Home Analysis V-Nova: The strategy behind PERSEUS 2 – and why there’s still room...

V-Nova: The strategy behind PERSEUS 2 – and why there’s still room for lots more compression efficiency

Share on

At IBC, V-Nova was putting its next-generation PERSEUS 2 technology, first demonstrated at NAB earlier this year, through its paces. Guido Meardi, CEO and Co-Founder of V-Nova, explained that although the new platform was obviously capable of delivering significantly better quality at the same bit-rates than its predecessor, this had not been the only focus of the platform’s recent evolution.

“We know that in today’s world, people want to stream more and more channels. We needed to lower the cost-points,” said Meardi. While PERSEUS was fast, V-Nova knew it could make PERSEUS 2 even faster, and that translated into several benefits. “The first one is very, very low cost of encoding, low processing-power utilisation, low electricity [demand], and high densities. […] Yes, it encodes Ultra HD p60 broadcast quality at 10Mbps, but the most important thing is, it costs so much less.”

The second benefit PERSEUS 2’s speed enabled was making it easier to use on all platforms, noted Meardi. In initial deployments, PERSEUS “was working very well in closed environments like an IPTV deployment or in app deployments.” But V-Nova realised that video delivery was moving to a more fragmented environment, where “HTML-5 was taking over. […] So with PERSEUS 2 we created and we are presenting at IBC, the possibility to decode PERSEUS within a browser, within all the HTML-5 ecosystem, using simple JavaScript decoding.”

Meanwhile, PERSEUS’s ability to work in two different modes, as a standalone video compression system [PERSEUS Pro] or piggy-backing on existing codecs to improve their performance [PERSEUS Plus], was expanded. Where PERSEUS Plus previously worked in tandem with H.264, at IBC the next-generation platform was shown also supplementing HEVC. “In general, PERSEUS 2 is amenable to any type of base encoding – so when AV1 becomes available, […] it will work also with AV1. It [also] works on top of VP9,” said Meardi. This means, he claimed, that for operators, “PERSEUS is actually the cheapest and simplest option to immediately achieve massive benefits – because it’s compatible with the whole eco-system, it’s a simple software upgrade and doesn’t change anything.”

Finally, V-Nova has optimised PERSEUS 2’s performance around specific operating points, to reflect the common bottlenecks in today’s video distribution networks. “We’ve understood those limits over the past two years and we optimised PERSEUS specifically for those operating points. That’s why when we say we do OTT 720p HD at 300 or 400KBps, that’s not just a number picked at random. […] If you want to do mobile video, very often the majority of people have a maximum 300, 400 or 500Kbps.” Meardi notes that even in central London in peak hours, “very often you’re below 600KBps even with 4G.”

But isn’t there a theoretical limit to how much V-Nova can continue to optimise PERSEUS’s video compression performance?

Meardi pointed to the video compression tools built into PERSEUS 2, which were augmented in February by the acquisition of the global patent portfolio of video imaging experts Faroudja Enterprises Inc. “We haven’t yet used all those tools, so we know perfectly well that we have double-digit improvements coming soon in a lot of operating points,” he claimed.

Meardi believes that talk of PERSEUS getting close to ‘theoretical limits’ is bound up with V-Nova’s historical unwillingness to engage with conventional, laboratory-based, video compression testing protocols. These are characterised by allowing encoders to carry out repeated ‘passes’ when processing a test video segment or by giving them unlimited time or unlimited power to perfect the perceived quality of the result, he charges.

“If you have unlimited time to compress and you don’t constrain the bit-rate when it’s a tough scene, you just have a spike in bit-rate – that’s no problem. It would compress very well, and the quality would be nice,” commented Meardi. But the resulting compressed video segment, spikes and all, would be impractical to actually transmit, he argued, “because [the bit-rates spikes] would overflow the buffers in the transmission system.”

PERSEUS, by contrast, “was built for real stuff. In our claims, we have always said, ‘in real-world conditions, it delivers two times, three times, the benefits, etc.’, not ‘in theoretical, unlimited, unconstrained, give-it-as-much-power-as-it-requires [conditions],” Meardi declared. With reference to theoretical testing, Meardi pointed out that “if we want to really serve the eco-system, we need to achieve great compression with a mobile phone, with a single chip.”

Meardi stressed that V-Nova’s mission has “always been that we want to offer next-generation performance, available everywhere at low cost, and low processing-power. If people have unlimited, unconstrained processing-power, yes, maybe they can equal our performance or even better,” he conceded. In which case, “maybe they belong to the part of the sector that will not adopt PERSEUS – we’re fine about that,” he concluded.

The PERSEUS 2 codec was awarded ‘Best Digital Video Processing Technology’ at this year’s CSI Awards at IBC2017.

Photo: Guido Meardi

Share on