TV/media listening and hearing aids
1
2
1,3
1,3
4
4
Olaf Strelcyk , Gurjit Singh , Lisa Standaert , Lori Rakita , Peter Derleth , Stefan Launer
1
2
3
4
Sonova U.S. Corporate Services, Warrenville, IL Phonak Canada, Toronto Phonak Audiology Research Center, Warrenville, IL Sonova AG, Stäfa, Switzerland Email: [email protected]
2) TV listening experiment
Average unaided HHIA scores (Newman et al., 1991) were higher for HA users than
for non-HA owners; total HHIA scores of 55 and 38 out of 100, respec vely
 Used automated subject elimina on to balance groups in terms of HHIA scores
and gender, resul ng in 173 non-HA owners and 173 HA users with average score of 46
Q: “Which of these problems do you encounter when watching TV/video?”
Research ques ons:
From our TV/media survey:
Non-HA owner
 85% consider hearing well when watching TV/video as “very important” or
“extremely important”
HA user
q15: Does a hearing problem cause you difficulty
when listening to TV or radio?
Sometimes
 3% “not at all sa sfied,” 15% “slightly sa sfied,” 42% “moderately sa sfied,”
34% “very sa sfied,” 7% “extremely sa sfied” with HAs when watching TV
Error bars show 95% confidence intervals
for the mean
No
q1 q2 q3 q4 q5 q6 q7 q8 q9 q10 q11 q12 q13 q14 q15 q16 q17 q18 q19 q20 q21 q22 q23 q24 q25
 TV listening is not just another speech-in-noise problem; some dis nc ve
features are reviewed below
q23: Does a hearing problem cause you to listen to TV or the radio less o en than you would like?
q17: Does any problem or difficulty with your hearing upset you at all?
q11: Does a hearing problem cause you difficulty in the movies or theaters?
 Yes, it does ma er
Media consump on per day
Viewing loca ons and devices
So far, there are very few studies on TV listening by hearing impaired (HI).
Q: “How many hours per day (on average) do you
watch or listen to the following media?”
Q1: “How o en do you watch TV or video in the following loca ons?”
Q2: “At home, how o en do you watch TV in the following rooms?
Q3: “How o en do you watch TV or video on the following devices?”
Here, we report on two studies, an online TV survey and a TV listening experiment.
4
Hours per day
Non-HA owner
Not just another speech-in-noise problem
3
HA user
 Factors reducing intelligibility: foreign accents and dialects, mumbling and
poor dic on, fast speech, background noise and effects (trend to life-like
ambient audio results in low SNR), background music, dialog in reverb
(VLV, 2011; HBB4all, 2015)
 TV’s large diversity of talkers, unfamiliar accents, speaking rates, and acousc scenes could be challenging for elderly and HI due to their reduced ability to compensate for changing talkers and speaking rates (Sommers, 1997)
 Reproduc on via flat screen TV with down-firing or rear-firing loudspeakers
in reverberant rooms
—Low direct-to-reverb ra o: Listener sits outside cri cal distance
—Diminished spa al cues unless using dedicated center speaker for dialog
(Shirley, 2013)
—Dynamic-range compression (DRC) is enabled by default on most TVs;
could lead to over-compression a er HA DRC
 Lip-reading not available when speaker not facing the camera or when
dubbed
 One advantage to real life: Closed cap oning
Broadcast Video on
TV
demand
Video
media
Radio
Music
Audiobooks
(no significant differences between non-HA and HA users)
 Respondents watch or listen to
TV/video on average 6 h/day
(consistent with Nielsen, 2015)
Use of RC to change volume
Types of loudspeakers
(no significant differences between non-HA and HA users)
 80%
use speakers in the TV
 32% use 2, 3, 5 or more external
speakers (consistent with CEA, 2014)
 14% use soundbars
(consistent with CEA, 2014)
 12%
use headphones
or TV ears (w/o HAs)
 1.2% use HA induc on loop systems
 3.5% use TV-to-HA streamers
Sa sfac on with TV loudness
Q1: How sa sfied are you typically with the loudness
of the TV/video when you set the volume yourself?
Q2: How sa sfied are you... when someone else with
good hearing sets the volume?
 Online
survey via Amazon Mechanical
Turk and SurveyMonkey in USA, 2015,
all data based on self-report
7%
 518
HI respondents, 256 non-HA owners
and 262 HA owners, 58% bilateral,
42% unilateral; PSAP owners not included
here
9%
10%
15%
36%
38%
30%
32%
35%
44%
 241
HA users who use HAs some mes/o en/always and 21 non-HA
users who use their HAs rarely/never when watching TV (the la er
are 3.3 mes more likely to use headphones or TV ears instead)
12%
2%
29%
42%
25%
10%
1%
21%
 73%
 Groups
 Sa sfac on is low when someone else
with good hearing sets TV loudness.
of HA users have more than one year experience, only 4% less than six
weeks; 74% use HAs more than 4 h/day, 6% less than 1 h/day
matched in age, median of 56 years, ranging from 51 to 66 (10th and
90th percen les)
10
0
Commercials are
excessively loud
I cannot
I need to pay full I get annoyed
I cannot
When I set the
I have great
When the
I cannot
understand attention to the
when I am
understand
TV volume,
difficulty
speech is set to understand the
everything the TV in order to watching TV and what is being others complain understanding a comfortable vocals in music
newscaster is
understand other people are said because
it's too loud
actors who level, loud sound videos or music
**: q < 0.01
saying
what is being
talking in the
background
speak with an effects tend to
programs
*: q < 0.05
said
room
music and sound
accent
be uncomforteffects are too
ably loud
(2 tests with mul ple-tes ng correc on
loud
by Benjamini and Hochberg, 1995)
100
90
80
70
60
50
40
30
20
10
0
Barking of I have difficulty I get tired
lip-reading because I have
dogs in the
to expend
when the
room is
particularly speaker is not much effort to
understand
facing the
disturbing
dialogs
camera, has
facial hair, or
when the lips
are out of sync
with the audio
When the
speech is set
to a comfortable level, I
cannot hear
soft sound
effects and
background
music
I cannot follow
TV dialogs
when the
conversation
switches from
one person to
another often
I hear the
Music sounds Women are Sound effects The TV audio
sounds more sounds as if
are less
less dynamic, more difficult
reverberant they originated
powerful, and to understand powerful,
in my head
(like in a
present, and
than men
alive than I
instead of
church or
alive than I
would like
would like bathroom vs. a from within
bedroom) than the room
I would like
Compared with HA users, more non-HA owners...
 cannot understand everything the newscaster is saying
(some men on background noise, conversa ons with field reporters, or fast speech as problema c)
 29% find barking of dogs par
Note: 37 to 47% of US households have a dog  2 out of 3 dog owners disturbed by barking?
29
Q1: When you are facing the TV or video, how difficult is it for you to understand speech on the TV or video?
Q2: When you are NOT facing the TV or video, how difficult is it... ?
“Only once a er switching on the device”
15
Note: The 21 non-HA users who do not use their aids when watching TV
(use TV ears instead) and also the 13 TV-to-HA streamer users tend to be
more sa sfied than the other HA users when someone else sets the TV
loudness.
 About 5% also men
on problems with closed cap oning in open responses
Understanding speech Understanding speech
when NOT facing TV
when facing TV
1%
6%
1%
8%
35%
32%
10%
21%
"Slightly difficult"
 Many have difficulty with
understanding speech when
they are not facing the TV or
video (absence of lip reading,
"Not at all difficult"
visual context and closed cap ons?)
6%
"Extremely difficult"
21%
"Very difficult"
"Moderately difficult"
42%
40%
40%
42%
24%
HA user
HA
user
Non-HA
Non-HA
owner
owner
3%
6%
Non-HA
Non-HA
owner
owner
HA
HA user
user
Q: When you have difficulty understanding speech on TV/video, what do you typically do to
“fix” the problem?
9
“O en”
14
“Some mes”
35
“Rarely”
“Never”
23
20
100
90
80
70
60
50
40
for fewer HA users to turn up volume of
TV (they turn up HAs instead)
 Only 3.5% change HA program
Mono
Stereo
Spatial
Mono
100
Stereo
Spatial
100
11 HI subjects
80
60
40
20
0
d
e
d
f me
ears
hea
dm
hea
o
y
n
y
y
i
t
h
m
Be
f ro n
of m ween
ve m
r
n
o
I
e
b
t
t
n
A
e
or b
e ce
h
n
t
I
In
8 NH subjects
80
60
40
20
0
d
e
d
f me
ears
hea
dm
hea
o
y
n
y
y
i
t
h
m
Be
f ro n
of m ween
ve m
r
n
o
I
e
b
t
t
n
A
e
or b
e ce
h
n
t
I
In
5 HI did not pass test-retest on dicho c noise s mulus (therefore excluded)
HI and NH show similar posi on ra ngs; both groups externalize sounds to
similar extents with Spa al enhancement
Conclusions

27.7%
30

20
10
on ra ngs for 5 clips (“I hear the sounds as if loudspeakers were placed…”)
 Trend
3.5%
1.2%
0
“Close the door”

“Mute TV”
“Tell everyone to shut up”
(no significant differences between non-HA and
HA users)
“Live with it”
“Read lips”
 58% use closed cap ons
some mes”, “o en” or “always”
 Posi
(no significant differences between non-HA and HA users)
Q: “How o en do you use closed cap oning/
sub tling when watching TV/video?”
“Always”
Consistent with NH preference ra ngs in Lorho and Zacharov (2004) and Lorho (2005)
18%
15%
Fixing problems with speech
(in percent)
Both HI and NH rate mono on average as least realis c
NH show trend to prefer stereo over spa al enhancement
26%
Use of closed cap oning
Non-HA owners and
HA users
ra ngs averaged across par cipants and 5 clips:
cularly disturbing
“At the beginning of a new program”
grams, channels or media
 Decrease sec ons that are too loud (effects/music)
 Adjust for external noise, e.g., street noise, AC, space
heater, cooking, dogs barking, kids playing, people talking
 Realism
they need to pay full a en on to understand
 have difficulty understanding accents
 complain about loud sound effects being too loud
 cannot understand vocals in music
Difficul es understanding speech
Reasons for volume changes (in order of frequency):
 Decrease/mute commercials
 Increase so sec ons to hear (dialog) be er
 Adjust volume for loudness differences between pro-
HI par cipants, bilateral symmetric,
IEC audiograms N3 to N5, median age 71 y
 8 NH par cipants, median age 35 y
 Film/news/sports clips of 60 to 75 sec
dura on played back from Blu-ray player
via Sony flat-screen TV
 Audio: TV stereo headphone out VST plugins  HA direct-audio-input (DAI)
 3 VST plugins switched by par cipant (bu ons randomized):
Mono, Stereo and Spa al enhancement, i.e., stereo Dolby ProLogic II Dolby
Headphone (Dolby Media Emulator v1.1)
 S mulus level: 60 dB(A) at DAI
 HA: Slow-ac ng DRC with NAL-NL2 for HI;
flat 0 to 5-dB inser on gain for NH
 Rated a ributes: realism (audiovisual) and
posi ons of sounds (audio only)
 indicate
31
(no significant differences between non-HA and HA users)
 16
Results
“Before/a er commercials”
6
a ribute ra ngs to running TV clips be used to evaluate processing strategies?
 Do HI and NH listeners perceive differences between mono, stereo and spa al enhancement in streamed TV presenta on?
Methods:
20
19
Extremely
satisfied
Very
satisfied
Moderately
satisfied
Slightly
satisfied
Not at all
satisfied
HA
HA
Non-HA
Non-HA
Non-HA
HA user
user
Non-HA
HA user
user
owner
owner
owner
owner
(no significant differences between non-HA and HA users)
www.sennheiser.com
30
“Every few minutes”
“Never”
 Can
40
Non-HA owners and
HA users
Sa sfac on when se ng Sa sfac on when someone
TV loudness oneself
else sets TV loudness
3%
**
(in percent)
www.bestsoundbarforthemoney.com
Note: 10% of the 68 eliminated HA users with higher
HHIA scores (average score of 77) use TV streamers
1) TV/media survey
Q: “While watching a program on TV/video, how o en
do you typically use the TV remote control to change
volume?”
*
50

0
*
60
(no significant differences between non-HA and HA users)
Respondents watch TV/video mostly
at home in the living room, some mes
bedroom; average room size ca. 20 sqm
They watch mostly on TV set, somemes on PC/laptop
**
70
2
1
**
80
Non-HA owner
HA user
Percentage of ratings
Yes
90
Percentage of ratings
Balanced HHIA scores:
100
Percentage of respondents
 Using ecological momentary assessments via mobile phones, Hasan et al.
(2014) found that conversa ons (33%) and listening to media (31%) were
the two most frequent listening ac vi es of hearing-aid (HA) wearers
Problems when watching TV
Percentage of respondents
 Nielsen (2015) reported that American adults of ages 50+ spend on average
6 h 45 min per day watching TV
TV/media survey — HHIA scores
Percentage of respondents
Does it matter?
Non-HA owner
HA user


TV/media listening is relevant but understudied
TV-to-HA streamer penetra on is low and few switch HA programs while watching TV
HA users report fewer problems with understanding speech on TV/video than
non-HA owners, but many s ll have problems, e.g., when not facing the TV/video
or when someone else sets the loudness
A ribute ra ngs to running clips can be used to evaluate processing strategies
Mono, stereo and spa al enhancement differ in terms of perceived realism and
posi on of sound image for both NH and aided HI listeners
References
CEA (2014). “16th Annual CE Ownership and Market Poten al Study.”
Hasan et al. (2014). “Evalua ng auditory contexts and their impacts on hearing aid outcomes with mobile phones. Pervasive Health ‘14, ICST, Brussels, Belgium. HBB4all (2015). “Online-Umfrage zur Sprachverständlichkeit von TV-Produk onen für Hörgeschädigte,” IRT and rbb. Lorho and Zacharov (2004). “Subjec ve evalua on of virtual home theatre sound systems for loudspeaker and headphones,” AES 116th Conv. Lorho (2005). “Evalua on of spa al enhancement systems for stereo headphone reproduc on by preference and a ribute ra ng,” AES 118th Conv. Newman et al. (1991). “Test-retest reliability of the Hearing Handicap Inventory for Adults,” Ear Hear., 12. Nielsen (2015). “The Total Audience Report Q1, 2015.”
Shirley (2013). “Improving Television Sound for People with Hearing Impairments,” PhD thesis, Salford, UK. Sommers (1997). “S mulus variability and spoken word recogni on. II. The effects of age and hearing impairment,” J. Acoust. Soc. Am., 101. VLV (2011). “VLV’s Audibility of Speech on Television Project will make a real difference,” VLV Press Release, h p://www.ips.org.uk/blog/wp-content/uploads/2011/03/Press_release_VLV-Audibility-Project-110314.pdf