It took four years of clandestine development, but Silicon Valley, Calif. start-up Pulsent Corp. Monday unveiled a breakthrough in
video compression that will enable businesses to stream broadcast-quality video over broadband networks at 1.1 megabytes-per-second
(Mbps), adding a crucial piece to the digital media distribution puzzle.
Pulsent claims its technology provides a 400-percent improvement in bandwidth and storage efficiency over current video compression
schemes such as MPEG-2, which transmits video at roughly 3 Mbps. The announcement could percolate interest among telecommunications and
service providers, which have been looking for ways to pique customer interest. It may also serve notice to Microsoft Corp. and RealNetworks Inc.
, who are makers of, and leaders in, proprietary video compression software.
What will telcos and ISPs be able to do with this technology? Deliver broadcast video over ADSL, which will quadruple the amount of
programming that can be stored on personal video recorders (PVRs), thereby increasing cable or satellite channel capacity. It should
also help with video-on-demand (VOD) and high-definition television (HDTV) adoption.
Pulsent Co-Founder and Chief Executive Officer Adityo Prakash noted simply that current methods of video compression have “run out
of gas.”
“20+ year-old block-based technologies like MPEG and its proprietary offshoots just can’t be improved significantly,” Prakash said.
“The industry has long needed a fundamental rethinking of how video is processed and delivered to enable the next generation of
services and applications. That is what we’ve achieved.”
So, what has Pulsent done differently? MPEG, which continues to evolve and is now facing uncertainty over the licensing schemes for
its Internet-charged standard MPEG-4, currently uses “block-based” approach to compressing digital video information. This divides a
video frame into arbitrary blocks and removes redundancies by attempting to match and reuse blocks from previous frames for the
current frame. Pulsent contends that despite improvements to motion compensation and residue coding, the evolution of MPEG has
reached its ceiling.
Pulsent proposes to evade such limitations by identifying the true structural elements or “objects,” in any video scene and modeling
their motion. Because these objects are natural, elastic constituent components of any image, they can directly correspond to parts
of real-world objects. Once identified, Pulsent contends the frame-to-frame motion of these “objects” can be far more accurately
modeled than with block-based approaches.
Pulsent’s frame-to-motion approach is patent pending, but the start-up stressed that its video payloads were designed to be
transmitted via the MPEG standard. Moreover, Pulsent products will support handling of other codecs such as MPEG, ensuring
compatibility with existing applications. The company also anticipates to be able to pass video between MPEG and Pulsent formats.
