Hypervideo
1.0 INTRODUCTION
On the World Wide Web, hypertext has provided a powerful mechanism for
organizing distributed multimedia as information text-based pages augmented with
image, sound and video elements. A broader basis is necessary to structure and
integrate a video-centric medium, where the notion of hot-links must be
redefined to consider the spatial and temporal flow of video frames. New
interaction paradigms must be developed to navigate a hierarchy of hyper-linked,
video-driven content called Hypervideo. Video object segmentation and object
recognition tracking technology can automatically tracks designated objects
throughout a series of consecutive frames. Embed time-driven events provide
hotspots into the digital video. These "hotspots" can then be linked to a
variety of media including URL addresses, other video/sound files, pop up text
boxes, and specified frames in the same video. The hotspot itself can be
highlighted, or an image can be superimposed over the hotspot, giving users a
clear indication of exactly where to click. The focus of hypervideo is on the
navigational issues related to browsing hyper-linked video nodes. Video-to-video
linking was demonstrated at the Interactive Cinema Group (MIT Media Lab). Image
mapping of timed events to frame location during video streaming establishes a
linked reaction.
In traditional hypertext, nodes, links and writing spaces
provide a structure for hypertext documents. In hypervideo, the framework is
complicated by the temporal and spatial nature of the medium. New structural and
navigational concepts have been developed to provide a unified approach towards
hypervideo. The video spaces in hypervideo require a dynamic representation that
encapsulates moving content, space, and time for navigation. Hypervideo requires
a hierarchy to permit the creation and navigation of a network of hyper-linked
video scenes. With the recent progress in video segmentation, object-tracking
and content analysis techniques, objects in the frame of moving video could be
automatically detected and tracked in real-time. Hot-spot linked to this
segmentation and tracking can jump the user to another video, graphic or text,
or show windows to all at once. An authoring tool for hypervideo must provide
suitable visualization and interaction techniques to make the authoring activity
intuitive for content developers. One approach to developing such a tool is to
consider the authoring process for creating hypervideo.
The tool should support
the different phases of conceptualization, video production, video editing and
post-production, and design of narrative sequences and link opportunities. Once
the video scenes are shot, the editing process could be facilitated by using the
preliminary scene descriptions previously outlined by the author. Edited shots
could then be transcribed with their precise times in the scene. When the edited
video is captured, the scenes could then point to the digital movies themselves.
Video scenes could be placed (drag-and-dropped) within specific narrative
sequences. The tool would permit manipulation of the time attributes of the
video scenes and their spatial positioning, using direct manipulation techniques
on the video itself. Visually creating and manipulating hotspots over the
relevant (moving) objects in the scene would generate the necessary position and
time coordinates. A 3-D authoring and navigation tool for hypervideo, could be
envisioned for multidimensional hypervideo information spaces.
2.0 AUTHORING TOOLS
Digital Renaissance, T.A.G, and Veon (formerly Ephyx) V-Active have developed
new interactive Web video tools that provide clicking in the middle of a video
clip to jump to another video segment, piece of text, audio clip, picture, or
Web page. Interactive video has real appeal for education, advertising banners,
news delivery, and electronic commerce. A QuickTime VR movie's hotspots can be
used to to trigger any sort of web action. Hotspots can link to other pages,
other movies, sounds, images - any kind of web media. Once you have a QTVR movie
with hotspots you can assign url's to the hotspots, by adding the HOTSPOT
parameter. For web sites using frames, an optional TARGET parameter can be
included for each hotspot.
Potential applications for hypervideo include
applicability to a wide range of experimental expression, film
conceptualization, training, simulation, and electronic news delivery. It is
clear that filmmakers, educators, corporate trainers, and media artists can
utilize the structural and navigational scheme of hypervideo in their work.
Broader research issues include 3-D representations of hypervideo. Video
Software Laboratory is evaluating video object segmentation and object
recognition tracking technology to automatically track designated objects
throughout a series of consecutive frames. This process will embed time-driven
hotspot events into the digital video. These hotspots can then be linked to a
variety of media including URL addresses, other video/sound files, pop up text
boxes, and specified frames in the same video. The hotspot itself can be
highlighted, or an image can be superimposed over the hotspot, giving users a
clear indication of exactly where to click.
A 3-D authoring and navigation tool
for hypervideo, could be envisioned for multidimensional hypervideo information
spaces. Video Software Laboratory is teaming with several developers to embed
streaming video in 3-D Web pages with interactive hyperlinked video hotspots. An
authoring process to create hypervideo, called Hypervideo in now in Alpha
testing. Other research into hypervideo is ongoing at InterACT and
Hypervideo-home.
3.0 VIDEO URL LINKING ISSUES
Spatial Linking: Creating an Image Map File for a Video Image Map Fill are
text files that have HTML-like tags, specifying the location and action of each
image map. All tags shown below are required unless otherwise stated. No
negative values for numbers are allowed. Merging an Image Map file with a Video
Image in Realvideo: When you have finished creating the Image Map file, you
merge the Image Map text file with a .rm file using the rmmerge tool that is
installed with the RealVideo Encoder. Then, you merge that .rm file with an
encoded audio (.ra) or video (.rm) file.
Temporal Linking using SMIL
[Synchronized Multimedia Integration Language] SMIL is a simple but powerful XML
compliant, markup language that coordinates when and how multimedia files play.
A SMIL file (file extension .smi) can be created with any text editor or word
processor than saves output as plain text with line breaks. An example of a
simple SMIL file that lists multiple media files played in sequence is shown
below. RealSystems supports SMIL and allows new data types to be added. SMIL
support time controlled URL jumping.
4.0 Hypervideo through JavaScript, AJAX, and JQuery
Linked Video Ads gives advertisers the impact and flexibility to connect brands
with consumers through ad placement using Flash, JavaScript, AJAX