BLOG: IP Video CODECs - how many is enough?

8th September 2015 BLOG: IP Video CODECs - how many is enough? by Tim Claman, Quantel and Snell CTO

At IBC this year, I expect many of us will experience a slightly unsettling sensation as we see the industry progress further away from the nicely constrained, homogeneous world of SDI, towards the asynchronous, heterogeneous universe of IP. Nowhere is IP's technological diversity more prevalent than in the realm of video compression formats.

Just as we are all beginning to get comfortable with exchanging packetized, uncompressed SDI over networks via SMPTE 2022-6, a vast array of video codecs is vying for relevance in the emerging IP landscape. The candidates include JPEG2000, H.264/AVC, VC-2, Sony IP Production Format, H.265/HEVC, and others. Who knows what new codec variants may decide to descend upon Amsterdam next week to join the fray?

A cynical observer might say this is yet another case of vendors vying for competitive advantage by promulgating their own proprietary technologies. We have all seen this movie before, many times. (File-based ENG cameras spring to mind.) Why not take this opportunity to converge on a single, universal codec?

To the contrary, a pragmatic observer might assert that different codecs are sensible for different applications. Light compression makes sense within a facility, where network bandwidth is relatively inexpensive. But a more complex codec makes more sense where the cost of the pipe itself is prohibitive when sending content streams across long distances, for example.

I think both perspectives are partly correct. As much as I find the utopian vision of a single-codec future incredibly appealing, it's clear to me that different applications require trade-offs between computational complexity, processing latency, and network bandwidth. So a one-size-fits-all codec is probably not in the cards. That said, I think we would all benefit if the industry could converge on a small number of open codec technologies. Fewer codec variants would mean less transcoding, less image degradation, lower complexity, and better interoperability all of which would contribute to lower operational costs and higher efficiency.

Industry groups like the Video Services Forum (VSF) are taking this issue head on. As members of the VSF, Snell and Quantel support their efforts to harness the power of end users and technology vendors coming together to evaluate technology choices, make recommendations, and codify best practices that help to align and inform the industry as a whole.

At Snell and Quantel, we don't have a proprietary codec that we are trying to foist on the industry. But we do have some informed opinions. We believe that light compression makes sense for signals within a facility, and that the ideal codec for this application would be computationally inexpensive, low in latency, visually lossless, and extensible to formats like UHDTV 4K & 8K. We also believe in codecs that are unencumbered by onerous licenses and royalty fees that typically get passed on to the end user. And we believe that open standards are essential to increase the velocity of industry advancement.

With all of these factors in mind, and having looked very closely at all the codecs available, we see SMPTE VC-2 as the ideal choice for light video compression within a facility. In the spirit of convergence, we encourage the industry to evaluate and adopt this proven, lightweight, open, SMPTE standardized, royalty-free codec.

In parallel, we are also hard at work implementing other codecs for interoperability, based on customer demand. In the end, while it's easy to dream about a single-codec future, we recognize that the newfound agility we are all striving for in the transition to IP can only be attained if we embrace the diversity that comes with heterogeneity. Even if that means embracing more than one (but hopefully not too many!) video codecs.

LINK:	http://www.snellgroup.com/news-and-events/press-releases/4092/blog-ip-...
	See more stories from snell

Most recent headlines