Tactual Frequency and Amplitude Discrimination with Fixed and Roving Background

Abstract

The authors present frequency and amplitude discrimination thresholds for motional and vibrational stimuli presented with and without the presence of roving background signals. Participants received on their left index fingertip sinusoidal displacement waveforms over a range of 2-200 Hz in frequency and 20-35 dB sensation level in amplitude. When the target stimulus was presented in isolation, the average Weber fraction for frequency was 0.18−0.25, and the average amplitude discrimination threshold was 2.2−2.5 dB. When roving background signals were presented with the targets and the participants were instructed to ignore the interfering backgrounds, the discrimination thresholds for both frequency and amplitude increased. The amount of increase depended on the nature of the background signals. The results are discussed in terms of their implications for tactual displays of speech in communication aids for the deaf.

Full text

Page 1

Tactual Frequency and Amplitude Discrimination
with Fixed and Roving Background

Ali Israr1, Hong Z. Tan1, and Charlotte M. Reed2
1Haptic Interface Research Laboratory, Purdue University, USA
2Research Laboratory of Electronics, MIT, USA
{israr, hongtan}@purdue.edu; cmreed@mit.edu


Abstract
present The authors frequency and amplitude
discrimination thresholds for motional and vibrational
stimuli presented with and without the presence of roving
background signals. Participants received on their left
index fingertip sinusoidal displacement waveforms over a
range of 2-200 Hz in frequency and 20-35 dB sensation
level in amplitude. When the target stimulus was presented
in isolation, the average Weber fraction for frequency was
0.18−0.25, and the average amplitude discrimination
threshold was 2.2−2.5 dB. When roving background
signals were presented with the targets and the participants
were instructed to ignore the interfering backgrounds, the
discrimination thresholds for both frequency and amplitude
increased. The amount of increase depended on the nature
of the background signals. The results are discussed in
terms of their implications for tactual displays of speech in
communication aids for the deaf.
1. Introduction
This work was motivated by our desire to use the sense
of touch as a substitute speech communication channel for
individuals with profound hearing impairment. A multi-
finger tactual stimulator capable of delivering displacement
waveforms from low-frequency large-amplitude kinesthetic
movements to high-frequency small-amplitude cutaneous
vibrations has been developed [1]. The device is capable of
transmitting information at a relatively high rate of 12
bits/sec [2], roughly the same rate that has been
demonstrated by dead-blind individuals who use a manual
method called Tadoma to “feel” speech [3]. The present
study was concerned with documenting tactual resolution
in the frequency-amplitude space for transmitting broad-
band modified speech signals through such a device.
Although previous studies have documented tactile
frequency and amplitude resolution (cf. [4]), few have
addressed the effects of background maskers on resolution
ability (although see [5]). The present set of experiments
provided a systematic measurement of frequency and
amplitude discrimination thresholds
displacement waveforms delivered to a fingertip with or
without the presence of other masking background signals.
for sinusoidal
2. Experiment Methods
The experimental apparatus was a three-channel
multidimensional tactual display called the Tactuator [1]
with a new controller that takes into account human
detection thresholds from dc to 300 Hz [6]. The controller
preserves the relative magnitudes of spectral components in
a broadband input signal in terms of sensation levels (SL;
dB above detection threshold).
Four participants (2 males and 2 females; 21-28 years
old; average age 24) took part in the present study. Six
target frequencies, 2 and 4 Hz in the low-frequency region,
15 and 30 Hz in the mid-frequency region, and 80 and 200
Hz in the high-frequency region (representing frequency
bands corresponding to motional, fluttering/rough and
vibrational sensations [2]), were selected at two amplitude
levels of 20 (A1) and 35 (A2) dB SL. The duration of the
stimulus was always 250 msec. A Hanning window with a
25-ms rise and fall time was applied to all signals. All
stimuli were presented to the left index fingertip.
A three-interval forced-choice paradigm with a one-up
three-down adaptive procedure [7,8] was used to measure
frequency and amplitude discrimination thresholds for each
of the 12 target signals (6 frequencies × 2 amplitudes). The
target stimuli were presented either in isolation (fixed
background: no-masker) or in the presence of one or two
maskers that occurred simultaneously with the target
(roving background: single-masker or double-masker). In
the single-masker condition, the masker was selected from
one of the two frequency regions of which the target was
not a member. In the double-masker condition, two
maskers were employed (one from each of the two
frequency regions of which the target was not a member).
The frequency of the masker was selected randomly from
10 equal steps between the two target frequencies for that
frequency region (i.e., 2−4 Hz, 15−30 Hz, or 80−200 Hz in
the low-, mid- or high-frequency regions, respectively).
The masker amplitude for a target presented at 20 or 35 dB
SL was randomly selected from 10 logarithmically-scaled
levels within 10−20 dB SL or 20−35 dB SL, respectively.
The masker frequency and amplitude were randomly
selected for each of the three intervals on each trial.
383
Symposium on Haptic Interfaces for
Virtual Environment and Teleoperator Systems 2006
March 25 - 26, Alexandria, Virginia, USA
1-4244-0226-3/06/$20.00 ©2006 IEEE

Page 2

Figure 1. Frequency discrimination results in Weber fractions.
3. Results
Frequency discrimination. Thresholds are shown in
terms of Weber fractions as (∆F)0/Fref (Fig. 1; top panel for
A1 = 20 dB SL, bottom panel for A2 = 35 dB SL). The
asterisks indicate statistically significant differences
between the results for roving background masker
conditions relative to the corresponding no-masker
conditions. For the no-masker condition and at A1, Weber
fractions ranged from 0.13 at 200 Hz to 0.38 at 4 Hz with
an average of 0.25. At A2, Weber fractions ranged from
0.14 at 30 Hz to 0.24 at 15 Hz with an average of 0.18. The
Weber fractions increased in the roving background
masker conditions, and more so for the double-masker
condition. At A1, the average values increased to 0.54 and
0.73 for the single- and double-masker conditions,
respectively. At A2, the average fractions increased to 0.36
and 0.66, respectively.
Amplitude discrimination. Thresholds are shown in Fig.
2 as (∆A)0 in dB. For the no-masker condition and at A1 =
20 dB SL, the thresholds ranged from 1.8 dB at 30 Hz to
3.0 dB at 2 Hz with an average of 2.5 dB. At A2 = 35 dB
SL, the thresholds ranged from 1.7 dB at 15 Hz to 2.7 dB
at 200 Hz with an average of 2.2 dB. Detrimental effects of
roving background masking conditions were observed in
the amplitude discrimination experiments as well. At both
A1 and A2, roving effects were again larger in the double-
masker condition than in the single-masker condition. The
average values of (∆A)0 in dB for the single- and double-
masker conditions were 3.46 and 3.96 dB at A1, and 3.49
and 4.19 dB at A2, respectively.



Figure 2. Amplitude discrimination results in (∆ ∆A)0 in dB.
4. Discussion
Our results obtained in the no-masker conditions for
both frequency and amplitude discrimination were similar
to those reported previously [4]. In general, lower
thresholds were found at the higher amplitude A2 than at
A1. Thresholds were higher for the double-masker
compared to the single-masker conditions. There were
differential effects for background maskers, that is, some
maskers are more effective than others with a given target
signal. For example, low-frequency targets were not
significantly affected by high-frequency maskers and vice
versa. The results provide quantitative resolution data for
designing multi-attribute signals for speech communication
through the Tactuator, and suggest the sparing use of more
than two frequency components on a single digit. In view
of the relative independence of low- and high-frequency
signals, they can be used simultaneously to encode separate
or redundant speech features to improve information
transmission.
Acknowledgments
This research was supported by research grant No. R01-
DC00126 from the National Institute on Deafness and
Other Communication Disorders, National Institutes of
Health.

References
[1] H. Z. Tan and W. M. Rabinowitz, “A new multi-finger tactual
display,” Proc. of the International Symposium on Haptic
Interfaces for Virtual Environment and Teleoperator Systems,
515-522, 1996.
384

Page 3

[2] H. Z. Tan, N. I. Durlach, C. M. Reed, and W. M. Rabinowitz,
“Information transmission with a multifinger tactual display,”
Perception and Psychophysics, 61, 993-1008, 1999.
[3] C. M. Reed, N. I. Durlach, and L. D. Braida, “Research on
Tactile Communication of Speech: A Review,” ASHA
Monographs No. 20, 1982.
[4] R. T. Verrillo and G. A. Gescheider, “Perception via the
sense of touch,” in Tactile Aids for the Hearing Impaired (I. R.
Summers, Ed), London: Whurr Publishers, 1-36, 1992.
[5] M. A. Rinker, J. C. Craig, and L. E. Bernstein, “Amplitude
and period discrimination of haptic stimuli,” Journal of the
Acoustical society of America, 104, 453-463, 1998.
[6] A. Israr, P. H. Meckl, and H. Z. Tan, “A two DOF controller
for a multi-finger tactual display using a loop-shaping technique,”
Proc. of the ASME IMECE, 1083-1089, 2004.
[7] M. R. Leek, “Adaptive procedures in psychophysical
research,” Perception and Psychophysics, 63, 1279-1292, 2001.
[8] A. J. Brisben, S. S. Hsiao, and K. O. Johnson, “Detection of
vibration transmitted through an object grasped in the hand,”
Journal of Neurophysiology, 81, 1548-1558, 1999.
385