With more people streaming video content over the internet than ever, especially during the pandemic, data centers are working in overdrive to support demand.
However, with fierce competition between streaming rivals placing speed and video quality at the top of the priority list, few are considering the potential impact on the environment.
To find out more, TechRadar Pro spoke to Guido Meardi, CEO at UK-based V-Nova, a company that claims to have developed a solution that will dramatically enhance compression formats, to hear more about the delicate balance required to satisfy the call for improved video quality and the need to build a sustainable future.
How does your LCEVC standard relate to MPEG-5 EVC? Is it a proprietary subset of the standard?
LCEVC (or MPEG-5 part 2) is a standalone standard by MPEG, distinct from EVC. It differs from other approaches by being a “low complexity enhancement” (the first three key letters of the acronym Low Complexity Enhancement Video Coding) for any of the existing or future video compression technologies.
Interestingly, it can enhance MPEG standards including AVC, HEVC, VVC and EVC and other technologies such as Google’s VP9 and AOM’s AV1. In contrast, EVC or “Essential Video Coding” is a fully self-contained video codec that can be enhanced by LCEVC.
Is V-Nova powering Amazon Prime?
V-Nova is collaborating with Amazon Web Services (AWS) on a number of different projects and with different technologies. V-Nova’s SMPTE VC-6 compression standard [is used] to power live remote-production workflows for broadcasters much more cost-effectively, removing the need for very expensive leased connections. We also provide an AWS-enabled LCEVC transcoding platform for file-based and live video streaming.
What does a small outfit like V-Nova offer that the existing giants (e.g. Facebook, Microsoft, Amazon) don’t?
The companies mentioned are all potential users of V-Nova technologies and specifically our highly optimized implementation of the new MPEG-5 LCEVC standard.
By deploying LCEVC they can substantially enhance their video delivery to consumers on a broad range of devices, providing higher quality at up to 40% lower bitrates while significantly reducing the energy consumption of video delivery at the same time. We currently have a number of major trials and early-stage deployments underway with streaming giants, broadcasters and video service providers in different markets such as sports, news, gaming, real-time communication and social.
Since the early days of media encoding (MPEG-1), how far have we moved down that rabbit hole? How good are the codecs compared to the original 1993 ones?
The progress with digital video compression has been steady and impressive ever since the early days of MPEG-1. In general each major generation of codecs has doubled, or come close to doubling, the possible compression ratio (or in other words halving the bitrate required). For example, MPEG-4 AVC/H.264 reduced bitrates by about half compared to MPEG-2, HEVC/H.265 achieved similar gains over AVC/H.264 and the very latest codecs like VVC and AV1 are once again making gains typically of 30-40% over HEVC.
The downside though as mentioned before is that each of these generational gains has come at the cost of massive increases in processing complexity and compute/energy costs per pixel (typically around 10x), with newer displays calling for an ever-greater number of pixels.
Do you see a hard upper limit to the compression ratio? How are current hardware innovations helping?
The extent to which each generation is improving things appears to be gradually decreasing but there is undoubtedly more to come, especially since the goalposts of what needs to be compressed are moving. The latest processors and AI continue to help, but Moore’s Law is quite simply not enough to keep pace with the skyrocketing energy consumption driven by the “perfect storm” of more complex codecs, a growing number of services and higher video resolutions. This is where we see MPEG-5 LCEVC contributing significantly to make the use of high resolutions and the latest codecs more feasible and sustainable.
Compression might be seen as a brute force approach. Are there any alternatives that have been explored?
There really isn’t any alternative to compression. As the resolution and quality of video that viewers expect has increased over time, it has demanded ongoing improvements in compression to ensure that it can be effectively and reliably delivered to our devices over the available bandwidth we have wherever we are. From entertainment services, video conferencing solutions and virtual reality to military surveillance drones and remote medicine, compression is absolutely core to any good video experience.