The Cambridge aided loudness profile (CALP): item choices and data for unaided listening by young and older normal-hearing listeners and older hearing-impaired listeners

Abstract

Objective

To develop an improved version of the profile of aided loudness (PAL), intended for assessment of the appropriateness of the loudness of everyday sounds.

Design

Initially, 16 participants with a range of ages and degrees of hearing loss indicated whether they encountered each situation described in the PAL and how specific they considered the description to be. Based on the responses, most situations from the PAL were eliminated and new situations were introduced, giving the Cambridge Aided Loudness Profile (CALP). The CALP was administered to 80 young and 22 older participants with normal hearing, who rated the loudness of each situation and satisfaction with this loudness (as for the original PAL). Satisfaction was strongly negatively correlated with loudness, suggesting that satisfaction was largely based on loudness. The CALP was then administered to 32 new young normal-hearing participants and 49 older participants with hearing loss, most of whom used hearing aids, who rated loudness and the appropriateness of loudness.

Results

Some situations were rated as loud but appropriate in loudness, indicating that the CALP can distinguish these aspects.

Conclusions

The CALP questions were understood by all participants. The CALP may be useful for assessing the appropriateness of loudness.

Keywords:

• Loudness
• satisfaction with loudness
• appropriateness of loudness
• aided listening
• evaluation of hearing aids

Introduction

The primary goal of hearing aids is to increase the audibility of weak sounds while providing appropriate loudness for sounds of medium and high intensity (Dillon 2012; Moore 2007). The focus has been on the audibility and loudness of speech sounds (Dillon 2012), but environmental sounds should also have an appropriate loudness. The profile of aided loudness (PAL, Mueller and Palmer 1998; Palmer et al. 1999) was designed to assess the perception of the loudness of everyday sounds for users of hearing aids. Respondents are asked to rate a series of items (e.g. a door slamming, a lawn mower) in terms of: (1) Loudness, on a scale from 0 (=“Do not hear”) to 7 (=“Uncomfortably loud”); (2) Satisfaction with loudness, on a scale from 1 (=“Not good at all”) to 5 (=“Just right”). For example, if a specific sound has a loudness that is rated as “Soft” and the preferred loudness for that sound is “Soft,” the satisfaction rating should be “Just right”, while if the preferred loudness is “Comfortable”, the rating might be “OK” or “Not too good”.

Moore and Füllgrabe (2010) used the PAL to assess loudness perception for hearing-impaired participants before and after wearing hearing aids for about four weeks. The hearing aids had an extended high-frequency response and were fitted using the CAMEQ2-HF fitting method (Moore et al. 2010). Many participants had not experienced all of the situations described in the PAL during the period they wore the aids. Some also complained that the situations were not described precisely enough. For example, the description often did not specify the distance of the sound source; a dog barking might be loud when nearby but soft when far away.

While the concept behind the PAL seems useful, the problems noted above limit its applicability and accuracy. The goal of the present study was to develop a new version of the PAL based on situations that were encountered by a majority of participants and using more specific questions, so that loudness judgements would be more accurate and more consistent across participants. We chose situations that would be encountered reasonably often in most countries and for which sound levels would be reasonably consistent. Data for the new version of the PAL, called the Cambridge aided loudness profile (CALP), were obtained for British participants with both normal and impaired hearing.

Initial selection of situations for the CALP

The PAL was given to a randomly selected group of 16 people (age range 18–74 years, pure-tone average (PTA) threshold across 0.5–4 kHz, 5–55 dB HL). They were given the PAL in paper form and asked to fill in their responses either in the laboratory or at home (this was also the case for the other studies described in this paper). They were asked to comment on how generally applicable, specific and/or representative of everyday life they considered each situation to be and were asked to give written suggestions for new situations falling into the categories “soft”, “average” and “loud”. Based on the responses, eight of the 12 situations in the PAL were omitted, because the situation was not encountered (“a marching band”), the situation was applicable only to one gender (“An electric razor”), or the situation was not sufficiently specified (“A dog barking”—is it a large or small dog and how far away is it?).

Twelve new situations were selected that were suggested more than once and that seemed sensible and well defined. Four situations were retained from the PAL, but the wording of two of them was modified to improve clarity and to make the situation more specific (e.g. “A door slamming” was changed to “A door slamming, standing a few feet away”). Three situations were taken from the APHAB (Cox and Alexander 1995), but two were modified to make them more specific. The resulting CALP contained 19 situations (see Appendix). For brevity, the situations are referred to as “items”. Participants received written instructions on how to complete the CALP.

Evaluations with young and older normal-hearing participants

Participants

Initially, two groups completed the CALP:

1. Young with normal hearing (YNH). There were 54 female and 26 male participants with ages ranging from 18 to 38 years (mean = 25 years; standard deviation, SD = 5 years). All participants had audiometric thresholds ≤20 dB hearing level (HL) in both ears for all audiometric frequencies between 125 and 8000 Hz (including 3000 and 6000 Hz) (British Society of Audiology 2004).

2. Older with normal hearing (ONH). There were 21 female participants and 1 male participant, all with no reported hearing difficulties. Their ages ranged from 62 to 80 years (mean = 68 years; SD = 5 years). Their audiometric thresholds were ≤20 dB HL for both ears for frequencies up to and including 6000 Hz; some had hearing loss (up to 45 dB) at 8000 Hz.

Instructions

For each item, participants were asked to rate both loudness and satisfaction with loudness, as for the original PAL. The instructions were:

1. For loudness: “Please rate the subjective level of loudness (volume) for the following situations or sounds (from ‘Do not hear’ to ‘Uncomfortably loud’). Give only one rating per situation/sound.”

2. For satisfaction: “Please also rate your satisfaction with the loudness indicated in the previous section, that is how happy you are with the loudness of the following situations or sounds (from ‘Not good at all’ to ‘Just right’). Give only one rating per situation/sound”.

Loudness ratings for normal-hearing participants

Participants were able to give answers for all items, since they had encountered the situations described, or ones very similar to them. Figure 1 shows the mean loudness ratings for each group. Except for item 18 (“Hum from the motor of a fridge”), the differences in mean rating across groups were always less than one unit on the seven-point scale. However, based on an independent-samples t-test, the difference in mean ratings between the two groups was significant [t(58.7) = −3.07, p = 0.003]; the ONH participants gave slightly higher mean ratings than the YNH participants.

Figure 1. Mean loudness ratings for YNH (filled symbols) and ONH (open symbols) participants for each CALP item. Error bars indicate ±1 SD across participants. Squares at the far right show overall means.

The variability of loudness judgements across participants, indicated by the error bars, varied across items. The large variability for item 18 may reflect the fact that the loudness of a fridge motor depends on the make and model of the fridge.

Satisfaction ratings of loudness for normal-hearing participants

Figure 2 shows the mean ratings of satisfaction with loudness. The mean ratings were similar across groups for each item; the largest difference was about one 0.5 scale units. Based on an independent-samples t-test, the difference in mean scores between the two groups was not significant [t(100) = −0.13, p = 0.90, two-tailed]. Thus, satisfaction with loudness was unaffected by age.

Figure 2. As Figure 1, but for ratings of satisfaction with loudness. For this scale, higher numbers indicate greater satisfaction.

Some items received low satisfaction ratings, especially item 15 (“Siren of fire truck or police car driving close by”). Presumably, the low satisfaction reflects the fact that the sound was judged as “Loud” or “Uncomfortably loud” by most participants. The concept behind the original PAL was that satisfaction should be related to the appropriateness of loudness; some sounds, such as the siren of a police car, are meant to be loud. Hence, the siren should have been rated as “Okay”. Evidently, our participants did not interpret the instructions in this way. The imagined siren was judged as “Loud” or “Uncomfortably loud”, and this was associated with a mean satisfaction rating of “Not too good”. The participants probably judged satisfaction partly on a pleasantness-aversiveness scale. The likely influence of loudness/aversion on the satisfaction judgements is unfortunate, as a poor rating of satisfaction with loudness might be given for an item that is heard with a similar loudness to that for people with normal hearing (listening unaided).

Further evaluation with YNH participants based on appropriateness judgements

Rationale and instructions

To examine the distinction between satisfaction with loudness and appropriateness of loudness, 32 new YNH participants (nine female) were recruited. Their mean age was 24 years (range = 18 to 38 years, SD = 5 years). They were asked to complete the CALP, but rating loudness and the appropriateness of the loudness of each item. The instructions for loudness ratings were the same as described above. The instructions for appropriateness were: “Please rate also how appropriate the loudness (volume) is for each of the situations or sounds. Here, we are not asking about how much you like the loudness, but about whether you feel that the loudness is reasonable given the situation. For example, it might be appropriate for an alarm sound to be loud.” The rating scale used the same categories as for the “satisfaction” scale described earlier.

Loudness and appropriateness ratings for YNH participants

The grey symbols in Figure 3 show the mean appropriateness rating (referred to as the right axis) for each item, plotted as a function of the mean loudness rating for that item. For comparison, the results for satisfaction ratings (left axis of Figure 3), as shown in Figure 2, are plotted against the loudness ratings shown in Figure 1. The satisfaction ratings were highly correlated with the loudness ratings (for both groups, r = −0.92, p < 0.001). Thus, the satisfaction ratings provide little information above that provided by the loudness ratings. However, for some items, the satisfaction ratings were low, but the appropriateness ratings were reasonably high. This is shown more clearly in Figure 4, which shows the mean and SD of the satisfaction ratings (filled symbols) and appropriateness ratings (grey symbols) for each item. Especially large discrepancies between satisfaction and appropriateness occurred for items 1 (“A door slamming, standing a few feet away”) and 15 (“Siren of a fire truck or police car driving close by”). The lowest mean appropriateness rating was for item 4 (“A food blender, standing next to it”), perhaps because there is considerable variability in the sound levels produced by food blenders.

Figure 3. Mean satisfaction (left ordinate) and appropriateness (right ordinate) ratings for loudness plotted as a function of mean loudness rating for each of the three groups (see figure key). Each symbol represents one of the items. Item numbers are given next to symbols for items for which the satisfaction and appropriateness ratings were markedly different.

Figure 4. Mean ratings of appropriateness for YNH participants (light grey symbols). Ratings of satisfaction are replotted from Figure 2 (filled black symbols).

For hearing-aid applications, judgements of appropriateness seem more relevant than judgements of satisfaction. A hearing aid must lead to a level of loudness that is appropriate for a variety of listening situations. For further evaluation of the CALP using hearing-impaired participants, we gathered ratings of loudness and appropriateness rather than loudness and satisfaction.

Ratings of loudness and appropriateness by hearing-impaired participants

Our next goal was to gather data for the CALP for older participants with hearing loss, denoted OHL.

Participants

There were 49 OHL participants. The PTA for the audiometric frequencies of 500, 1000, 2000 and 4000 Hz, averaged across the two ears, was used to divide the participants into a “mild” group (PTA < 40 dB; 16 female, 12 male; mean age = 70 years, SD = 7 years, range = 49–85 years) and a “moderate” group (40 dB ≤ PTA ≤ 70 dB; 12 female, 9 male; mean age = 71 years, SD = 11 years, range = 46–89 years). The absolute value of the interaural difference in PTA was, on average, 6 (SD = 4) and 4 (SD = 3) dB for the mild and moderate groups, respectively. The mean audiometric thresholds for the two groups are shown in Figure 5. Thirty six percent of the mild group wore a hearing aid unilaterally (30%) or bilaterally (70%). The average duration of use was four years, with a minimum of six months. All aids were of the behind-the-ear (BTE) type. Ninety five percent of the moderate group wore a hearing aid either unilaterally (20%) or bilaterally (80%). The average duration of use was 12 years, with a minimum of one year. Eighty percent of the aids were of BTE type, and 20% were of in-the-ear type.

Figure 5. Mean audiometric thresholds (in dB HL) for the four frequencies used to compute the PTA for 28 participants with mild hearing loss (continuous line) and 21 participants with moderate hearing loss (broken line). Shaded areas around the mean indicate ±1 SD. The two panels show thresholds for the left and right ears, respectively.

Instructions

Ratings for the loudness and the appropriateness of loudness for each item in the CALP were obtained using the same instructions as above, except that participants who normally used hearing aids were asked to give ratings for their experience when not wearing hearing aids.

Loudness ratings for OHL participants

Figure 6 shows the mean loudness ratings for the mild group (filled symbols) and the moderate group (open symbols). The pattern of the ratings was similar across groups, but the moderate group gave lower mean loudness ratings than the mild group for every item. The difference in mean ratings across groups was significant: t(47) = 5.31, p < 0.001, two-tailed). The pattern of the ratings was similar to that for the YNH and ONH groups, as shown in Figure 1. However, loudness ratings for the OHL groups were somewhat below those for the YNH and ONH groups. The overall mean ratings were 4.0 (“Comfortable”) for the YNH and ONH groups and 3.5 and 2.3 for the mild and moderate groups, respectively.

Figure 6. Mean loudness ratings for the mild group (filled symbols) and the moderate group (open symbols). Otherwise as in Figure 1.

Appropriateness ratings for OHL participants

Figure 7 shows the ratings of the appropriateness of loudness for the OHL participants. For the mild group, the mean ratings mostly fell in the range 3 (Okay) to 4 (“Pretty good”). For the moderate group, most mean ratings fell close to 3, but a few fell closer to 2 (“Not too good”). The lowest mean appropriateness rating for the moderate group was for item 3 (“A religious leader/presenter during a sermon/lecture in a larger space, using a microphone”).

Figure 7. As Figure 6, but for ratings of appropriateness.

Discussion

Unlike for the original PAL, YNH, ONH and OHL participants were consistently able to give responses for all items in the CALP. The pattern of loudness ratings across items was very similar for the different participant groups, but mean loudness ratings tended to decrease with increasing hearing loss.

In the original PAL, the sounds were grouped into three categories, “soft”, “average” and “loud”. The target loudness ratings for sounds in these three categories were 2, 4 and 6, respectively. Here, these categories are referred to by the more physical descriptors of “weak”, “average” and “strong”. We consider next the subsets of items that can be considered representative of these categories. Items were excluded if the SD of ratings across participants was unusually large. The mean loudness and appropriateness ratings across the items selected for each category are shown for each group in Table 1.

Table 1. Means of loudness and appropriateness ratings for each group for subsets of items representing the categories weak, average and strong.

Category	Items	Group	Loudness	Appropriateness
Weak	2, 5, 13	YNH	1.6	4.1
		ONH	1.7	n/a
		Mild	1.7	3.8
		Moderate	0.9	2.8
Average	9, 11, 12, 17	YNH	3.9	3.8
		ONH	4.1	n/a
		Mild	3.6	3.5
		Moderate	2.4	2.8
Strong	1, 4, 7, 15	YNH	6.0	3.5
		ONH	5.8	n/a
		Mild	5.2	3.4
		Moderate	4.0	3.1

n/a: not available.

Items 2, 5 and 13 received mean loudness ratings close to or a little below 2 for all groups and represent the category “weak”. Appropriateness was high (close to or above 4) for the YNH and ONH groups and decreased with increasing hearing loss, presumably because the sounds were judged as too soft for the hearing-impaired groups when listening unaided. One would expect appropriateness for weak sounds to increase for hearing-impaired participants when listening via hearing aids, but this remains to be investigated.

Items 9, 11, 12 and 17 received mean ratings close to 4 (“Comfortable”) for the YNH and ONH groups, a little below 4 for the mild group, and close to 3 for the moderate group. These items represent the category “average”. Appropriateness was high (close to 4) for the YNH and ONH groups, and for the mild group, but dropped to 3 for the moderate group, again presumably because the sounds were judged as too soft for the latter when listening unaided.

Items 1, 4, 7 and 15 received mean ratings close to 6 (“Loud”) for the YNH and ONH groups, close to 5 for the mild group, and close to 4 for the moderate group. These items represent the category “strong”. Appropriateness was judged reasonably close to 3 (“Okay”) for all groups. Appropriateness for hearing-impaired participants might go down if they were fitted with hearing aids that provided gain even for strong sounds.

Based on these findings, a shortened version of the CALP could be used based on items 1, 2, 4, 5, 7, 9, 11, 12, 13, 15 and 17. This would contain at least three items representing each category of sound—weak, average and strong.

Limitations

It is not clear whether the results are applicable to participants from countries other than the UK. Also, the CALP depends on the participants having an accurate memory of loudness for the items in the CALP. This could be checked by comparing CALP responses to those for “ecological momentary assessment”, which obtains loudness judgements in “live” situations (Glista et al. 2021). Finally, no data were obtained on the test-retest repeatability of the CALP or on the ability of the CALP to assess individual changes in appropriateness of loudness produced by hearing aids. Further research is needed to address these issues.

Conclusions

The CALP was developed to assess the loudness and appropriateness of loudness of everyday sounds, based on the ideas behind the PAL. The items in the CALP represented situations that were familiar to all participants and were well defined. When participants were asked to judge satisfaction (as for the original PAL), the responses were probably partly influenced by loudness/aversion. When participants were asked to judge appropriateness, the responses seemed to be less affected by loudness/aversion. Items in the CALP representing the categories of weak, average and strong sounds were identified. The CALP may be suitable for assessing the effects of hearing aids on the loudness and appropriateness of loudness of everyday sounds, but this remains to be determined.

Ethical approval

This study was approved by the Cambridge Research Ethics Committee. All participants gave written informed consent.

Acknowledgments

Martin McKinny is thanked for helpful discussions. Maaike van Eeckhoutte and an anonymous reviewer gave helpful comments on an earlier version of this paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by Starkey (USA), the Medical Research Council (UK, Grant NO. G0701870), and a Wolfson College, Cambridge Junior Research Fellowship to author CF.