Applications

Video Conferencing Trends

Video conferencing is transitioning to HD and beyond to enable life-like, efficient interaction. However, to allow for true broad market adoption of these "Telepresence" systems, total cost and bandwidth requirements need to come down.

Video conferencing systems deploy many levels of processing, including camera sensor, audio/video codecs, and various levels of video and display processing. There is also increased focus on new features, such as speaker detection, content-adaptive compression (e.g. faces), scalability, improved resiliency, 3D, and integration of mobile devices that is requiring an increasing degree of flexibility.

Current solutions commonly use a complicated integration of multiple DSPs, FPGAs, and ASSPs, and with development costs are getting out of hand, designers are forced to make difficult tradeoffs.

SPI's Technology Benefits

The Storm-1 platform simplifies the implementation and product management of  high-quality video conferencing enpoints and MCU systems. The unified C-programmable platform reduces development effort, enables code reuse across product lines, and slashes BOM cost compared to multi-DSP and FPGA based approaches. In addition, the software-based solution enables easy field upgrades and fixes without having to wait for respins of FPGA or ASSP hardware.

Storm-1 is able to simultaneously process duplex HD H.264 streams while providing audio processing, network connectitivity, display processing including scaling, overlays and compositing for multiple HD and UXGA screens, all in one single chip.

The scalable platform is suitable a range of video conferencing applications, from low-cost CIF endpoints, to HD telepresence systems and MCU transcoders.



HD Endpoint Example:
Single-chip 720p30 HD H.264 duplex system

Feature Highlights:

• Software programmability provides flexibility to support any codec, resolution with audio/video processing.
• Scalable platform, capable of multi-channel 1080p60 H.264 encode and decode
• Low-cost solution integrates preprocessing, codecs, scaling, compositing, alpha-blending overlays, and network processing
• Standards-based Linux platform and C code implementation
• Software-based codec assures interoperability and is tuned for low bitrate, low latency and error resiliency
• Flexibility to support combinations of multiple streams and framerates
• Intelligent Region-of-Interest based compression reduces bitrates
• Camera sensor Bayer RGB demosaicing and color processing
• Display processing with alpha-blending, overlays, scaling and Picture-In-Picture
• Processor headroom enables advanced processing such as audio beamforming, high-quality scaling, and content-adaptive compression

Contact SPI or your local reseller for more details on the video conferencing solutions, request a demonstration, applicable benchmarks and demonstrations.