Transcript How loud is allowed? It`s déjà vu all over again!

How loud is allowed?
It’s déjà vu all over again!
Ruth Bentler, Ph.D.
Dept of Speech Pathology & Audiology
The University of Iowa
[email protected]
A little history…









Clinical effort 1973-1980s
Dissertation topic, 1987
Bentler & Pavlovic, 1988
Bentler, Abbas, and Pavlovic, 1990
Gad & Bentler, 1993
Ricketts & Bentler, 1994
Bentler & Cooley, 2001
Bentler & Nelson, 2001
Warner & Bentler, 2002
Terms




UCL
ULC
LDL
TD : my choice, because it isn’t
always loudness the patient is
reacting to; could be another
dimension such as harshness,
tinniness, annoyance, etc
Kochkin report (HJ, 2000)


Only 44% were satisfied when asked
about “comfort for loud sounds”
Of the 25 categories on the survey related
to hearing aid performance and different
listening environments, only two items
received lower satisfaction rating: use in
noisy situations, and listening in a large
group
Can we rely upon manufacturers to get
it right?

We can’t rely upon them to agree!
(more later)
Can we rely on formulas to get it right?
…more later on this topic…
Problems with inappropriately set output
…too high…
…or too low (reduced DR)
Measurement issues

Test-retest reliability




Must use structured approach
Must average several runs
1 in 5 cannot do the task
What to do with the HL values when
you get them?
Variability across hearing levels
Variability across gender, age
How does all this compare to the
famous “Pascoe data”?




Recall the graph…
Used in a number of manufacturerbased prescriptive formulas
Pascoe: For hearing levels of 20 to
60 dB HL, 2cc TDs range from 105
to 110 dB (or 10 dB higher than our
large data set.
Pascoe used a protocol that pushed
the TD upward.
Pascoe quote

“Several ascending sequences are
presented at each frequency,
usually starting at progressively
higher levels. This procedure forces
the ‘thresholds of discomfort’
towards higher levels than initially
chosen …is terminated when the
discomfort judgment is not raised
any further.” (page 132)
Pascoe data
Are these clinical measures valid?




i.e, do they relate to real world loudness
experiences?
Probably
Earlier Filion & Margolis data
Recent Munro & Patel data


Significant positive correlation between ratings
(traffic, wind, eg) and difference between
RESR and measured TDs
No such relationship for short duration sounds
(cutlery, door slamming, eg)
What about in situ measures of TD?


i.e., using the patient’s own hearing
aid as the sound generator
Ben Hornsby (Vanderbilt) found that
in situ measures agreed quite well
with manufacturer recommended
OSLP, but varied by as much as 17
dB for the same loss!
Loudness Summation



For years we have believed that TDs
for HI are higher than those of NH
May apply to pure tone stimuli,
which is not real-world!
Due to loudness summation being
greater in HI than NH, evidence of
TDs for complex sounds being lower
for HI than NH
Binaural Summation


Seems to depend upon the level of
input, ranging from 3 dB at
threshold to 6 dB at moderate or
higher level inputs
Seems to encourage the use of
binaural correction for fitting
hearing aids
Power (or channel) summation

The sum of the outputs from independent
channels must be considered





e.g., adding dBs for equal or unequal inputs
e.g, two independent channels of 110 dB
output will equal 113 dB output
e.g., four independent channels of 110 dB
output will equal 116 dB ouput
IF all channels outputting level
simultaneously!
Several suggestions:
Power (channel) summation, cont.

Bentler & Pavlovic derived
regression equation for loudness
plus power summation:


Reduction = 3.95 + 12.88log(n), where
n equals number of channels
Dillon suggestion:


5 dB for 2 channels
9 dB for 4 channels
Why not just predict?

NAL efforts

OSPL3FA = .3T3FA +88.9 dBSPL (<60 HL)
OSPL3FA = .54T3FA +74.3 dBSPL (>60 HL)

Where T is threshold in HL


Results in placement of maximum output ½
way between data of LDL and level that will
saturate speech signal
What about prediction?
Bentler and Cooley (2001, Ear & Hearing)



For hearing levels below 60 dB only
1% of the variance was accounted
for by HL
For hearing levels above 60 dB,
11% of the variance accounted for
by HL
TDs equivalent to CONTOUR results
in high frequencies only
Children?


No reason to believe their TDs are
different than those of adults
Can reliably test children over the
age of 7 or 8 MA

May need alternate approach (crossmodality matching, sad faces, etc)
What about mixed loss?

Not a lot of research in this realm,
but Harvey Dillon’s suggestion of
.875 the air/bone gap sounds
reasonable!
Clinical procedure







Several frequencies in ascending
approach (750 Hz and 3000 Hz best
predictors of TD for complex sounds)
Convert to 2cc values
Consider binaural summation (5dB?)
Consider power (channel) summation (5
dB for 2 CH; 9 dB for 4 CH)
RESR
Real life obnoxious sounds!
Follow-up APHAB, IOI-HA etc
RETSPLs


Determine own
6cc






500
750
1000
2000
3000
4000
13.5
8.5
7.5
11
9.5
10.5
2cc
500
750
1000
2000
3000
4000
8.5
5.0
3.5
6.5
5.5
1.5
Final Words


Don’t count on the subject to
acclimate/adjust to excessive
output
Ask for precise situations of
discomfort in follow up
appointments
Questions?