Filtered Pulse Trains Research Articles

Using binaural beat sensitivity to explore mechanisms of bimodal temporal envelope beat sensitivity

Current cochlear implant (CI) fitting strategies aim to maximize speech perception through the CI by allocating all spectral information across the electrode array without regard to tonotopic placement of each electrode along the basilar membrane. For patients with considerable residual hearing in the non-implanted ear, this approach may not be optimal for binaural hearing. This study aims to explore fitting procedures in which CI maps better complement information from the acoustic ear by reducing the frequency mismatch between them. We investigate the mechanisms of binaural temporal-envelope beat sensitivity in normal-hearing listeners using bandpass filtered pulse trains with parameters including stimulus level, filter bandwidth, filter slope, and spectral overlap using bandpass filtered pulse trains. We find the minimum baseline interaural timing difference and spectral mismatch that normal-hearing listeners can tolerate while maintaining their ability to detect interaural timing differences. Initial results consistently demonstrate maximum sensitivity to binaural beats when place of stimulation is matched across ears. The outcomes of this study will provide new information on binaural interactions in normal-hearing listeners and guide methodology for incoming single-sided-deafness patients as we adjust their CI maps in an effort to reduce the frequency-mismatch. [Work supported by NIH grant F32DC016815-01.]

Temporal Regularity Detection and Rate Discrimination in Cochlear-Implant Listeners

Cochlear implants (CIs) convey fundamental-frequency information using primarily temporal cues. However, temporal pitch perception in CI users is weak and, when measured using rate discrimination tasks, deteriorates markedly as the rate increases beyond 300 pulses-per-second. Rate pitch may be weak because the electrical stimulation of the surviving neural population of the implant recipient may not allow accurate coding of inter-pulse time intervals. If so, this phenomenon should prevent listeners from detecting when a pulse train is physically temporally jittered. Performance in a jitter detection task was compared to that in a rate-pitch discrimination task. Stimuli were delivered using direct stimulation in cochlear implants, on a mid-array and an apical electrode, and at two different rates (100 and 300 pps). Average performance on both tasks was worse at the higher pulse rate and did not depend on electrode. However, there was a large variability across and within listeners that did not correlate between the two tasks, suggesting that rate-pitch judgement and regularity detection are to some extent limited by task-specific processes. Simulations with filtered pulse trains presented to NH listeners yielded broadly similar results, except that, for the rate discrimination task, the difference between performance with 100- and 300-pps base rates was smaller than observed for CI users.

Further examination of complex pitch perception in the absence of a place--rate match

Oxenham et al. [Proc. Nat. Acad. Sci. 101, 1421-1425 (2004)] reported that listeners cannot derive a "missing fundamental" from three transposed tones having high carrier frequencies and harmonically related low-frequency modulators. This finding was attributed to complex pitch perception requiring correct tonotopic representation but could have been due to the very high modulator rate difference limens (DLs) observed for individual transposed tones. Experiments 1 and 2 showed that much lower DLs could be obtained for bandpass-filtered pulse trains than for transposed tones with repetition rates of 100 or 300 pps; however, DLs were still larger than for low-frequency pure tones. Experiment 3 presented three pulse trains filtered between 1375 and 1875, 3900 and 5400, and 7800 and 10 800 Hz simultaneously with a pink-noise background. Listeners could not compare the "missing fundamental" of a stimulus in which the pulse rates were, respectively, 150, 225, and 300 pps, to one where all pulse trains had a rate of 75 pps, even though they could compare a 150 + 225 + 300 Hz complex tone to a 75-Hz pure tone. Hence although filtered pulse trains can produce fairly good pitch perception of simple stimuli having low repetition rates and high-frequency spectral content, no evidence that such stimuli enable complex pitch perception in the absence of a place-rate match was found.

Frequency selectivity of contralateral residual acoustic hearing in bimodal cochlear implant users, and limitations on the ability to match the pitch of electric and acoustic stimuli

Objective: To assess the reliability of across-ear, acoustic-electric pitch/timbre comparisons for determining effective characteristic frequencies of cochlear implant electrodes. Study sample: Nine CI users with contralateral residual acoustic hearing. Design: Absolute acoustic thresholds in the unimplanted ear were measured and frequency selectivity was assessed via psychophysical tuning curves. An adjustment method was used to match the percepts elicited by pulse trains on individual electrodes with various acoustic signals (pure tones, narrow-band noises, and bandpass filtered pulse trains). The starting frequency of the acoustic signal was roved and matches were obtained at different loudness levels. Results: Acoustic frequency selectivity varied widely. Two subjects showed clear evidence of frequency selectivity extending above 500 Hz. Only these subjects produced consistent pitch matches over repeated measurements. For other subjects, the acoustic frequency eventually selected tended to correlate with the initially presented frequency. There was limited evidence of level effects and these were inconsistent across subjects and electrodes. Conclusions: Across-modality pitch/timbre matching appears unlikely to provide a generally applicable method for determining the effective characteristic frequencies of cochlear implant electrodes. Frequency selectivity above 500 Hz may be necessary for consistent pitch/timbre matches.

Limitations on rate discrimination.

We investigated the limits of temporal pitch processing under conditions where the place and rate of stimulation on the basilar membrane were independent. Stimuli were harmonic complexes passed through a fixed bandpass filter and resembled filtered pulse trains. The task was to detect a difference in F0. When the harmonics were filtered between 3900-5400 Hz, presented monaurally, and summed in sine phase, subjects could perform the task at all FOs studied. However, when the pulse rate was doubled by summing components in alternating phase, thresholds increased with increasing F0 until the task was impossible at F0 = 300 Hz (pulse rate=600 pps). Thresholds improved again at higher FOs, presumably because some harmonics became resolved. The F0 at which this breakdown occurred decreased when the complexes were filtered into a lower frequency region, and increased when they were filtered into a higher region. In the highest region tested (7800-10800 Hz), all listeners could detect an increase of less than about 20% re: a pulse rate of 600 pps for alternating-phase complexes. Presenting a copy of the standard (lower-F0) stimulus to the contralateral ear during all intervals of a forced-choice trial improved performance markedly under conditions where monaural rate discrimination was very poor. This showed that temporal information is present in the auditory nerve that is unavailable to the temporal pitch mechanism, but which is accessible when a binaural cue is available. The results are compared to the inability of most cochlear implantees to detect increases in the rate of electrical pulse trains above about 300 pps. It is concluded that this inability is unlikely to result entirely from a central pitch limitation, because, with analogous acoustic stimulation, normal listeners can perform the task at substantially higher rates.

Temporal pitch mechanisms in acoustic and electric hearing.

Two experiments investigated pitch perception for stimuli where the place of excitation was held constant. Experiment 1 used pulse trains in which the interpulse interval alternated between 4 and 6 ms. In experiment 1a these "4-6" pulse trains were bandpass filtered between 3900 and 5300 Hz and presented acoustically against a noise background to normal listeners. The rate of an isochronous pulse train (in which all the interpulse intervals were equal) was adjusted so that its pitch matched that of the "4-6" stimulus. The pitch matches were distributed unimodally, had a mean of 5.7 ms, and never corresponded to either 4 or to 10 ms (the period of the stimulus). In experiment 1b the pulse trains were presented both acoustically to normal listeners and electrically to users of the LAURA cochlear implant, via a single channel of their device. A forced-choice procedure was used to measure psychometric functions, in which subjects judged whether the 4-6 stimulus was higher or lower in pitch than isochronous pulse trains having periods of 3, 4, 5, 6, or 7 ms. For both groups of listeners, the point of subjective equality corresponded to a period of 5.6 to 5.7 ms. Experiment 1c confirmed that these psychometric functions were monotonic over the range 4-12 ms. In experiment 2, normal listeners adjusted the rate of an isochronous filtered pulse train to match the pitch of mixtures of pulse trains having rates of F1 and F2 Hz, passed through the same bandpass filter (3900-5400 Hz). The ratio F2/F1 was 1.29 and F1 was either 70, 92, 109, or 124 Hz. Matches were always close to F2 Hz. It is concluded that the results of both experiments are inconsistent with models of pitch perception which rely on higher-order intervals. Together with those of other published data on purely temporal pitch perception, the data are consistent with a model in which only first-order interpulse intervals contribute to pitch, and in which, over the range 0-12 ms, longer intervals receive higher weights than short intervals.

Study of effect of low-pass filtering for channels amplitude estimation on speech intelligibility by acoustic simulation of SPEAK strategy

The conventional way of amplitude envelope estimation for stimulation pulse train generation for selected channels within multi-channel cochlear implant speech processors is low-pass filtering of half or full wave rectified (HWR or FWR) frequency bands of the corresponding channels. To study the necessity of low-pass filtering (LPF) for amplitude envelope generation, acoustic simulation of the SPEAK speech strategy was implemented. Stimulation pulse patterns for selected channels are estimated without LPF and compared to those estimated by the conventional way. Acoustic stimuli are regenerated using an overlap adding technique of filtered pulse trains. Speech intelligibility tests on CV and VCV of Japanese within normal-hearing subjects were conducted to measure the effect of LPF. A one-way ANOVA indicates that the LPF with a cutoff frequency 160 Hz for amplitude envelope estimation of selected channels negatively affected the speech intelligibility.

Pitch of envelope-modulated rippled noise

Both amplitude modulation (AM) and the delay-and-add process of rippled-noise (RN) generation can impart a pitch to wideband Gaussian noise. For AM, and possibly also RN, the basis of the pitch percept derives from temporal processing of the stimulus. In the present study, the pitch of envelope-modulated rippled noise was matched to filtered pulse trains. RN was generated with a delay of either 5 or 10 ms with either one or eight iterations and a gain of either plus or minus one. AM rates ranged from 50 to 200 Hz. Without rippling the carrier, accuracy in pitch matching to AM rate was fairly consistent over the range of rates studied with most deviations occurring as ‘‘octave errors.’’ With envelope modulation of RN carriers, pitch matches were distributed between pitches associated with the AM rate and fine-structure delay, in some cases with a noticeable increase in ‘‘octave errors.’’ Even with eight iterations in RN generation, an influence of AM rate on pitch matches was obtained. Pitch-match distributions can be roughly accounted for by consideration of the largest positive peak of both the fine-structure and envelope autocorrelation functions. [Work supported by NIH and AFOSR.]

The effect of mean rate cues on the pitch of filtered pulse trains.

Listeners judged which of two bandpass filtered (3900–5400 Hz) pulse trains had the higher pitch. In the control condition, where the two trains differed only in repetition rate, two cues might produce a pitch difference. The interval between any two pulses is a multiple of a common period, which is shorter for the train with the faster rate (the ‘‘F0 cue’’). This stimulus also contains more pulses (the ‘‘mean rate cue’’). Generally, temporal pitch models focus on the F0 cue. The mean rate cue was investigated by, in one condition, deleting a proportion (1−P) of the pulses in each train, and manipulating P so that the mean rate (P⋅F0) was equal in the two halves of each trial [R. A. Dobie and N. Dillier, Hear. Res. 18, 41–45 (1985)]. Listeners identified the stimulus having the higher (original) F0 as possessing the higher pitch less reliably than in the control condition, and less reliably than in a condition where the mean rate difference was exaggerated. The conclusion that the mean rate cue can influence pitch, was confirmed a second experiment, where listeners adjusted the F0s of two pulse trains with different P’s to have equal pitches.

Masking between Spatially Separated Sounds

The development of optimal three-dimensional auditory displays requires a more complete understanding of the interactions among spatially separated sounds. Free-field masking was investigated as a function of the spatial separation between signal and masker sounds within the horizontal, frontal, and median planes. The detectability of filtered pulse trains in the presence of noise maskers was measured using a cued, two-alternative, forced-choice, adaptive staircase procedure. Signal and masker combinations in low (below 2.3 kHz), middle (1.0–8.5 kHz), and high (above 3.5 kHz) frequency regions were examined. As the sound sources were separated within the horizontal plane, signal detectability increased dramatically. Similar improvement in detectability was observed within the frontal plane. As suggested by traditional binaural models, interaural time cues and interaural intensity cues are likely to play a major role in mediating masking release in both the horizontal and frontal planes. Because no interaural cues exist for stimuli presented within the median plane, traditional models would not predict a release from masking when the stimuli are separated within this plane. However, with high frequency signals, masking release similar to that observed in the horizontal and frontal planes could be observed in the median plane. The current literature suggests that sound localization in the median plane may depend on direction-specific spectral cues that are introduced by the pinna at high frequencies. The masking release observed here may also depend on these “pinna cues.”

The perceptual segregation of simultaneous auditory signals: pulse train segregation and vowel segregation.

In the experiments reported here, we attempted to find out more about how the auditory system is able to separate two simultaneous harmonic sounds. Previous research (Halikia & Bregman, 1984a, 1984b; Scheffers, 1983a) had indicated that a difference in fundamental frequency (F0) between two simultaneous vowel sounds improves their separate identification. In the present experiments, we looked at the effect of F0s that changed as a function of time. In Experiment 1, pairs of unfiltered or filtered pulse trains were used. Some were steady-state, and others had gliding F0s; different F0 separations were also used. The subjects had to indicate whether they had heard one or two sounds. The results showed that increased F0 differences and gliding F0s facilitated the perceptual separation of simultaneous sounds. In Experiments 2 and 3, simultaneous synthesized vowels were used on frequency contours that were steady-state, gliding in parallel (parallel glides), or gliding in opposite directions (crossing glides). The results showed that crossing glides led to significantly better vowel identification than did steady-state F0s. Also, in certain cases, crossing glides were more effective than parallel glides. The superior effect of the crossing glides could be due to the common frequency modulation of the harmonics within each component of the vowel pair and the consequent decorrelation of the harmonics between the two simultaneous vowels.

Discrimination of permuted orders of tones having the same pitch but different qualities

Untrained subjects judged whether two sequences of tones were different or the same. Each sequence consisted of three tonal items (A = filtered pulse train, B = sinusoidal tone, C = filtered square wave) which were arranged either in the form [ABCABC⋯A] followed by [ACBACB⋯A] (different) or [ABCABC⋯A] followed by [ABCABC⋯A] (same). Each tonal item had the same waveform repetition frequency (800 Hz) and pitch, but possessed a distinctive quality. Duration of items was the same within sequence pairs, and was varied systematically from item durations of several seconds (permitting easy verbal labeling of components in the proper order) down to 10 ms (components could not be recognized). Accuracy of judgments was well above chance at each item duration. It is suggested that discrimination of differences in the order of items was accomplished through a naming of individual components in order of occurrence for longer item durations, and through recognition of qualitative differences between sequences for brief...

Rate discrimination of high‐pass filtered pulse trains

Rate discrimination of high‐pass filtered pulse trains

Adaptation effects in binaural experiments

Experiments were conducted to examine some of the effects of imbalanced interaural parameters on binaural lateralization. We examined adaptation both to stimuli with interaural amplitude differences and to stimuli with interaural time differences. To measure the adaptation over time, 30‐trial runs were made with each run lasting about 15 min. On each trial, the subject was required to adjust the interaural time delay from an initial random value to a value that resulted in a centered image. For a low‐pass filtered pulse train with an interaural amplitude difference of 10 dB, the amount of interaural time delay needed to center the image changed from 250 μs at the outset to 100 μsec at the end of the run. These results are consistent with earlier results using a paradigm nearly identical to ours [M. P. Bertrand, Proc. R. Soc. Med. 65, 3–4 (Sept. 1972)]. For the same low‐pass filtered pulse train with no interaural amplitude differences, but with the initial random delay restricted to one ear, the amount of time delay needed to center the image changed by as much as 50 μs during the course of a run, i.e., at the end of a run the subjects no longer heard a diotic presentation (zero delay) in the center of their heads. Two possible mechanisms of adaptation will be discussed. [Work supported by NIH.]

The Case of the “Missing MLD”

The decrease in the size of the MLD with signal frequency is presumably caused by the inability of the auditory fibers to follow details of the waveform at high frequencies. The ear appreciates slow frequency envelope variations even at high carrier frequency, however, as the residue attests. Using amplitude modulation, we attempted to obtain an MLD by manipulating the phase of a 250-Hz envelope on a 2000-Hz carrier. Essentially no MLD was observed. Attempting to create a more salient low-frequency pitch sensation we used a high-pass filtered pulse train. Delaying the pulse train, 250 ips, by half a period in the opposite ear had essentially no effect on the detectability of the binaural stimulus in wide-band masking noise. Our procedure is different from that used by Small who showed that simply inverting the waveforms at the two ears produced no MLD. These results replicate some of a more extensive series of measurements carried out by Dr. G. Bruce Henning. [Research supported by the National Institutes of Health, Public Health Service, U. S. Department of Health, Education, and Welfare, and by a research grant awarded to the Center for Human Information Processing.

Noise masking of a change in residue pitch.

A comparison was made of the effectiveness of noise in masking (1) a change in the frequency of a sinusoid, and (2) a change in the repetition rate of a pulse train that had no low-frequency energy. The sinusoid and the pulse train had the same pitch. It was found that, whereas the change in the pitch of the sinusoid was better masked by low-frequency noise, the change in the residue pitch produced by the filtered pulse train was more easily masked by high-frequency noise. These results replicate the findings of Small and Campbell (1961). We conclude as they did that the basis for the residue pitch of the pulse train is its high-frequency components. In a second experiment, a pair of sinusoids was used to produce the residue pitch. The same masking technique was employed, and again the results indicate that the residue pitch is not based on a nonlinear distortion component, but rather is based on information being transmitted through high-frequency components (2100 Hz).

Correction to 'On the Approach of a Filtered Pulse Train to a Stationary Gaussian Process'

Correction to 'On the Approach of a Filtered Pulse Train to a Stationary Gaussian Process'

On the approach of a filtered pulse train to a stationary Gaussian process

A narrow-band process is conveniently characterized in terms of a complex envelope whose magnitude is the envelope, and whose angle is the phase variation of the actual narrow-band process. When the narrow-band process is normally distributed, the complex envelope has the properties of a complex normally distributed process. This paper investigates the approach to the complex normally distributed form of the complex envelope of the output of a narrow-band filter when the input is wide-band non-Gaussian noise of a certain class, and the bandwidth of the narrow-band filter approaches zero. The non-Gaussian input consists of a train of pulses having identical waveshapes, but random amplitudes and phases. While the derivations assume statistical independence between pulses, it is shown that the results are valid for a certain interesting class of dependent pulses. The Central Limit Theorem is proved in the multidimensional case for the output process.