Binary Format for Scenes

from Wikipedia, the free encyclopedia

Binary Format for Scenes ( BIFS ) is a description language based on VRML 97 and standardized in MPEG-4 part 11 ( ISO / IEC 14496-11 , "Scene description and application engine") for two- and three-dimensional multimedia audiovisual interactive content. It is coded in binary .

Interactive content, scene concept and object-based coding

While MPEG-1 and MPEG-2 were still the standards for encoding audio and video data only, MPEG-4 was planned from the start as a tool for encoding and transmitting (not necessarily) interactive audiovisual content . When MPEG-4 is mentioned in the following, the scene description aspect is always meant, i.e. H. The audiovisual content is described as a so-called scene, which can consist of natural (e.g. video sequences or recorded audio tracks) and synthetic objects (e.g. 2D and 3D graphics). In addition, the concept of object-based coding is used for the first time in MPEG-4, in which the individual objects in a scene (e.g. background, actors, 3D and 2D objects, speech, background music, etc.) are coded and transmitted separately . The scene is then reassembled during playback.

The advantages of object-based coding and transmission:

  • It enables efficient transmission and storage of multimedia content.
For example, it is sufficient to transfer the background of the scene as an image once (as long as the camera is not moved).
  • A suitable codec from one of the MPEG standards can be used for each object (e.g. voice codec, still image codec, video codec, etc.).
  • Individual objects can easily be reused in the production of other scenes.
  • The scene can be adapted in the playback device to its capabilities.
  • Interactive content is possible.
The viewer can move freely or restrictedly in three-dimensional scenes. Actions (e.g. starting a video and / or audio clip, forwarding to a web shop where you can get the item in question, etc.) are also conceivable when clicking on an object in the scene.

terminology

The terminology was adopted from VRML . BIFS scenes have a scene graph , a hierarchical data structure whose individual elements are called nodes . Properties of the nodes are described in fields . Data types and value ranges are defined for fields .

There are nodes for visible objects (e.g. rectangles, cylinders) as well as for defining their properties such as color, texture or position. There are also nodes for positioning sound sources in the scene and for audio signal processing (see below under "AudioBIFS" ).

Applied Technologies

In order to seamlessly integrate the actors in a scene, the MPEG-4 video standard (ISO / IEC 14496-2) enables so-called shaped video objects to be encoded for the first time . A binary or grayscale mask is encoded in the video data stream together with the actual image content, so that when the scene is put together, the background is visible at the point where there is no actor. Blue screen technology is used to generate such video objects . Shaped video objects are integrated into the BIFS scene as textures .

In order to be able to integrate three-dimensional objects (e.g. objects modeled in 3D programs) into the scene, BIFS references the entire VRML standard so that all possibilities that are given with VRML can also be used in BIFS scenes.

MPEG-J defines a Java interface that allows access to the objects in the scene. This enables complete interactive applications (e.g. travel guides) with Java and MPEG-4. The bytecode of the application (also known as “MPEGlet”) is transmitted or saved together with the scene.

With AudioBIFS the part of BIFS is called, enables a complete audio signal processing. This makes it possible to position sound sources in a virtual room, apply effects to audio data or even describe rooms acoustically. A separate language SAOL ( structured audio orchestra language ) is standardized in MPEG-4 Audio to describe the effects to be used , with the help of which all conceivable signal processing operations can be described. For a realistic simulation of acoustic environments, objects in the scene can be provided with acoustic properties ( frequency-dependent transmission and reflection ) so that these can be taken into account during playback ( virtual acoustics ).

Web links