WP 3 : Data and Metadata acquisition

Data and Metadata acquisition is made of one platform and 2 devices each designed to load and translate the USE into GUI, guidance and process control for the content production: Semantic camcorder 2.0, Camera Mate and Ingest platform.

Ingest platform presentation

Functional Description

The Ingest Tool is a software running on PC that allows producing audiovisual elements, as described by the USE. It allows searching, and acquiring Audiovisual (AV) elements even from an archiving site, a web2.0 site or from an optical block as included in Skamera. Ingest is a part of the tools connected to the Unique Semantic Bus. It is used to complete a USE. The acquired content can be either “life” produced with a Camera2.0, or documentary elements found on an Audiovisual archiving system.

Technical work

The development of this platform consists in creating a “rich client” communicating with servers. It is essentially software development in a standard environment like windows or Linux and running on PCs.

 

Semantic Camcorder 2.0 & Camera Mate

Functional Description

These devices are portable nomadic AV production tools of which the principal difference is that the optical block is built-in or not. This acts principally on their housing and design. In terms of features: all basic functions are the same. Only some advanced functions where the optical block is piloted automatically by the Processing Unit are dedicated to the camcorder.

Both systems are made to download first the empty USE through a build-in wired or wireless interface. Then both products include a complete set of functions allowing to edit script and modify structure of clip, shoot and acquire audio and associated metadata, post produce including, titling, subtitling, dubbing, adding transition effects, etc. The result is a filled in USE (almost clip or part of clip) could be then uploaded to archiving server or transcoded as a mounted clip for local streaming or DVD burning.

The capture of video uses the ingest function presented above.

The camcorder and Mate also include a powerful ergonomy using large screen and enhanced interactivity making process simple for an amateur but giving him all assistance to work as a professional and obtain the best result. As described in the USE depending on ontology adapted to the content and interpreted by the system, the guide acts as Screen Play Assistant, Director’s assistant, audio engineer assistant, etc.

To realize these functions, two main categories of algorithms will be integrated to the skamera and skamera mate 2.0:

  • the actual pre-indexing algorithms allowing the labelling of video sequences with metadata perceived on the environment in close vicinity, and coming from embedded sensors like accelerometers, GPS, etc.
  • the “video syntax and grammar verification” ensuring that the user has correctly followed the indications given by the “video form” (for example zoom in or out recommendations are effectively followed, etc.), and linking them to the sensors data.

The camcorder will also include new features of optical block control like realization of automatic effects initialized by AV semantic: framing, zooming, panning … These will be managed using high level algorithm analyzing the pictures and directly controlling the optical block.

Technical work

The project consists of 5 main parts:

  • Ergonomic study and validation
  • Study and realization of the electronic Central Unit Board based on Da Vinci (bi-Processor architecture ) of Texas Instruments
  • Development of the application firmware (ARM processor, OS Linux)
  • Development of the algorithm and compression tools (DSP – OS TI)
  • Design and prototyping of housing for both camcorder2.0 and Mate

After developing each of the elements proposed above, obtaining a working prototype that will be used on real cases requires the integration of bricks and their improvement.