Subject: Re: phoneme recognition and stimulus length From: Bill Budd <bill(at)PSY.UWA.EDU.AU> Date: Tue, 20 Mar 2001 16:14:52 +0000Dick You raise an important issue in terms of the inconsistencies between the backward recognition masking functions obtained in some psychophysical studies, although I'm not sure that this point of controvesy really does suggest that backward recognition masking is not relevant to the 250ms+ durations Lazlo discusses. Essentially these inconsistencies relate to backward recognition masking studies that show threshold elevations as long as Lazlo's magic 250 ms (Massaro et al, see also Cowan 1984) and others showing more modest functions of masking delay with threshold elevations barely reaching 100 ms test-masker delays (Leshowitz, Watson and others). These shorter functons being more like more like those obtained in backward detection masking rather than backward recognition masking (a distinction perhaps not unrelated to the energetical/informational distinction recently discussed on this list). As you may be suggesting (I think?), some have argued that since seemingly minor variations in the backward masking procedure could reduce the magnitude and degree of masking and that therefore the backward recognition masking at test-mask delays greater than about 100ms may not reflect strictly perceptual processes (Watson, Kelly, & Wroton 1976; Leshowitz & Cudahy, 1973). However an alternative view argues that these inconsistencies can be adequately explained in that the extent of recognition masking is closely governed by the quality of the representation in memory against which the acoustic input of each trial is compared. If memory for the sensory properties of the test tone alternatives are in a highly stable state, then a relatively modest sampling of the input (in early or pre-perceptual sensory memory) will be sufficient to ensure accurate performance (Hawkins and Presson 1977b; Massaro & Idson, 1977). Indeed there is good evidence to support the proposal that both the magnitude and duration of masking will be reduced only if procedures which decrease the amount of test tone sensory information required for correct recognition are used. In fact, Loeb and Holding directly demonstrated that the amount of backward interference of test tone decreases in direct proportion to increases in practice time (1975). Further, decreasing stimulus uncertainty by randomising within block test-mask SOA, test tone frequency or ear of presentation between, rather than within, experimental blocks also significantly reduces the magnitude of the interfering effect of the masking tone (Massaro, 1975; Watson et al., 1976; Yost, Berg, & Thomas, 1976). The importance of varying stimulus conditions within rather than between experimental blocks was demonstrated by Hawkins et al. showing that when test-mask laterality was varied between stimulus blocks, only ipsilateral maskers reduced test tone recognition in backward recognition masking, while no masking was observed for either binaural or contralateral stimuli (Hawkins, Thomas, Presson, Cozic, & Brookmire, 1974). The absence of contralateral masking is more consistent with backward masking of detection rather than recognition. However when ear of presentation was varied within stimulus blocks, masking was observed under ipsilateral, contralateral and binaural masking conditions (Hawkins & Presson, 1977a). It was argued that the reduction in stimulus uncertainty in the blocked condition increased the stability of tonal memory and therefore reduced the potentially interfering effect of the masking stimulus. This need to avoid the potential influence of subject strategies and/or confounding by variations in trial to trial tonal memory by increasing stimulus uncertainty on backward recognition masking tasks as well as other perceptual tasks is generally acknowledged (Massaro, 1975; Rammsayer & Leutner, 1996). In my (admittedly limited!)experience, when these methodological issues are addressed, the 250ms asymptote is a robust result that is easily replicated using the same basic design as Massaro. In terms of Lazlo's question, I think the basic recogniton masking effect is relevant, in fact I think it was Cowan (1984;1995) who originally pointed out that the correspondence between the duration of the process/store underlying the masking effect and the average length of a syllable was not entirely coincidental! Cowan, N. (1984). On short and long auditory stores. Psychological Bulletin, 96(2), 341-370. Cowan, N. (1995). Attention and memory: An integrated framework. (Vol. 26). New York: Oxford University Press. Hawkins, H. L., & Presson, H. C. (1977a). Masking and perceptual selectivity in auditory recognition. In D. S (Ed.), Attention and Performance VI. Hillsdale, N.J.: Erlbaum. Hawkins, H. L., & Presson, H. C. (1977b). Masking level and attention in auditory recognition. In D. S (Ed.), Attention and Performance VII. Hillsdale, N.J.: Erlbaum. Hawkins, H. L., Thomas, G., Presson, J., Cozic, A., & Brookmire, D. (1974). Tonal specificity and masking in auditory recognition. Journal of Experimental Psychology., 103, 530-538. Kallman, H. J., Beckstead, J. W., & Cameron, P. A. (1987). Ipsilateral and contralateral masking of duration. Perception and Psychophysics., 43(31-37.). Kallman, H. J., & Morris, M. D. (1984). Backward recogntion masking as a function of ear of mask presentation. Perception and Psychophysics., 35(4), 379-384. Leshowitz, D., & Cudahy, E. (1973). Frequecny discrimination in the presence of another tone. Journal of the Acoustical Society of America., 54, 882-887. Loeb, M., & Holding, D. (1975). Backward interference by tones or noise in pitch perception as a function of practice. Perception and Psychophysics., 18, 205-209. Massaro, D. W. (1970). Preperceptual auditory images. Journal of Experimental Psychology, 85(3), 411-417. Massaro, D. W. (1972). Preperceptual images, processing time, and perceptual units in auditory perception. Psychological Review, 79(2), 124-145. Massaro, D. W. (1975). Backward recognition masking. Journal of the Acoustical Society of America, 58(5), 1059-1065. Rammsayer, T. H., & Leutner, D. (1996). Temporal discrimination as a function of masker duration. Perception and Psychophysics., 58(8), 1213-1223. Massaro, D. W., & Idson, W. L. (1977). Backward recognition masking in relative pitch judgments. Perceptual & Motor Skills, 45(1), 87-97. Watson, C. S., Kelly, W. J., & Wroton, H. W. (1976). Factors in the discrimination of tonal patterns. II. Selectrive attention and learning under various levels of stimulus uncertainty. Journal of the Acoustical Society of America., 60, 1176-1186. Yost, W. A., Berg, K., & Thomas, G. B. (1976). Frequency recognition in temporal interference task: A comparison among four psychophysical procedures. Perception and Psychophysics., 20, 353-359. At 12:38 19/03/01 -0500, you wrote: > Laszlo > In papers of the 1970s, Dom Massaro (in several papers) and Dave >Sparks both showed interference (drop in percent correct) in pitch >identification of tones just above a threshold difference apart by a >delayed following tone of intermediate frequency. Performance improved to >an asymptotic level (but not to unmasked level) a with an ISI of 250 to 250 >ms. Others [Barry Leshowitz & Ed Cudahy, (several papers in 1970s), and a >paper by Chuck Watson & Bill Wroton (1980s)] measured delta-F threshold as >a function of masker delay. At most, best, threshold performance was >elevated out to approximately 110 ms. I am not sure that backward >recognition masking is necessarily all that relevant to what is being >discussed. > Target tone duration is relevant to backward recognition masking. Don >Ronken (early 70s) found a type of backward recognition masking when >examining the frequency DL for brief tones - a trailing tone interfered >with performance, but a preceding tone, if anything, improved >performance. Finally, remember that the spectral bandwidth of stimulus is >an inverse function of duration, and the perception of brief segments of >tones are heard as clicks, tone pips, and finally as tones when duration is >increased. If dealing with brief stimuli, Roy Patterson's work on minimum >duration for different types of recognition is also definitely >relevant. However, the questions being discussed seem to be about >syllable-length stimuli, and thus time parameters that are longer than >those in the literature I have tried to summarize. > Dick Pastore > > >At 05:39 PM 3/19/01 +0100, you wrote: >>On Mon, 19 Mar 2001, Jont Allen wrote: >> >> > I dont know if this is what you are getting at, but have you looked at the >> > paper by Sadaoki Furui "On the role of spectral transitions for speech >> perception." >> > JASA, Oct. 1986, page 1016+ >> > >>It partly answers my questions, as they used truncated syllables as >>stimuli. But I would be more interested in the other direction: >>is there any additional gain in recognition performace if we use stimuli >>LONGER than a syllable? >>What made me wonder about this is the "backwards recognition masking" >>experiments of Massaro (unfortunately, I don't have the original papers, >>only a half-page review in a Ph.D. thesis by Brian Kingsbury). Their >>results say that masking has no effect if the target is longer than a >>syllable or if there is at least a syllable-long silent interval between >>the target and the masker. I would >>need a reinforcement of these results, but possibly from the opposite >>direction (i.e. not how recognition deteriorates from backwards masking >>but how recognition improves from "forward helping" - so to say). >> >> Laszlo Toth >> Hungarian Academy of Sciences * >> Research Group on Artificial Intelligence * "Failure only begins >> e-mail: tothl(at)inf.u-szeged.hu * when you stop trying" >> http://www.inf.u-szeged.hu/~tothl * > >--------------------------------------------------------------------------- ----- >Richard E. Pastore >Professor of Psychology and Linguistics >Binghamton University (SUNY University Center) >Binghamton, NY 13902-6000 > >http://psychology.binghamton.edu/faculty/pastore.htm >