Gary L. Gibian

Clinical trials of a hybrid adaptive beamformer (HAB) for improved speech understanding in noise

A wearable, hybrid adaptive beamformer (HAB) device has been developed using a four‐microphone array and a combination of adaptive and fixed‐weight beamforming. The HAB is being evaluated with hard‐of‐hearing (HoH) listeners in a variety of environments. Pilot tests used HINT test sentences and noise from separate loudspeakers (106‐deg angle) in an audiological booth (AI‐weighted direct‐to‐reverberant ratio 9.1 dB). Measured sentence reception thresholds (SRTs) indicated that the HAB provided 14.5, 16.1, and 12.8 dB improvement over the single microphone for two normal‐hearing and one HoH listener. Results will be reported for ongoing objective and subjective testing of elderly listeners with mild to moderate hearing losses. Subjects are fitted monaurally with a commercially available behind‐the‐ear hearing aid with a dual microphone array following the NAL‐R fitting algorithm. Comparisons are made between listeners’ SRTs in noise, using: (a) conventional single‐microphone, (b) dual microphone, and (c) fo...

Noise reduction using a multimicrophone array for automatic speech recognition on a handheld computer

A four‐microphone array and signal‐processing card have been integrated with a handheld computer such that the integrated device can be carried in and operated with one hand. Automatic speech recognition (ASR) was added to the USAMRMC/TATRCs Battlefield Medical Information System (BMIST) software using an approach that does not require modifying the original code, to produce a Speech‐Capable Personal Digital Assistant (SCPDA). Noise reduction was added to allow operation in noisier environments, using the previously reported Hybrid Adaptive Beamformer (HAB) algorithm. Tests demonstrated benefits of the array over the HP/COMPAQ‐iPAQ built‐in shielded microphone for noise reduction and automatic speech recognition. In electroacoustic and human testing including voice control and voice annotation, the array provided substantial benefit over the built‐in microphone. The benefit varied from about 5 dB (worst‐case scenario, diffuse noise) to about 20 dB (best‐case scenario, directional noise). Future work is expected to produce more rugged SCPDA prototypes for user evaluations, revise the design based on user feedback and real‐world testing, and possibly to allow hands‐free use by using ASR to replace the push‐to‐talk switch, providing feedback aurally and/or via a head‐up display. [Work supported by the U.S. Army Medical Research and Materiel Command (USAMRMC), Contract No. DAMD17‐02‐C‐0112.]

Directional hearing aid using hybrid adaptive beamformer (HAB) and binaural ITE array

A directional hearing aid algorithm called the Hybrid Adaptive Beamformer (HAB), developed for NIH/NIA, can be applied to many different microphone array configurations. In this project the HAB algorithm was applied to a new array employing in‐the‐ear microphones at each ear (HAB‐ITE), to see if previous HAB performance could be achieved with a more cosmetically acceptable package. With diotic output, the average benefit in threshold SNR was 10.9 dB for three HoH and 11.7 dB for five normal‐hearing subjects. These results are slightly better than previous results of equivalent tests with a 3‐in. array. With an innovative binaural fitting, a small benefit beyond that provided by diotic adaptive beamforming was observed: 12.5 dB for HoH and 13.3 dB for normal‐hearing subjects, a 1.6 dB improvement over the diotic presentation. Subjectively, the binaural fitting preserved binaural hearing abilities, giving the user a sense of space, and providing left–right localization. Thus the goal of creating an adaptive...

Physical correlates of perceptual similarity among synthesized approximations to selected targets

Experiments are described to evaluate the relative merits of several measures for predicting the perceived closeness of synthesized approximations to selected steady‐state portions of musical tones (“targets”). Approximations were synthesized by means of audio‐rate frequency modulation [Chowning, J. Audio Eng. Soc. 21, 526–534 (1973)] using a computer program developed in a previous paper [Gibian et al., J. Acoust. Soc. Am. Suppl. 1 81, S46 (1987), and Audio Engineering Society Pre‐print ♯2380]. The importance of including level‐ and frequency‐dependent error weightings according to the Fletcher‐Munson curves will be assessed. This information can be useful to composers who are blending electronic and traditional instruments in mixed ensembles.

Progress toward more intuitive programming of FM synthesizers

As was reported last year, work has been done to make the programming of new tone colors on FM synthesizers more intuitive for composers and musicians. The logarithmic relations between modulator, carrier, and envelope level parameters on commercially available instruments and corresponding index of modulation and output amplitudes in standard FM synthesis equations have been empirically determined. The relations have been incorporated into a tutorial computer program that displays the waveform and corresponding spectrum after synthesis parameters are entered in either form. With the program, users can gain familiarity with the spectra of FM‐synthesized sounds, which can be quite complicated for synthesizers using four or six oscillators in various configurations. Work is currently being done to calibrate feedback and envelope rate parameters and to develop software to go in the opposite direction, determining FM parameters to approximate spectra with complicated and time‐varying shapes. Ultimately, such ...

Chasing the lost chord with an Apple computer

The twentieth century has seen the rapid growth of electronic synthesis and reproduction of music, to the point where “a major part of recent popular and commercial music would not exist without synthesizers and…the recording studio…has itself become a musical instrument…” [Gordon Mumma, in Electronic Music (Allen Strange, 1983)]. The Audio Technology Program at The American University includes a course in Digital Music Synthesis based in part on the Soundchaser system manufactured by Passport Designs for the Apple computer equipped with Mountain digital circuit cards. Measurements of performance parameters (e.g., the discrete nature of the attack, decay, and release) will be discussed and musical excerpts will be presented.

Automobile interior noise: psychoacoustic difference thresholds in octave‐band sound pressure levels

The present study investigates human response to spectral changes in automobile interior noise, which characteristically has strong low‐frequency content and much less high‐frequency content. Specifically, we determined the smallest increases and decreases in sound pressure level in each octave band of automobile interior noise that can be detected by typical vehicle occupants. Increments and decrements in 1‐dB steps in each octave band with center frequencies from 63 Hz to 4 kHz were presented to a group of automobile users, using a two‐alternative forced‐choice procedure. Results indicate that smaller changes (2–3 dB) can be detected in the higher‐frequency octave bands (2 and 4 kHz), while larger changes (5–7 dB) are required for detection in the lower‐frequency octave bands (63 and 125 Hz). The difference thresholds do not appear to depend on overall sound pressure level in the range 60 to 75 dB A.