When we last spoke to the Alliance for Open Media (AOMedia) just before Christmas, they assured us that everything was going along nicely and that their first license free codec would be out by March this year.
Just to be clear, the arrival of a license free codec would change the entire video landscape – taking money out of R&D among traditional codec developers – but at the same time, shifting video power into the hands of a number of chip and internet leaders, away from traditional CE businesses and studios.
Instead of a new codec every ten years, it would mean an update once or twice a year, to run on existing chips that did not need replacing – a practise far more in line with the pace of cloud developments.
It is against this background that we have to see the opportunities for V-Nova, which this week said it has a second generation of its Perseus codec, two years after launching the first.
To be clear this is just a next gen product, a little better compression, which works with existing HEVC set ups, HTML 5.0 browsers and this does not yet include the improvements which may come from its latest acquisition, of the technology behind Faroudja Enterprises, bought back in February. This will likely come as a pre-processing element to add to Perseus 2, which will create even larger compression gains.
The issue here is that tiny V-Nova has lined up against the power of the AOMedia leaders – Cisco, Microsoft, Google, Netflix, Amazon, Intel and a few others – and if that effort was ready to engage with the video community, it would make little sense to continue to invest in an independent codec company which plans to live on paid licenses.
But something has come off the wheels at AOMedia and two weeks of emailing them has not resulted in a statement. However, outcomes from AOMedia meetings are known to have pushed the arrival of this license free codec to at least the end of this year, which is when it plans to freeze its code base. Commercial implementations may take another year, leaving a 2 year window for HEVC to become established and for V-Nova to carve a space for itself.
The key claims for Perseus 2 are that it will allow video to be delivered in 100 kbps over mobile and that 300 kbps will be sufficient in more advanced mobile regions, and that 1 Mbps will lead to what it calls a monetizable mobile stream.
It also says that it can squeeze video into 2 Mbps for DSL customers and that UHD movies begin at around 4 Mbps and go up to 6 Mbps, with live sport scaling to 10 Mbps. V-Nova also claims that version 2 software libraries have also been streamlined to deliver a lighter and faster codec that reduces computational load – so longer battery life in portable devices.
When you see the V-Nova plan unfurling, we can see that it has initially targeted video contribution feeds at Sky Italia, then shifted to markets where either satellite is in short supply, or where 4G and even 3G networks are not yet installed. This represents the low hanging fruit in the sector, offering a trade-off between codecs and expensive capex.
It is hard to sell to an operator which is planning to leap into 5G, complete with 1 Gbps connections, that you can save a few Mbps for him. But offering video to a network which can only muster 100 Kbps, is a real capex saver.
But the codec environment is not just AOMedia, it is for the most part represented by two HEVC factions, one from MPEG LA which is backed by Apple, the BBC, ETRI, Fujitsu, Hitachi Maxell, Humax, JVC Kenwood,
Korean Broadcasting, Korea Telecom, NEC and NTT, Orange, Samsung, South Korea Telecom, Vidyo and a handful of university and research institutes
The key drivers for this are the handset leaders of Apple and Samsung, Korea and Japanese operators and their suppliers.
Lined up against them are more traditional codec patent holders within HEVC Advance in Dolby Labs, GE Video Compression, HFI Innovation – part of MediaTek, and Philips, Mitsubishi and Warner Brothers
The key name missing from both factions is Technicolor which left HEVC Advance last February and immediately claims that it signed a deal with one of the world’s largest device manufacturers – likely to be one of the big three handset makers Apple, Samsung or Huawei.
But even getting quality mobile video at HD levels in a monetizable form at 1 Mbps in 4G is a huge achievement – as these are typically 5 Mbps right now.
Other applications could be to extend the reach of DSL further distances from the exchange, something which could have huge benefits in the US, France and Spain in particular, which have huge chunks of their community which cannot get a broadband single fast enough to support online video.
We spoke in a pre-briefing to Fabio Murra, SVP Product and Marketing at V-Nova. “The new version gives better quality and even more savings. It can be downloaded to existing customers as an over the air update,” Murra told us.
V-Nova began life two years ago and at the press conference everyone asked out loud if this could really be possible and Faultline while certain that it could be done, was dubious that V-Nova had actually achieved this – mostly because it refused to say precisely how it had managed it.
It had already won a Sky Italia contribution feed contract. “Now that relationship has evolved,” said Murra, “and last April we moved onto its IPTV platform.” This is the one that feeds the Telecom Italia relationship with Sky content, including TIM mobiles, through a Harmonic head end, sending video at 4 Mbps.
We are aware that this service only goes to around 100,000 set top boxes in Italy and that it’s not intended to be the best quality on the market, but everyone in Sky will have their eyes on its performance, and there is a possibility that the codec could become standard throughout the organisation, although it has already made huge investments in an Elemental head end in the UK.
Since then Eutelsat, the satellite provider for Sky, has taken a shareholding in V-Nova in September last year, and plans to bring the technology to bear on its video distribution platform for UHD around Europe.
The only other known customer is the recently signed Thaicom, which operates a series of satellites over Thailand, which says it will also be an early user for UHD delivery.
Thaicom is a special case – to be fair right now it is considered a basket case – it dropped revenues for each of the last two years, last years by 12%, after closures of DTH customers and its usage of existing satellites fell to around 67%. It likely sees the use of a more efficient codec as something which can entirely change its fortunes, and rush new clients onto UHD. Most homes in the region however cannot use UHD signals, and are barely getting to grips with HD. Its four satellite are old and out of date, and it is rushing to put more satellites in the air.
“Perseus is neutral to delivery of multiple types of UHD and does not interfere with HDR delivery. UHD movies can be squeezed into 4 Mbps with Perseus, maybe 6 Mbps, and live sports into 9 to 10 Mbps,” said Murra.
“We’re not suggesting this is quite the same quality as DVR and Blu-ray which puts out full UHD at 120 Mbps, but the quality remains high and this can lead any operator to massive savings. A satellite operator could double its channel density for instance. And cloud operators can halve the size of their server farms, with savings in electricity, and storage.”
“Version 2 of Perseus will also mean that we can put the codec onto any HTML 5.0 platform, not just into controlled pay TV, but into any app and any website. We work with all appropriate charging models and obviously, it is a very different subscription price to use our technology on the web rather than for an OTT or pay TV platform.”
Murra agreed that the market had been slower than he had anticipated, “The market was slow because they did not believe it could be done. Now we have broader platform support this will not be so much of a problem. We have it working with the Imagine products, on Microsoft’s Azure cloud and on Windows and Amazon can run it on AWS. Devices which previously used JPEG 2000, can get the same quality with Perseus, but with costs 70% lower.” The market would have believed it more if the company had better explained its technology.
Mura was also was adamant that AOMedia does not represent competition. We don’t have to compete with AOMedia, it can run as a “turbocharger” for the first AOMedia codec, AVI.”
One of V-Nova’s first wins was an integration with Alcatel, as it was, into its cloud DVR platform. Has that relationship continued under Nokia’s ownership of Alcatel? The fact that a quote from Nokia was included in the press release, is encouraging, talking about looking forward to the additional advantages of Perseus 2.
“Operators were promised a 50% improvement in throughput with HEVC, and what most of them are seeing is closer to 20%,” claimed Murra.
“We don’t see AOMedia’s AV1 shipping into commercial deployments for 2 years,” said Murra, It has publicly admitted that it will not freeze the code base until the end of 2017. They are still internally debating which transforms they will use. That’s what we were doing in 2010.”
Perseus 2 will first be used in its native mode for Contribution, but for distribution platforms it needs to have an MPEG encapsulation or HEVC present to make it compatible with distribution as it is set up today.
“There needs to be a codec present with Perseus 2 it can be H.264, VP9, HEVC or AV1. Perseus is still the cheapest option because when you buy a video system, you still need a particular implementation. In HEVC and H.264 you still pay royalties, and for Perseus you pay less. In AOMedia’s AV1 you will pay implementation costs and you will still pay someone to look after it all.”
The company claims that it has been profitable, but with new investment it has accelerated its R&D and that means it has slipped back out of profit for a while.
“We see the market taking off before AV1 comes on the scene, and it will take off first in China and the Far East where everyone is more innovative in video,” he said. The company says it has many more customers than it has stated in public, but has signed non-disclosure agreements which mean it cannot reveal all of its partners.