1
00:00:07,400 --> 00:00:11,200
EASIT: Easy Access
for Social Inclusion Training.

2
00:00:12,300 --> 00:00:15,550
Unit 3B. Easy to understand
and audio description.

3
00:00:15,550 --> 00:00:17,400
Element 3. Technical aspects.

4
00:00:17,400 --> 00:00:20,300
Interview with professionals:
Joel Snyder.

5
00:00:20,650 --> 00:00:24,350
In this video interview
the AD professional Joel Snyder

6
00:00:24,350 --> 00:00:27,300
will discuss some aspects
regarding AD.

7
00:00:27,300 --> 00:00:30,650
He will share his ideas on how
these can contribute

8
00:00:30,650 --> 00:00:33,730
to making AD easier to understand.

9
00:00:35,750 --> 00:00:38,450
Hello, this is Joel Snyder.

10
00:00:38,450 --> 00:00:43,800
I am the president
of Audio Description Associates,

11
00:00:43,800 --> 00:00:47,800
in the USA.
I am the director and founder

12
00:00:47,800 --> 00:00:53,000
of the AD project
for the American Council of the Blind.

13
00:00:54,500 --> 00:01:00,500
Are there specific technical aspects
that could improve

14
00:01:00,500 --> 00:01:06,300
the comprehensibility
and listenability of AD?

15
00:01:06,300 --> 00:01:10,350
That can make it usable
also for people

16
00:01:10,350 --> 00:01:14,120
with cognitive, intellectual
and learning disabilities,

17
00:01:14,120 --> 00:01:16,430
whose processing times
are slower?

18
00:01:16,500 --> 00:01:22,800
I think so. There are technical
and programmatic aspects

19
00:01:22,800 --> 00:01:28,600
or aspects of execution of AD
that can accomplish that,

20
00:01:28,600 --> 00:01:32,700
that can increase
comprehensibility and listenability

21
00:01:32,700 --> 00:01:37,600
for people with intellectual
or learning disabilities.

22
00:01:38,500 --> 00:01:44,800
On a technical side,
there is a way to adjust pace.

23
00:01:46,000 --> 00:01:52,000
That is a critical element
of listening and listenability,

24
00:01:52,800 --> 00:01:56,900
especially for people
with cognitive disabilities.

25
00:01:57,000 --> 00:02:00,000
Slowing down the pace.

26
00:02:00,000 --> 00:02:05,700
That can be accomplished
in the AD through the voicing.

27
00:02:07,750 --> 00:02:12,750
For six years I produced
and supervised the AD

28
00:02:12,750 --> 00:02:15,900
for the American children show
“Sesame Street”,

29
00:02:15,900 --> 00:02:20,200
which is geared towards three
to five year-olds.

30
00:02:21,800 --> 00:02:27,800
That audience requires
a vocabulary and reading level,

31
00:02:27,800 --> 00:02:31,500
that is lower
than the general public.

32
00:02:31,500 --> 00:02:35,500
We were aware of that
in the writing of the description

33
00:02:35,500 --> 00:02:40,400
and I think the same principle
would apply

34
00:02:40,400 --> 00:02:45,800
to making description that is
more listenable for all people.

35
00:02:45,800 --> 00:02:49,300
That is a key point.

36
00:02:49,500 --> 00:02:55,400
By making audio description
more applicable

37
00:02:55,400 --> 00:02:59,700
and more reachable
by more people,

38
00:02:59,700 --> 00:03:03,400
I think we build a visibility of AD

39
00:03:03,400 --> 00:03:08,000
and increase the network of people
who understand

40
00:03:08,000 --> 00:03:10,300
and appreciate audio description.

41
00:03:10,300 --> 00:03:15,800
The audience of people
with learning disabilities

42
00:03:15,800 --> 00:03:18,200
is an important part of that.

43
00:03:18,200 --> 00:03:22,800
But the audience can also
include the general public,

44
00:03:22,800 --> 00:03:26,000
Audio description can be great
for them as well:

45
00:03:26,000 --> 00:03:28,000
they can be in the kitchen,

46
00:03:28,000 --> 00:03:31,700
while the tv is on in the living room,
they do not miss a bit.

47
00:03:31,700 --> 00:03:36,000
I think there are technical
and problematic aspects,

48
00:03:36,000 --> 00:03:41,500
aspects of execution,
that can make AD more listenable.

49
00:03:42,900 --> 00:03:46,800
Do you think that extended ADs
could be a solution

50
00:03:46,800 --> 00:03:50,650
for people with intellectual
or learning disabilities,

51
00:03:50,650 --> 00:03:53,000
whose processing times
are slower?

52
00:03:53,000 --> 00:03:58,300
I do. Extended ADs,
as long as the concept

53
00:03:58,300 --> 00:04:02,900
is done in a way
that is easy to access and understand.

54
00:04:03,200 --> 00:04:07,800
Most extended AD
involves clicking hyper links

55
00:04:07,800 --> 00:04:09,900
or switching
to another screen

56
00:04:09,900 --> 00:04:13,000
or another audio file
that supplements

57
00:04:13,000 --> 00:04:19,000
the AD that exists
on a base level.

58
00:04:19,300 --> 00:04:24,800
It is important to remember
that AD, describers,

59
00:04:25,000 --> 00:04:29,900
are about describing,
not explaining.

60
00:04:29,900 --> 00:04:34,800
I think the explaining
and the elaboration on a concept

61
00:04:34,800 --> 00:04:39,700
is best left to a teacher,

62
00:04:39,700 --> 00:04:43,200
someone who is working
with a given student.

63
00:04:43,200 --> 00:04:48,900
The AD does not need to be charged
with that aspect.

64
00:04:48,900 --> 00:04:53,600
It may provide additional description
in the extended AD

65
00:04:53,600 --> 00:04:58,300
but it is not about explaining
what something is.

66
00:04:58,300 --> 00:05:03,600
We do not explain, we describe.
We do not tell, we show.

67
00:05:05,000 --> 00:05:10,300
Would TTS AD work better
than human voiced AD

68
00:05:10,300 --> 00:05:15,400
for the same audience?
I do not think so.

69
00:05:15,400 --> 00:05:20,100
TTS (text-to-speech)
is always improving

70
00:05:20,100 --> 00:05:23,100
but it is still not at a point,

71
00:05:23,100 --> 00:05:28,400
not today, November 2020

72
00:05:28,400 --> 00:05:31,800
where it can make the
distinctions,

73
00:05:31,800 --> 00:05:37,600
provide the nuance
that increases listenability.

74
00:05:39,000 --> 00:05:42,600
Perhaps for some
dry documentaries,

75
00:05:42,600 --> 00:05:48,600
the reading of a book
or in a non-fiction text.

76
00:05:50,300 --> 00:05:55,000
But even there,
we make meaning with our voices.

77
00:05:55,000 --> 00:05:59,300
I have not met
a text-to-speech engine

78
00:05:59,300 --> 00:06:01,900
that can yet provide
the same nuance

79
00:06:01,900 --> 00:06:04,800
that a trained audio describer
can provide.

80
00:06:04,800 --> 00:06:08,600
When I train audio describers
we cover vocal skills

81
00:06:08,600 --> 00:06:12,000
and we use one example,
for instance:

82
00:06:12,000 --> 00:06:14,600
“Woman without her man is a savage”

83
00:06:14,600 --> 00:06:19,000
That sounds misogynistic,
right?

84
00:06:19,000 --> 00:06:21,000
Who would agree with that?

85
00:06:21,000 --> 00:06:24,200
But say it back
using your voice, no punctuation.

86
00:06:24,200 --> 00:06:28,300
An audio describer can insert
the punctuation themselves

87
00:06:28,300 --> 00:06:31,200
because they make meaning
with their voice.

88
00:06:31,200 --> 00:06:36,200
So it comes out:
“Woman: without her, man is a savage”.

89
00:06:37,400 --> 00:06:40,900
I do not know that a text-to-speech engine
can be taught

90
00:06:40,900 --> 00:06:45,300
to make that distinction
and to bring that nuance.

91
00:06:45,300 --> 00:06:47,600
The other element is tone.

92
00:06:47,600 --> 00:06:53,500
The best audio describers employ
what I call “concinence”,

93
00:06:53,500 --> 00:06:57,850
the tone of voice is in concinence
with what is described.

94
00:06:57,850 --> 00:07:02,200
If it is a happy scene,
there is an upbeat quality.

95
00:07:02,200 --> 00:07:06,600
If it is a serious scene,
there is a sober quality to the voice.

96
00:07:06,600 --> 00:07:11,700
I am not sure that text-to-speech
can be aware

97
00:07:11,700 --> 00:07:17,100
of what happens
over the throughline of a video,

98
00:07:17,100 --> 00:07:22,100
to understand that
“this part needs to be this way”

99
00:07:22,100 --> 00:07:26,250
because of what is on the screen
and what comes later.

100
00:07:26,250 --> 00:07:30,200
That kind of nuance and subtlety
is important

101
00:07:30,200 --> 00:07:33,600
and is something that
a good writer of description

102
00:07:33,600 --> 00:07:37,000
and a good voicer of
description can provide.

103
00:07:37,800 --> 00:07:40,500
This video was prepared
by Elisa Perego,

104
00:07:40,500 --> 00:07:43,500
produced by
Angelika De Markis, Laura Marini,

105
00:07:43,500 --> 00:07:46,700
Annalisa Navetta
from University of Trieste.

106
00:07:46,700 --> 00:07:48,700
Narrator: Annalisa Navetta.