask frequencies of 880 and 1060 Hz were tested using selected onset disparities. Onset disparities of 18 ms in the detection condition and 160 ms in the recognition condition were used.

III. RESULTS--EXPERIMENT 2

Figure 2 contrasts.the effects of the 520- and 700-Hz masks upon recognition and detection performance. Detection performance improved with a remote mask (700 Hz) and recognition performance actually becomes worse with this s•tme mask. Similar conclusions are

apparent from an inspection of Fig. 3 in which a systematic decrease in recognition performance and a systematic increase in detection performance is observed as the mask frequency is increased.

IV. DISCUSSION

The stimulus parameters used in the present study were such that typical BDM performance was observed. Yet, when the subjects were required to recognize rather than merely detect the acoustic target, much longer delay intervals were required for reliable performance. Moreover, a remote mask produced greater masking than a near mask. That these results occurred

with the same subjects, at the same state of training, and with identical stimuli indicates that procedural differences alone cannot account for the differences between BDM and Bi•M.

J. Acoust. Soc. Am., Vol. 63, No. 4, April 1978


1217 Letters to the Editor 1217

100

DErECTI ON 00

c)

z

• 70

6O

520 7O0 880 1060

MASK FREQUENCY

FIG. 3. Percent correct as a function of mask frequency at a target-mask delay interval of 18 ms in the detection task (untilled circle) and 160 ms in the recognition condition (filled circles).

Our results concerning increased BRM with a remote mask are in agreement with data reported by Hawkins et al. (1974) and Massaro, Cohen, and Idson (1976).

However, using vastly different stimulus parameters, task conditions, and states of training, other investigators (Divenyl and Hirsh, 1975; Loeb and Holding, 1975; Massaro, 1970; Sparks, 1976) have reported results which conflict not only with our own but with each other. Unfortunately, without being able to measure BDM under similar circumstances it is •mpossible to determine if such procedural variables would produce similarly paradoxical changes in BDM performance.

Sparks (1976) and Loeb and Holding (1975) have shown that the BRM effect can be reduced or even eliminated

with extensive practice. Similarly, Watson, Kelly, and Wroton (1976) and Yost, Berg, and Thomas (1976) re- port that under conditions of minimal stimulus uncertainty the BRM effect can be virtually eliminated. Sim- ilar findings have been reported by Pfafflin (1968) in the case of simultaneous detection masking.

We found that a generalized improvement in performance occurred over the initial course of training. The target-mask delay required for 75% correct performance decreased by 25% and 22% in the detection and recognition tasks, respectively. It remains to be seen whether this generalized change in performance would be upheld under the extremes of training imposed by other investigators (e.g., Sparks, 1976).

1Performance in the detection and recognition tasks was also measured without a mask under otherwise identical conditions.

2In the recognition condition the event presented in a given trial could be either the 500- or 540-Hz tone. In the detec-

tion condition the event presented on a given trial could be either the tonal target or nothing. The trailing mask was never presented as part of the feedback,

3The delay interval required for 75% correct performance was determined by connecting the points with straight lines and interpolating.

Cudahy, E., and Leshowitz, B. (1974). "Effects of a contralateral interference tone on auditory recognition," Percept. Psychophys. 15, 16-20.

Divenyl, P. L., and Hirsch, I. J. (1975). "The effect of blanking on the identification of temporal order of three- tone sequences," Percept. Psychophys. 17, 246-252.

Elliott, L. L. (1962a). "Backward masking: monotic and dichotic conditions," J. Acoust. Soc. Am. 34, 1108-1115.

Elliott, L. L. (1962b). "Backward and forward masking of probe tones of different frequencies," J. Acoust. Soc. Am. 34, 1116-1117.

Elliott, L. L. (1967). "Development of auditory narrow band frequency contours," J. Acoust. Soc. Am. 42, 143-153.

Hawkins, H. L., Thomas, G. B., Presson, J. C., Cozic, A., and Brookmire, D. (1974). "Precategorical selective attention and tonal specificity in auditory recognition," J. Exp. Psychol. 103, 530-538.

Leshowitz, B., and Cudahy, E. (1973). "Frequency discrimination in the presence of another tone," J. Acoust. Soc. Am. 54, 882-887.

Loeb, M., and Holding, D. H. (1975). "Backward interference by tones or noise in pitch perception as a function of practice," Percept. Psychophys. 18, 205-208.

Massaro, D. (1970). '"Preperceptual auditory images," J. Exp. Psychol. 85, 411-417.

Massaro, D. (1971). "Effect of masking tone duration on preperceptual auditory images," J. Exp. Psychol. 87, 146-148.

Massaro, D. (1972). "Preperceptual images, processing time, and perceptual units in auditory perception," Psychol. Rev. 79, 124-145.

Massaro, D. (1973). "A comparison of forward versus backward recognition masking," J. Exp. Psychol. 100, 434-436.

Massaro, D. (1975). "Backward recognition masking," J. Acoust. Soc. Am. 58, 1059-1065.

Massaro, D., Cohen, M. M., and Idson, W. L, (1976). "Rec- ognition masking of auditory lateralization and pitch judg- ments," J. Acoust. Soc. Am. 59, 434-440.

Pfaffiin, S. M. (1967). "Detection of auditory signal in re- stricted sets of reproducible noise," J. Acoust. Soc. Am. 43, 487--49O.

Sparks, D. W. (1976). "Temporal recognition masking--or interference ?" J. Acoust. Soc. Am. 60, 1347-1353.

Yost, W. A., Berg, K., and Thomas, G. B. (1976). "Fre- quency recognition in temporal interference tasks: a comparison amohg four psychophysical procedures," Percept. Psychophys. 20, 353-35 9

Watson, C. S., Kelly, W. J., and Wroton, H. W. (1976). "Factors in the discrimination of tonal patterns. H. Selective attention and learning under various levels of stimulus uncertainty," J. Acoust. Soc. Am. 60, 1176-1186.

J. Acoust. Soc. Am., Vol. 63, No. 4, April 1978


backward masking: detection versus recognition

Documents