Zenverge has leveraged the company’s extensive expertise in algorithms and ASIC design to create a revolutionary, high-performance, low-latency video pipeline capable of transcoding and decoding
HD content up to 4 times faster than real time or lower resolution content up to 40 times faster. In addition, the Zenverge solution can simultaneously transcode up to 4 independent HD streams or up to
16 independent SD streams along with supporting full encode of HD content. By designing the video features as a single pipeline, the architecture can scale with process technology. In the future,
it will be possible to process HD even faster, process more streams of HD, and at much higher resolutions (e.g. future TV formats such as 4Kx2K used by digital cinema today and planned for broadcast in Japan
starting in 2011).
Zenverge has filed 21 patents related to this architecture to date with many more in the pipe.
Over the years, digital media IC designers have attempted to solve problems such as multi-format decode by building large media processors that employ separate pipelines and memory subsystems for
each supported format. As the number of video and audio formats and the requirements for multiple channels of HD have grown along with the need for transcoding, designers of these ICs have added a
high degree of functional duplication that requires
large IC footprints and larger amounts of memory. Moreover, these media processors need high levels of power. To overcome these obstacles, engineers have to find a new way to solve this explosion in
design complexity and to support performance levels capable of driving multiple HD streams.
To overcome these obstacles, Zenverge has created highly optimized function units that take advantage of the commonalities across the various formats, as well as across ecnoding and decoding. For example, all of the popular video formats,
including MPEG-2 and H.264, use block-based motion compensation. The Zenverge video pipeline has a single motion compensation unit that works across all formats. This very efficient approach to multi-format
video processing allows Zenverge to make a smaller, cheaper, and lower-power IC. In addition, the Zenverge architecture uses more than 12 distinct techniques to accelerate performance. Below are some of the fundamental problems that must be
solved for 4x HD performance:
1. Efficient multi-format support
2. H.264 CABAC acceleration
3. Optimized memory bandwidth
4. Reuse of compressed-domain information
For more details, please refer to our white paper "The Need for Smart, High-Performance Transcoding"
in the White Paper section of this web site. Please note that access to the full article requires secure customer login which is available to qualified recepients by contacting Zenverge.
Technology Overview
Behold the power of ZEN.
The world's highest performance digital media chip architecture.
ZEN (Zenverge Entertainment Nexus) was built from the ground up to meet the soaring performance requirements of the next generation of digital media applications leveraging patent pending, innovative ideas from a
team of industry veterans and using the latest advances in ASIC design. ZEN is our proprietary media transformation and delivery technology designed to address the most demanding requirements of a fast growing digital
media convergence market with a single, flexible and enduring architecture.
