In the near future a household will have
access to about 500 channels through a satellite connection. This implies more than 12.000
programmes in one day, and 9 minutes of channel hopping to go through all channels. Even
at this moment, having only about 30 to 40 channels, most of us only regularly watch about
7 channels. Simply because it is too much trouble to find every programme of interest on
all channels.
Since the chance of simultaneous broadcast of interesting programmes is getting bigger
with the number of channels, the urge for recording programmes is growing. Another major
reason to record programmes is the fact that people want to keep and own content (mostly
series and movies).
Thus, two related problems, from a user point of view, can be identified: finding the
programmes one wants to watch, and finding the programmes one wants to keep.
On the other hand the broadcasters want their programmes to end up being viewed and
recorded.
This implies that it is in the users interest as well as the broadcasters’ interest
to make an easy and fitting match between content (programme) and user. Therefore, the
focus in developing the STORit system is to provide ways to help the user find programmes
of interest to watch (view) and to record (keep).
At the user end of the broadcast delivery chain, in this case that would be the actual
STORit box, three ways of helping the user to find the content of interest are
implemented:
At the broadcasters end of the broadcast delivery chain additional Metadata (data, broadcast along with the actual content, that contains descriptors of that content) and services are implemented, again to help the user to find the content of interest:
Subsets actually are filters that are under full control of the user. When a subset is active, only the content (programmes), which fit the description of that subset, are presented to the user. One of the subsets is called ‘all’; this subset can not be altered, since it simply encapsulates all content. Furthermore the system comes with a number of default subsets based on high level genre information conform the Metadata descriptors used:
The user can alter, remove and add new subsets. Each subset has numerous entries that can be edited:
Only the name of the subset is not used for
filtering purposes, it is only used in order to allow the user to recognise and
distinguish the subsets.
The description of a subset, on a system level, consist of AND/OR rules based in the
content of the entries that can be edited. Basically AND rules are used for entries on
different levels, whereas OR rules are used for entries on the same level.
The different bullet-points in the list above have a mutual AND relationship. The entries
within an each bullet-point have a mutual OR relationship, except for genre and sub-genre.
For the latter an AND rule must be used, because the sub-genre is on a different level: a
sublevel to genre. Furthermore, each entry can consists of several items. For example
multiple genres can appear in the description: movies, entertainment, etcetera. These
items have a mutual OR relation.
Thus very simple subsets can be defined as well as very complicated subsets. For example a
subsets can have a description such as ‘genre: movies, sub-genre: western’,
which can be written as:
This subset will only show the western movies.
A subset can also can have a
more complex description such as: ‘genre: movies, sub-genre: western, sci-fi and
horror, genre: entertainment, sub-genre: factual, language: English, sub-title language:
English, day of the week: Friday, Saturday’, which can be written as:
This subset will show all western movies,
science fiction movies , horror movies and factual entertainment, either English spoken or
with English subtitles, on Fridays and Saturdays.
When a subset is activated, it could be possible that the Electronic Programme Guide (EPG)
does not show any content in some at some points in time. A user then might get lost in
the user interface. Therefore a zoomed out view is presented at the same time, providing
help in navigating the EPG. When navigating within the EPG a highlight provides feedback
on the current location in the EPG, a second highlight simultaneously provides feedback on
the location in the zoomed out overview. In this way a direct link is established between
the EPG, call it the field of vision, and the zoomed out overview, call it the radar
vision.
A problem in using user profiles is the so
called ’cold start’ problem. At first time use the system has no knowledge on
the user whatsoever. Of course the system could state all kinds of questions or present a
list for the user to fill in, but this is not in line with the experience people have in
the television domain. Even in the PC domain where questionnaires are more accepted, it
would be a bothersome task.
A solution to this problem are the Lifestyle icons (L-icons).
Furthermore (L-icons) are implemented to enhance the interaction and lighten up the user
interface. A L-icon will present itself by telling its name and its personal interests (in
programmes). The representation, behaviour, name and interests of a L-icon all are in line
with a specific lifestyle. A user can select a personal L-icon, probably one that is the
closest match: just as in making friends, similar interests form a base of familiarity and
trust.
When asked, a L-icon can state its opinion on programmes and behave accordingly. Since the
user has selected a personal L-icon based on similar interests, the user is likely to be
more interested in programmes the L-icon likes and less interested in programmes the
L-icon dislikes. Thus, the user can decide for himself what to do with the remarks made by
the L-icon: ignore or use, but ideally the user does mind what the L-icon states, they are
friends after all.
13 Different L-icons are implemented, based on lifestyles including kids and teenagers,
both boys and girls.
Each L-icon is presented using 6 different animations, 1 introduction and 5 expressions:
The neutral state can also be used when no programme is selected to state an opinion on.
In the following table the relation between these different states and the Metadata is given:
| Expression | Relation with DVB-SI metadata genre | |
| Excited | General,
but not humour, comedy, news, current affairs, instruction
or religion Good match with profile |
|
| Happy / fun | Humour
and comedy, but not news, current affairs, instruction
or religion Good match with profile |
|
| Serious | News,
current affairs, instruction and religion,
but not humour or comedy Good match with profile |
|
| Neutral | General Not so good match with profile |
|
| Not interested / dislike / neutral | General Poor match with profile |
|
The L-icons are 3D models and fully
animated, adding an emotional dimension to their stated opinions. Not only to enlighten
the user interface but also because emotional expressions are recognised in the blink of
an eye and thus enhance the communication of their opinions towards the actual user.
For the IBC demonstrator, the L-icon profiles where entered by hand, partly based on
existing market research, partly based on gut feeling.
The L-icons as implemented are a starting point; more precise descriptions for L-icons are
needed depending on region if the STORit box would be an actual product for sale in the
local shops.
Figure 2 : The
STORit EPG, with in the top part the zoomed out view, in the middle the actual EPG and at
the bottom
part some additional information on the programme currently selected.
Figure 3 : The
STORit EPG, with an activated sub-set. Here it is clearly visible
how the zoomed out view supports the use of the subsets.
Figure 4 : The
STORit EPG, with an activated sub-set and a L-icon
stating its opinion on the currently selected programme.
Figure 5 : During
regular television viewing, the user still can display additional information
on the programme being watched, and still ask the L-icon its opinion on that programme as
well.
Figure 6 : A screenshot of the set-up menu where the user can select a (different) L-icon.