Applications

Video Conferencing Trends

The new generation of high-definition video conferencing is becoming a must for fast moving companies. At the same time, the cost of these systems needs to come down to allow a broad market adoption.

Video conferencing deploys many levels of image processing, starting with image sensor pipeline processing, video/audio encode and decode and graphics processing. Examples of advanced processing include 360 degree panoramic views with head-size equalization, and codecs with improved error resiliency.

DSPs, and to some degree in combination with FPGAs, has been the traditional approach to implement these functions, but with accelerating performance demands there's a need for a new class of solution.

SPI's Technology Benefits

SPI's stream processors bridge the performance gap between DSP and FPGAs while a standard C programming environment improves productivity, shortens time to market and preserves the software investment.

A number of compute-intensive tasks can run on Storm-1:

• Image sensor pipeline
  
  
• Interaction & experience; ambient light compensation, 360 degree stitching, head-size equalization, advanced 4K upscaling, speaker detection, gesture recognition
  
  
• Video pre-processing and compression; Motion-compensated noise filtering, 720p60 / 1080p60 H.264/RTVideo compression, scalable video codec (SVC)
  
  
• Display processing; 3D UI, source changes, scaling, OSD
  
  
• Audio beamforming, echo cancellation, sound direction detection

A single Storm-1 SP16HP-G220 can replace several traditional DSPs, allowing for 720p encode and decode, sensor processing, scaling and OSD.