Your browser is not supported. Please update it.

8 September 2022

Codec experts clash over scope and scale of VVC opportunity

The volume of live audience interaction during Faultline’s latest webinar justified our captioning of the event – “There is never a dull day in the world of video codecs.”

Aside from some prickly questions around patent royalties, another highlight of the panel came when V-Nova’s Fabio Murra, SVP of Product and Marketing, disagreed with MediaKind’s Carl Furgusson, Head of Portfolio Development, on the opportunities ahead for VVC (H.266).

This came about after Faultline observed how there doesn’t seem to be much love for VVC from the market right now, excluding some niche 8K UHD use cases often from broadcasters flexing muscles during major live sporting events.

Fergusson mentioned that the one place likely to see adoption of VVC is in 8K broadcast, which ruffled a few of Murra’s feathers, given V-Nova’s first-hand experience of VVC being standardized in Brazil’s SBTVD 3.0 DTT system for both 4K and 8K, with LCEVC as an enhancement layer.

“We expect some traction for VVC to come in Brazil first. It will take a few years, but it will be a bit larger than perhaps Carl predicts,” observed Murra. “Brazil and the Latin American countries are leading the way.”

Of course, with MediaKind betting its best R&D dollars on VVC being the “pinnacle of codecs to aim for,” Fergusson would make such a statement on VVC, while knowing that MediaKind could roll it back to EVC if there proves to be any future demand.

The scale of MediaKind commitment to VVC means the former Ericsson Media Solutions business has spent a grand total of zero R&D dollars on other codecs, whether AV1, EVC, or LCEVC, as MediaKind is fundamentally waiting for market demand.

“Netflix and others, to some degree, build their own technology, so there is no opportunity to sell encoders,” added Fergusson.

Another contentious comment from Murra came regarding EVC. While other panelists hastened to agree or disagree with the recent comments of Facebook’s Direct of Video Coding, David Ronca (see separate story in this issue), Murra contended, “maybe EVC is a bit pointless.”

Murra went on to suggest that EVC was forged from the fires of MPEG by the powers that be to ensure that others behaved when it came to VVC royalties. As we said, there is never a dull day in the world of video codecs.

Synamedia’s Jan de Cock, Director of Codec Development, on the other hand, described EVC as having a “noble purpose” while admitting that EVC is not a standard that Synamedia is investing a lot in – at least not until EVC can prove that it fills a use case that AV1, VVC or HEVC does not.

“We are heavily involved in VVC and see a big future for AV1, which is more targeting VoD OTT type scenarios, while VVC support is being heavily pushed or pulled by the broadcast world,” commented de Cock.

That said, companies including Synamedia are also showing that HEVC works really well for 8K, streaming to 3 or 4-year-old TV sets which are perfectly capable of handling HEVC 8K.

“AV1 is going to help push down the bitrate a bit, but VVC will help push it down significantly,” added de Cock.

With Apple still stalling on AV1 support, MediaKind’s Fergusson made an almost desperate remark that it would take a meteoric move – something of Apple’s size – to make a sweeping commitment to VVC, which would change the streaming landscape forever. Not only would it penetrate Apple’s ecosystem, but Android chipset makers would soon follow suit.

Faultline could not sit down with this panel of codec experts without broaching the subject of sustainability in video streaming. This has stemmed from people in the CDN space throwing shade over seemingly major strides in reducing power consumption from the encoding camp, when often the watts are just being pumped further down the video delivery pipe, for consumption elsewhere.

De Cock argues that power consumption is going down every year on a per-channel basis, by about 30% a year, according to some unspecified data. Improvements have come through a combination of software optimizations and also more powerful CPUs. Synamedia is also working to become more eco-friendly on the CDN side too, supported by the recent acquisition of Quortex, bringing just-in-time operations that help to reduce CO2 emissions and power usage, as well as reducing costs.

Meanwhile, V-Nova’s Murra has a research paper hot off the press showing LCEVC’s battery efficiency gains when used with SVT-AV1. “We wanted to show there are trade-offs between computational requirements of a codec and performance. AV1 is more computationally intensive, but more efficient on the transport side, so the CDN, caching, storage all benefit. It’s a trade-off that is much better with LCEVC – with less computational power and bitrate savings with LCEVC on top of SVT-AV1,” Murra explains.

Murra urged service providers to look at how they use a codec in terms of sustainability. Whether using their own infrastructure, using HW encoding, targeting a few set tops, or targeting millions of devices, and then draw conclusions by testing the available codec technologies out there on the market.

An important point was made for overlooking the decode side, because we’re often talking about hundreds of watts on the encoding side. “Every device now is a decoding device, so efficiencies count. Obviously, this benefits users too if they can go on longer watching video without having to recharge,” added Murra.

Meanwhile, Fergusson isn’t onboard with claims like 10x compute power for each generation of codec, as this assumes you have taken all the available toolsets of a standard to, in theory, achieve the maximum possible bitrate savings. That is a maybe truer for file-based encodes, admits Fergusson, but with live encoding, you have to work with a finite amount of processing power.

“If an on-prem device is sitting on an appliance-type server, there are only so many cores in that server. In the cloud, there is a limit to what node sizes you can take in the cloud. When doing live encoding, there is a trade-off between what tools we use and how much bitrate savings are made. Less tools = less compute power,” explained Fergusson.

Finally, a word from our wildcard on the panel, Janne Pelkonen, CTO at IdeaNova, which brings encoding expertise and security services to the avionics and telehealth industries, as well as media and entertainment.

Sustainability is an interesting area for IdeaNova. A lot of use cases in the avionics world involve VoD, as you’d expect, so IdeaNova is encoding a relatively small amount of titles which are not necessarily the dominant cost in this context – let alone the dominant source of CO2 emissions.

“Decoder efficiency is a big factor. Regarding AV1, software decoders are pretty good these days and hardware is coming shortly in Android devices, so that should help the sustainability factor,” said Pelkonen.

You can register to view the recording of Faultline’s Video Codecs webinar panel in its entirety here.