Engineering topics
Publications and source records attributed to Begault, Durand R..
Auditory thresholds for 10 subjects were obtained for speech stimuli reverberation. The reverberation was produced and manipulated by 3-D audio modeling based on an actual room. The independent variables were octave-band-filtering (bypassed, 0.25 - 2.0 kHz Fc) and reverberation time (0.2- 1.1 sec). An ANOVA revealed significant effects (threshold range: -19 to -35 dB re 60 dB SRL).
This viewgraph presentation provides information on a study the goal of which was to determine the auditory threshold for speech reverberation using a specific room model. Procedures and hardware used are detailed as are the participants in the study.
In recent work, a microphone array consisting of an omnidirectional microphone and colocated dipole microphones having orthogonally aligned dipole axes was used to examine the directional nature of a room impulse response. The arrival of significant reflections was indicated by peaks in the power of the omnidirectional microphone response; reflection direction of arrival was revealed by comparing zero-lag crosscorrelations between the omnidirectional response and the dipole responses to the omnidirectional response power to estimate arrival direction cosines with respect to the dipole axes.
The challenges facing successful multimedia presentation depend largely on the expectations of the designer and end user for a given application. Perceptual limitations in distance, elevation and azimuth sound source simulation differ significantly between headphone and cross-talk cancellation loudspeaker listening and therefore must be considered. Simulation of an environmental context is desirable but the quality depends on processing resources and lack of interaction with the host acoustical environment. While techniques such as data reduction of head-related transfer functions have been used widely to improve simulation fidelity, another approach involves determining thresholds for environmental acoustic events. Psychoacoustic studies relevant to this approach are reviewed in consideration of multimedia applications
Current virtual acoustic displays for teleconferencing and virtual reality are usually limited to very simple or non-existent renderings of reverberation, a fundamental part of the acoustic environmental context that is encountered in day-to-day hearing. Several research efforts have produced results that suggest that environmental cues dramatically improve perceptual performance within virtual acoustic displays, and that is possible to manipulate signal processing parameters to effectively reproduce important aspects of virtual acoustic perception in real-time. However, the computational resources for rendering reverberation remain formidable. Our efforts at NASA Ames have been focused using a several perceptual threshold metrics, to determine how various "trade-offs" might be made in real-time acoustic rendering. This includes both original work and confirmation of existing data that was obtained in real rather than virtual environments. The talk will consider the importance of using individualized versus generalized pinnae cues (the "Head-Related Transfer Function"); the use of head movement cues; threshold data for early reflections and late reverberation; and consideration of the necessary accuracy for measuring and rendering octave-band absorption characteristics of various wall surfaces. In addition, a consideration of the analysis-synthesis of the reverberation within "everyday spaces" (offices, conference rooms) will be contrasted to the commonly used paradigm of concert hall spaces.
This talk will overview the basic technologies related to the creation of virtual acoustic images, and the potential of including spatial auditory displays in human-machine interfaces. Research into the perceptual error inherent in both natural and virtual spatial hearing is reviewed, since the formation of improved technologies is tied to psychoacoustic research. This includes a discussion of Head Related Transfer Function (HRTF) measurement techniques (the HRTF provides important perceptual cues within a virtual acoustic display). Many commercial applications of virtual acoustics have so far focused on games and entertainment ; in this review, other types of applications are examined, including aeronautic safety, voice communications, virtual reality, and room acoustic simulation. In particular, the notion that realistic simulation is optimized within a virtual acoustic display when head motion and reverberation cues are included within a perceptual model.
The use of a soundfield microphone for examining the directional nature of a room impulse response was reported recently. By cross-correlating monopole and co-located dipole microphone signals aligned with left-right, up-down, and front-back axes, a sense of signal direction of arrival is revealed. The current study is concerned with the array's ability to detect individual reflections and directions of arrival, as a function of the cross-correlation window duration. If is window is too long, weak reflections are overlooked; if too short, spurious detections result. Guidelines are presented for setting the window width according to perceptual criteria. Formulas are presented describing the accuracy with which direction of arrival can be estimated as a function of room specifics and measurement noise. The direction of arrival of early reflections is more accurately determined than that of later reflections which are quieter and more numerous. The transition from a fairly directional sound field at the beginning of the room impulse response to a uni-directional diffuse field is examined. Finally, it is shown that measurements from additional dipole orientations can significantly improve the ability to detect reflections and estimate their directions of arrival.
These are 3 chapters that will appear in a book titled "Building Acoustical Design", edited by Charles Salter. They are designed to introduce the reader to fundamental concepts of acoustics, particularly as they relate to the built environment. "Fundamentals of Acoustics" reviews basic concepts of sound waveform frequency, pressure, and phase. "Psychoacoustics and Hearing" discusses the human interpretation sound pressure as loudness, particularly as a function of frequency. "Acoustic Measurements" gives a simple overview of the time and frequency weightings for sound pressure measurements that are used in acoustical work.
An optimal approach to auditory display design for commercial aircraft would utilize both spatialized ("3-D") audio techniques and active noise cancellation for safer operations. Results from several aircraft simulator studies conducted at NASA Ames Research Center are reviewed, including Traffic alert and Collision Avoidance System (TCAS) warnings, spoken orientation "beacons" for gate identification and collision avoidance on the ground, and hardware for improved speech intelligibility. The implications of hearing loss amongst pilots is also considered.
Our work in virtual reality systems at NASA Ames Research Center includes the area of aurally-guided visual search, using specially-designed audio cues and spatial audio processing (also known as virtual or "3-D audio") techniques (Begault, 1994). Previous studies at Ames had revealed that use of 3-D audio for Traffic Collision Avoidance System (TCAS) advisories significantly reduced head-down time, compared to a head-down map display (0.5 sec advantage) or no display at all (2.2 sec advantage) (Begault, 1993, 1995; Begault & Pittman, 1994; see Wenzel, 1994, for an audio demo). Since the crew must keep their head up and looking out the window as much as possible when taxiing under low-visibility conditions, and the potential for "blunder" is increased under such conditions, it was sensible to evaluate the audio spatial cueing for a prototype audio ground collision avoidance warning (GCAW) system, and a 3-D audio guidance system. Results were favorable for GCAW, but not for the audio guidance system.
Synthetic head related transfer functions (HRTF's) for imposing reprogramable spatial cues to a plurality of audio input signals included, for example, in multiple narrow-band audio communications signals received simultaneously are generated and stored in interchangeable programmable read only memories (PROM's) which store both head related transfer function impulse response data and source positional information for a plurality of desired virtual source locations. The analog inputs of the audio signals are filtered and converted to digital signals from which synthetic head related transfer functions are generated in the form of linear phase finite impulse response filters. The outputs of the impulse response filters are subsequently reconverted to analog signals, filtered, mixed and fed to a pair of headphones.
A spatial auditory display for multiple speech communications was developed at NASA/Ames Research Center. Input is spatialized by the use of simplifiedhead-related transfer functions, adapted for FIR filtering on Motorola 56001 digital signal processors. Hardware and firmware design implementations are overviewed for the initial prototype developed for NASA-Kennedy Space Center. An adaptive staircase method was used to determine intelligibility levels of four-letter call signs used by launch personnel at NASA against diotic speech babble. Spatial positions at 30 degree azimuth increments were evaluated. The results from eight subjects showed a maximum intelligibility improvement of about 6-7 dB when the signal was spatialized to 60 or 90 degree azimuth positions.
The advantage of a head up auditory display was evaluated in an experiment designed to measure and compare the acquisition time for capturing visual targets under two conditions: Standard head down traffic collision avoidance system (TCAS) display, and three-dimensional (3-D) audio TCAS presentation. Ten commercial airline crews were tested under full mission simulation conditions at the NASA Ames Crew-Vehicle Systems Research Facility Advanced Concepts Flight Simulator. Scenario software generated targets corresponding to aircraft which activated a 3-D aural advisory or a TCAS advisory. Results showed a significant difference in target acquisition time between the two conditions, favoring the 3-D audio TCAS condition by 500 ms.
A spatial auditory display was designed for separating the multiple communication channels usually heard over one ear to different virtual auditory positions. The single 19 foot rack mount device utilizes digital filtering algorithms to separate up to four communication channels. The filters use four different binaural transfer functions, synthesized from actual outer ear measurements, to impose localization cues on the incoming sound. Hardware design features include 'fail-safe' operation in the case of power loss, and microphone/headset interfaces to the mobile launch communication system in use at KSC. An experiment designed to verify the intelligibility advantage of the display used 130 different call signs taken from the communications protocol used at NASA KSC. A 6 to 7 dB intelligibility advantage was found when multiple channels were spatially displayed, compared to monaural listening. The findings suggest that the use of a spatial auditory display could enhance both occupational and operational safety and efficiency of NASA operations.
Speech intelligibility was evaluated using a virtual acoustic ("3-D audio") display using the method specified by ANSI. Ten subjects were evaluated with stimuli either unfiltered or low-pass filtered at 4 kHz. Results show virtual acoustic techniques are advantageous for both full-bandwidth (44.1 kHz srate) and low (8 kHz srate) bandwidth "telephone-grade" teleconferencing systems.
Synthetic head related transfer functions (HRTF's) for imposing reprogrammable spatial cues to a plurality of audio input signals included, for example, in multiple narrow-band audio communications signals received simultaneously are generated and stored in interchangeable programmable read only memories (PROM's) which store both head related transfer function impulse response data and source positional information for a plurality of desired virtual source locations. The analog inputs of the audio signals are filtered and converted to digital signals from which synthetic head related transfer functions are generated in the form of linear phase finite impulse response filters. The outputs of the impulse response filters are subsequently reconverted to analog signals, filtered, mixed, and fed to a pair of headphones.
Three-dimensional acoustic display systems have recently been developed that synthesize virtual sound sources over headphones based on filtering by head-related transfer functions (HRTFs), the direction-dependent spectral changes caused primarily by the pinnae. In this study, 11 inexperienced subjects judged the apparent spatial location of headphone-presented speech stimuli filtered with nonindividualized HRTFs. About half of the subjects 'pulled' their judgments toward either the median or the lateral-vertical planes, and estimates were almost always elevated. Individual differences were pronounced for the distance judgments; 15 to 46 percent of stimuli were heard inside the head, with the shortest estimates near the median plane. The results suggest that most listeners can obtain useful azimuth information from speech stimuli filtered by nonindividualized HRTFs. Measurements of localization error and reversal rates are comparable with a previous study that used broadband noise stimuli.
A spatial auditory display was used to convolve speech stimuli, consisting of 130 different call signs used in the communications protocol of NASA's John F. Kennedy Space Center, to different virtual auditory positions. An adaptive staircase method was used to determine intelligibility levels of the signal against diotic speech babble, with spatial positions at 30 deg azimuth increments. Non-individualized, minimum-phase approximations of head-related transfer functions were used. The results showed a maximal intelligibility improvement of about 6 dB when the signal was spatialized to 60 deg or 90 deg azimuth positions.