You are on page 1of 5

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/301555299

Most comfortable listening levels for listening through earphones and in a


sound field

Article · January 1995

CITATIONS READS

0 179

2 authors:

Tomasz Letowski Diana C Emanuel


Sound Research Towson University
232 PUBLICATIONS   1,719 CITATIONS    21 PUBLICATIONS   284 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Bone Conduction Microphones View project

BC Equal-Loudness View project

All content following this page was uploaded by Tomasz Letowski on 28 January 2017.

The user has requested enhancement of the downloaded file.


15th International Congress on Acoustics
Trondheim, Norw ay 26 - 30 June 1995

MOST COMFORTABLE LISTBNING LEVELS FOR LISTENING


TROUGH EARPIIONES AND IN A SOUND FIELD
Tomasz Letowski, Diana W. Emanuel*
U.S. Army research Laboratory Aberdeen Proving Ground, MD 21005
*Towson State University, Towson, MA 21003.

SUMMARY

Most comfortable listening (MCL) levels for connected speech were measured for: (l) monaural
earphone listening, (2) binaural earphone listening, (3) monaural sound field listening, and (4)
binaural sound field listening. Two groups of normal-hearing listeners, age 18-26 (Group 1) and
58-75 (Group 2), participated. Three types of earphones were used: Telephonics TDH-49
supra-aural earphones (both groups), Sennheiser HD-250 circumaural earphones (Group 1
only), and Etymotic ER-3A insert earphones (Group 1 only). In all listening conditions and for
both groups of subjects, the binaural MCL levels were lower than monaural levels by about 4 to
6 dB. Soundfield levels were lower than the TDH-39 earphone levels by approximately 3 to 6
dB. The MCL levels for older adults were similar to those for younger adults for the same
speech material. They went up, however, about 4 to 5 dB when the speech material was
replaced with a TV movie. Results of this study may be useful in selecting and predicting MCL
levels and in comparing MCL data obtained in various listening conditions.

INTRODUCTION

Most comfortable listening (MCL) level is frequently used as the level of sound presentation in
audiology, speech science, psychoacoustics, and sound quality studies. There is, however, a lack
of consensus on how the MCL level depends on the listening conditions, instruction, age of the
listener, or test procedure. For example, the MCL levels for monaural earphone listening to
speech signals have been reported as low as 48 dB SPL (Ventry, Woods, Rubin, and Hill, 1971)
and as high as 85 dB SPL (Richards, 1982). It is generally accepted that MCL is a range of levels
rather than a single level, but this range must be relatively narrow to have a practical value.

The purpose of this study was: (1) to measure and compare MCL levels for connected speech
presented in a sound field and with a variety of earphones and (2) to determine whether MCL
level changes with age. The direct impetus for this study was the authors' attempts to use the
MCL level for speech in noise as a measure of tolerance to noise and a criterion for attenuation
ofspeech by hearing protectors.

METHOD

Two groups of subjects, age 18-26 (Group 1) and 58-75 (Group 2) participated in the study. All
listeners had hearing threshold levels 20 dB HL or better at audiometric frequencies from 250H2
to 4kIlz. The difference between hearing thresholds in right and left ears never exceeded l0 dB

-269 -
at a given frequency. All subjects in both groups were female. The subjects in Group 1 were
itudents of The Pennsylvania State University while the subjects in Group 2 were volunteers
selected from over 200 respondents to a newspaper advertisement.

All tests were conducted in a listening booth (Suttle Bl) having ambient noise level suitable for
ears opentesting (ANSI S3.l-1991). The subject was seated in the center of the booth facing a
loudspeaker at a 1.5 m distance (sound field listening) or wearing earphones (earphone listening).
In the sound field, MCL levels were measured for binaural (both groups) and monaural (Group I
only) conditions. The subject was asked not to move during testing and to hold her head in one
position. Three sets of earphones were used in earphone listening tests: Telephonics TDH-49
supra-aural earphones, Sennheiser HD-250 circumaural earphones, and Etymotic ER-3A insert
earphones. Subjects in Gr<iup 2 used only the TDH-49 earphones. In the monaural condition the
preferred ear was used and the other ear was occluded with an EARrM plug and a Bilsom earmuff
(sound field listening) or covered with an inactive earphone (earphone listening). During the tests
with ER-3A earphones, the foam tips were inserted 12 to 15 mm from the entrance to the ear
canal.

The speech signal was a commercial recording (Auditec, St. Louis) of a story read by a male
talker. The talker had a general American accent and read the text with flat inflection. The signal
was played from a CD player through a Beltone 2000 audiometer and delivered to either the
loudspeaker or the earphones. In addition, the subjects in Group 2 listened to the soundtrack of
the movie "The Princess Bride" and watched the movie on the TV screen. The movie was played
from a VCR and the TV was located in front of the subject and below the loudspeaker level. The
soundtrack was delivered through the loudspeaker or the earphones, depending on the listening
condition.

The MCL levels were obtained using a 2dB-upll0dB-down ascending variant of the method of
limits. Four seconds of speech were presented at each level. The subject's task was to indicate
whether the speech signal was too soft or too loud to be comfortable. The optimum listening level
was obtained three times and the median value was recorded as the MCL value.

The Beltone 2000 audiometer and all transducers were calibrated prior to data collection (ANSI
53.6-1989). The calibration signals were an octave noise band centered at 1000 IIz and the
speech signal used in this study. The calibration accuracy was verified at the end of the study and
the calibration error did not exceed I dB for any set of transducers. The TDH-49 and HD-250
earphones were calibrated using an IEC 318 coupler with a Type I Adaptor (I{D-250). The ER-
3A earphones were calibrated using an ANSI HA-2 coupler. The ER-3A foam plugs were 12 mm
long.

The data collection was part of a larger study on speech perception in noise. Each subject
participated in one listening session. The duration of the session was approximately 1.5 hours but
the MCL testing contributed to less than one-third of the session. The order of listening conditions
was semi-counterbalanced across subjects.

RESULTS AND DISCUSSION

Means and standard deviations for MCL levels obtained for young (Group l) and older adults
(Group 2) in various listening conditions are summarized in Tables 1 and 2, respectively. The
mean monaural and binaural MCL levels for young adults and the sound-field listening situation
vrere found to be about 64 and 60 dB SPL, respectively. These values agree well with the level of
conversational speech, which is typically around 65 dB SPL. They also agree with binaural MCL
levels reported in our previous studies (Letowski, Magistro, and Ritter, 1994; Letowski,
Bursteirq Clarh Romanowski, and Sevec, 1955).

Table l. Means (M) and standard deviations (SD) for MCL levels reported by young
adults (Group l).

Monaural Listening Binaural Listening


M (dB SPL) SD (dB) M (dB SPL) SD (dB)
Sound Field 64.4 6.1 59.7 7.3
TDH.49 70.3 5.8 64.6 6.1
HD-250 65. I 9.4 61.3 8.5
ER-3A 61.0 7.6 56.4 7.5

Table 2. Means (M) and standard deviations (SD) for MCL levels reported by older
adults (Group 2).

Monaural Listening
dB SPL) SD(dB dB SPL) SD (dB
Speech Sound Field 57.7 6.5
TDH.39 66.9 6.4 61.4 6.8
TV Movie Sound Field - 61.4 6.8

The differences between monaural and binaural MCL levels measured for young adults were very
similar for all listening situations (loudspeaker, earphones) and stayed within 4 to 6 dB. The
differences between the soundfield and earphone MCL levels for the same listening condition
(monaural, binaural) were in the 3 to 6 dB range (see Table l). Please note, however, that these
differences are only true for the calibration procedures described in the previous section.

Comparison of the MCL levels for subjects in Group I and Group 2 for the same speech signal is
shown in Figure l. This comparison rweals that MCL levels for older adults are somewhat lower
than for young adults. This difference is not, however, statistically significant (F:1.08, dFl,l8;
p>0'3).Nevertheless, it indicates that the MCL levels for normal-hearing older adults are not
higher than for young listeners. The need for higher speech levels for older people seems only to
result from acquired hearing loss and not from increased age.

The story used as the speech signal in this study was related to the metric system and was read
without inflection by the talker. Such texc was probably not very exciting forthe older adults. As
a resllt, they might have done less than they could to follow the text and were concerned with its
loudness rather than its intelligibility. During the second taslg where the subjects were asked to
set their MCL levels for the soundtrack of an interesting movie, their MCL levels went up by
about 3 to 5 dB. The greatest increase was observed for monaural earphone listening whereas thl
smallest was found for binaural soundfield listening. The average size of this increasi corresponds
well to the difference of 4 to 5 dB between the MCL level for loudness and MCL level
for the
intelligibilir! reported by Hochberg (1975).. Please note, however, that the increase in the MCL
level for a TV program could be also caused by different mode of presentation (audio-video).
MCL Level (dB SPL)

TDH-99 TDH-39 Sound Field


Figure l. Most comfortable listening levels for young (Group l)
and older (Group 2) listeners. Data for connected speech signal.

CONCLUSIONS

The results reported in this study demonstrate that age alone does not affect MCL level for
connected speech. This level, however, can be easily affected by the listener's interest in
the
content of speech. The difference between monaural and binaurA fr4b1, levels for speech seems
to
be about 4 to 6 dB and independent of the listening situation.

ACKNOWLEDGMENTS

This study was supported by a grant from the AARP Andrus Foundation to the first author.

REFERENCES

American National Standards Institute. (1989). Specification for Audiometers. ANSI 53.6-19g9.
Washingtorq DC: ANSI.

American National Standards Institute. (1991). Methods for the ma:<imum permissible ambient
noise levels in audiometric rooms. AI{SI 53.l-1991. washingtorq DC: AI.{SL

Hochberg I. (1975). Most comfortable listening for the loudness and intelligibility of speech.
Audiologt 14,27-33.

Letowski T, Magisto D, and Ritter A. (1994). Most comfortable listening level and speech
attenuation by hearing protectors. J. occup. Safety and Ergonomics (in press).

Letowski T, Burstein N, Clark J, Romanowski L, and Sevec A (1995). Most comfortable


loudness shift as a measure of speech attenuation by hearing proteciors. Am. Indust. Hygiene
Assoc..r. (in press).

Richards AM. (1982). Estimating the listening-comfort of speech. Hearing Instruments33, 38-39.

Ventry IM, Woods RW, Rubin lv! and Hill W. (1971). Most comfortable loudness for pure tones,
noise, and speech J. Acoust. Soc. Am.49. lg05-1913.

View publication stats

You might also like