180 likes | 309 Vues
In March 1999, the Broadcast.GL project was presented to enhance OpenGL for broadcast-enabled receivers. This specification aimed to create a conformant profile and cross-platform API to facilitate the integration of synchronized media, enabling new interactive broadcasting capabilities. The project includes contributions from key industry players like ATI, NVIDIA, and Sony, focusing on addressing the growing needs of digital set-top boxes and internet devices. With monthly meetings and a timeline set for public specification by May 1, the project emphasizes high-performance media, compositing, and support for multiple surface types.
E N D
“Broadcast GL” Presentation to OpenGL ARB March 1999 Rob Glidden robg@quadramix.com
Outline • Overview • Status • Relation to ARB
What is BGL? • “Conformant profile/extension of OpenGL for broadcast-enabled receivers” • Launched Jan 1999 • Goal of public specification by May 1
ATI Broadcom 3Dlabs Metabyte NVIDIA Philips PLATINUM Quadramix Reader LLC SONY Stellar Semiconductor Wind River Participants
Why BGL? • New devices • Digital set tops • Internet devices • Consumer media “appliances” • New capabilities • Rapidly expanding graphics • High-bandwidth communication interfaces • New standards • “Interactive broadcast services” • ATSC/DASE, AICI, DVB, MHEG, W3C, Web3D, ATVEF, MPEG4 Cross platform “media API” needed
Schedule • Launched Jan 1999, monthly meetings • Feb 1 Requirements, data evaluation • Mar 1 1st internal draft • Apr 1 Public review draft • May 1 Proposal 1.0
Requirements • 1 Compositor • Cross-platform API for display of synchronous integrated media • 2 Surfaces • Video, 2D, 3D, text, and vector content to be composited • 3 Build on existing 2D, video APIs • 4 3D Profile • Profile/adaptation of OpenGL for broadcast environments
1.0 Compositor • 1.1 Compositing of multiple surfaces • 1.2 Run-time performance profiling for scaleable content • 1.3 Performance optimizations must not require content awareness
2.0 Surfaces • 2.1 Support multiple rendering contexts • 2.2 Support synchronized composition • 2.3 Support domain-specific rendering methodologies • 2.4 Multiple surface composition • 2.5 Surfaces as buffers
3.0 Build on existing APIs • 3.1 Enable transition from existing video, 2D, and vector/font rendering APIs • 3.2 Support animated images
4.0 Profiling/Adapting OpenGL • 4.1 Operate in embedded contexts • 4.2 Operate in footprint-constrained contexts • 4.3 Support range of rendering architectures • 4.4 Enable high-performance, feature-rich media • 4.5 Conformant
“Profiling” • Tried application-based profiles • Profiled OpenGL calls • GL games, VRML browsers • Did not work • No common architecture theme • Considering “architecture” profiles • Color formats, display lists, buffers • Device constraints are moving target • 32 meg designs appearing
“Surfacing API” • Two proposals • “OpenGL agnostic” • Surfaces are not OpenGL buffers • Could be implemented without OpenGL • “OpenGL centric” • Uses OpenGL buffers • Requires OpenGL • Currently merging proposals • Leaning to “OpenGL agnostic” • Contacts with UGL, SVG, MHEG, others
Issue: Long-Term Home • No group decision yet • Time-to-market, IP concerns expressed • Options • 3D Profile in ARB? • Assumed continued OpenGL licensing • Surfacing API • ARB, or other standards organization? • Possible IP-free or open source? • Need to avoid adding IP to existing 2D APIs • Depends on OpenGL “centric” v. “agnostic”? • Other options?
More Info • Rob Glidden - robg@quadramix.com • Andy Fischer - afischer@metabyte.com