FIG. 1: RECORDING APPARATUS
Helen J. Simon -- Helen@ski.org
Deborah Gilden -- debby@ski.org
John Brabyn -- Brabyn@ski.org
Al Lotze
The Smith-Kettlewell Eye Research Institute
2318 Fillmore St
San Francisco, CA, 94115
Harry Levitt -- harrylevitt@earthlink.net
Advanced Hearing Concepts
Bodega Bay, CA 94923
Popular Version of Paper 5aPP18
Presented Friday morning, May 18, 2012
163rd ASA Meeting, Hong Kong
INTRODUCTION
Blind pedestrians use their hearing to travel safely, independently and efficiently (a skill known as "orientation and mobility" [O&M] or "wayfinding"). Some of the auditory cues they use are clear and obvious, but others are subtle environmental sound cues.
These auditory cues may be
- Sounds generated by an object (e.g. car with engine running)
- Sounds reflected off of an object (e.g. environmental sounds in a corridor reflecting off walls and doors); these provide "echolocation"
- Sounds that are reduced in loudness due to blocking by an object (e.g. from a car); these are "sound shadows"
Hearing environmental sounds and noting sound shadows help blind pedestrians avoid obstacles, locate landmarks, make safe and straight street crossings, maintain a straight line of travel, remain oriented, and walk down corridors smoothly, etc. This study is an attempt to learn details about the acoustic nature of the cues received by the two ears that might provide useful environmental signals to blind people as they navigate. Good blind travelers often can detect if doorways along a corridor are open or closed, and we are interested in understanding the auditory cues that allow them to do this. To this end we have measured the characteristics of the sounds in the ears of blind pedestrians when traveling down a corridors with open and closed doors.
METHODS
With the equipment shown in Figure 1, we recorded the sounds reaching the ears of eight blind subjects while they walked along indoor corridors in an office building, with open and closed doors. These recordings were made through miniature microphones placed in the subjects' ear canals in order to capture the sounds that actually were available to them during the short walks. The subjects also wore a video camera with a 3600field-of-view, which was mounted on their heads. The recorded video allowed us to tell the exact surroundings and location of the subject for any point in the sound recordings. In other words, it synced the location in space with the sounds coming into the ears. Head and body movements were also recorded with a gyroscope and accelerometer also mounted on the head, and an additional gyroscope, which was worn on the torso.
|
FIG. 1: RECORDING APPARATUS |
RESULTS
Our results reported here show the acoustic information that we have recorded with subjects walking down a hallway past an open or closed doorway.
Figure 2 below shows the differences in the sound encountered in the hallway in the two scenarios above. The differences between the door-open and door-closed conditions are small (a maximum of 5.5 dB at 860Hz) with the bulk of the differences in the low frequencies (below 1500 Hz).
The low frequency dip in the ambient sound spectrum near a wall is presumably due to reflections from the passage walls. (Sound reflections are similar to mirror reflections. Mirror reflections increase the apparent brightness of a room; sound reflections can increase the apparent loudness of a space as well as change the acoustical characteristics of the sound.)
 |
FIG. 2: Recorded Sound (Intensity vs. Frequency) Adjacent to a Wall and an Open Door |
Figure 3 below shows the cross correlation of the environmental sound between the two ears for the same two conditions. (A cross-correlation is a measure of similarity of two waveforms as one of them is delayed in time by varying amounts).
 |
Fig 3: Cross Correlation of sound in left and right ear near a closed and open door |
Both cross correlation functions show a peak at approximately 600 microseconds (5sec), the amount of time it takes for a sound located to one side of the head to travel from one ear to the other. There is a higher (better) correlation for the closed-door condition, but a much broader peak (a wide range of time delays give a nearly equal match). This result indicates sound reflects from the closed door giving partial correlations over a range of time delays. The open-door correlation function has a sharper peak, but less overall correlation relative to the closed-door condition. This indicates that the correlation is best over a narrower range of time delays. This result is consistent with a lack of reflections from the closed door and/or sounds coming out of the adjacent room.
CONCLUSIONS
These data shown demonstrate the usefulness of cross correlation between the ears in the analysis of wayfinding cues; while the intensity vs. frequency differences are small (Fig 1) the differences between the cross correlation functions are quite large (Fig 2). Although the ear is very sensitive to interaural differences, previous research did not measure these differences. The results suggest that blind pedestrians use these differences in the sound reaching the two ears to help with foot travel. The differences in intensity vs. frequency at low frequencies confirm previous work regarding ambient room noise near a wall and an opening (Ashmead and Wall, 1999).
Implications of these findings for a visually impaired person with a hearing loss are that (1) hearing loss can be expected to interfere with safe, efficient, and effective foot travel and (2) special purpose hearing aids with low frequency amplification that do not distort time differences between the ears may be needed to preserve the auditory cues used by good travelers. Further study from this laboratory will test the importance of these cues with normally sighted and blind subjects.
REFERENCES
Ashmead DH, Wall RS. Auditory perception of walls via spectral variations in the ambient sound field. Journal of rehabilitation research and development. 1999;36(4):313-22.