Robert J. Porter

Multipath compensation in shallow water environments using a virtual receiver

An acoustic technique for compensation of signal distortion due to propagation is developed in analogy with an astronomical technique in which light from a bright “guide star” is used to correct atmospheric aberration of weaker objects that are nearby in the angular sense. The acoustic technique investigated here uses a vertical array to receive both the signal from a broadband “guide source” and an unknown “objective source” which propagates over a partially shared path. The algorithm is a spatial–temporal cross correlation of the two signals and is termed a “virtual receiver” as the output approximates the signal that the unknown source would produce at the location of the guide source. By strategically locating the guide source, many of the distorting effects of an unknown propagation region can be removed, including mode coupling and multipaths. The virtual receiver can also be combined with matched field processing techniques to estimate the unknown source location. Results are given for a variety of...

Discrimination of formant transition onset frequency: Psychoacoustic cues at short, moderate, and long durations

Two experiments determined the just noticeable difference (jnd) in onset frequency for speech formant transitions followed by a 1800-Hz steady state. Influences of transition duration (30, 45, 60, and 120 ms), transition-onset region (above or below 1800 Hz), and the rate of transition were examined. An overall improvement in discrimination with duration was observed suggesting better frequency resolution and, consequently, better use of pitch/timbre cues with longer transitions. In addition, falling transitions (with onsets above 1800 Hz) were better discriminated than rising, and changing onset to produce increments in transition rate-of-change in frequency yielded smaller jnds than changing onset to produce decrements. The shortest transitions displayed additional rate-related effects. This last observation may be due to differences in the degree of dispersion of activity in the cochlea when high-rate transitions are effectively treated as non-time-varying, wideband events. The other results may reflect mechanisms that extract the temporal envelopes of signals: Envelope slope and magnitude differences are proposed to provide discriminative cues that supplement or supplant weaker spectrally based pitch/timbre cues for transitions in the short-to-moderate duration range. It is speculated that these cues may also support some speech perceptual decisions.

Nonlinear Dynamical Systems in Speech Perception and Production

Research examining the nonlinear, dynamical-systems aspects of speech production and perception is reviewed. Examples are provided from studies of phonation, speech respiration, vocal tract acoustics, speech gestures, and auditory-phonetic perception. Results support a conclusion that nonlinear, dynamical system behaviors and principles are represented in speech production and perception at several levels of system behavior. The results are interpreted as suggesting that the communicative distinctions in speech are represented in task-relevant, order parameters which are preserved over a wide range of sets of task-irrelevant, control parameters. The control parameters, in turn, define the domains of behavior for each of the relevant acoustic, biological, or psychological systems. More specifically, it is proposed that speech order parameters may be understood in terms of collective variables defined, in production, over the degrees of, and opening/closing rates of, the vocal tract and, in perception, over the rates, amounts, and relative phases of the broad-band temporal modulations in speech signals.

Dichotic and monotic interactions between speech and nonspeech sounds at different stimulus onset asynchronies

Subjects’ identification of stop-vowel “targets” was obtained under monotic and dichotic, forward and backward, masking conditions. Masks, or “challenges,” were another stop-vowel or one of three nonspeech sounds similar to parts of a stop-vowel. Backward masking was greater than forward in dichotic conditions. Forward masking predominated monotically. Relative degree of masking for different challenges suggested that dichotic effects were predicated on interference with processing of a complex temporal array of auditory “features” of the targets, prior to phonetic decoding but subsequent to basic auditory analysis. Monotic effects seemed best interpreted as dependent on relative spectrum levels of nearly simultaneous portions of the two signals.

The Effect of Temporal Overlap on the Perception of Dichotically and Monotically Presented CV Syllables

When stop consonant‐vowel syllables, 390 msec in duration, are presented dichotically, with an onset asynchrony of 50–70 msec, subjects identify the temporally lagging stop more accurately than the leading. When the same pairs are presented monotically, the leading stop has the advantage. This difference has been interpreted in terms of central versus peripheral masking [M. Studdert‐Kennedy, D. Shankweiler, and S. Schulman, J. Acoust. Soc. Amer. 48, 599–602 (1970)]. Dichotically, the signals travel separate peripheral pathways, and central masking effects are revealed. Monotically, the lagging signal is masked peripherally by the overlapping leading signal. This experiment was designed to determine the effect that the elimination of the temporal overlap would have on the dichotic and monotic masking functions. Syllables 75 msec long were presented at six asynchronies between 0 and 100 msec. The dichotic masking function was unaffected, whereas the monotic lead advantages were substantially reduced at the ...

The “lag effect” in dichoptic viewing

Abstract Ten subjects were asked to report both of two different consonant-vowels (CVs) presented to the same foveal area, but in different eyes (“dichoptically”). Stimuli were presented at stimulus-onset asynchronies (SOAs) ranging from 0 to 150 msec in 25-msec steps. Correct identifications were significantly depressed for the eye receiving the leading stimulus at SOAs of 25 to 75 msec. Monoptic data from three subjects indicated no significant reductions in correct identification as a function of SOA. The dichoptic results can be understood in terms of current theories of visual backward masking and are similar to the “lag effect” observed with dichotic listening to speech stimuli. Similarity of results for the two modalities suggests a similar “two-process” explanation may underlie both phenomena.

More support for rate‐based discrimination of second formant transitions

Two prior studies [Besing et al., J. Acoust. Soc. Am. 81, 535 (1987); Porter et al., J. Acoust. Soc. Am. Suppl. 1 82, S81 (1987)] investigated differential thresholds for onset frequency of second formants (30‐ms transitions) and for transition duration (60‐ and 120‐ms transitions). The extent of both rising and falling (to 1800‐Hz) transition was systematically varied to yield differing rates‐of‐change for stimuli. Transition discrimination was examined for formants isolated and in a first formant context. The question addressed was whether discrimination varied as a function of the simple stimulus dimensions of frequency and time or in relation to the more complex domain of the rate‐of‐change of frequency over time. Several aspects of the results supported a rate‐of‐change interpretation. The present study further tests that interpretation by extending observations to discrimination of onset frequency for 60‐ and 120‐ms transitions. Results for six listeners indicate that onset‐frequency discrimination of 60‐ms transitions is also based on rate‐of‐frequency change as was the case for 30‐ms transitions; discrimination of 120‐ms transitions may be, in part, frequency based. In addition, rate‐based discrimination appears to be nonlinear in that higher rates‐of‐change yield smaller difference limens relative to a reference than do lower rates‐of‐change. [Supported by NIH‐NINCDS and the Louisiana Lions Eye Foundation.]

Searching for attractors in speech and nonspeech respiration

Our recent study of breathing during speech revealed that a nonlinear equation, originally proposed as a description of respiration during quiet breathing, may also provide a description of respiratory behavior during speech. This equation interrelates variables such as lung volume, airway resistance, and the duration of the utterance, and captures commonalities in the basic organization of the respiratory system during what have been considered different tasks or modes of behavior (i.e., speech and nonspeech). Further explorations of these data suggest that the mechanism underlying the behavior may be viewed as an attractor, possibly chaotic. The presence of such an attractor in respiratory patterns is supported by the findings of other researchers and is consistent with the adaptive nature of the system during speech and nonspeech tasks [e.g., M. P. Sammon and E. N. Bruce, J. Appl. Physiol. 70(4), 1748–1762 (1991)]. Descriptions of the attractor and some of its characteristics will be presented.

Examining a respiratory control model using an allometric analysis of speech breathing

The value of examining various aspects of speech production from a nonlinear, dynamic systems point of view has been suggested by recent reports [e.g., J. A. S. Kelso and B. Tuller, J. Physiol. 246, R928–R935 (1984)]. Utilizing principles from physical biology the relationships between speech breathing, length of utterance, and upper airway resistance were examined. A simplified biological state equation served as the model [P. N. Kugler and M. T. Turvey, Information, Natural Law, and the Self‐assembly of Rhythmic Movement (Erlbaum, Hillsdale, NJ, 1987)]. Speaking tasks were varied on reiterant versus meaningful speech, length of utterance, normal fundamental frequency speech, whispered speech, and monotonic speech. Upper airway resistance values were calculated as ratios according to differential airflow as measured by a Rothenberg mask. Duration and depth of inspiration and expiration were measured with a Respitrace, and acoustic information was obtained from digitized speech samples using the program C...

Dichotic and Monotic Masking of CV Syllables by CV Second Formants with Different Steady-State Durations

Listeners were asked to identify 295-ms synthetic CVs (45-ms transitions + 250-ms steady states) in the presence of monotically or dichotically competing isolated second-formant ‘bleats’ from CVs. Ble