Could the new Perseus codec usurp H.265/HEVC for future 4K distribution (and possibly even acquisition)?
V-NOVA, a UK-based technology company, announced its compression technology, Perseus, way back on April 1, which is a slightly odd decision if you think about it. Either way, it's not a joke, it does exist and the claim is that it has 50% better performance than H.265/HEVC. It's as well to be cautious, of course: there are many claims, but no whitepapers with technical detail substantiating those claims, although its award for 'Best in Show' from NewBay Media at this year's NAB Show somewhat supports the veracity of its claims.
Now, I've been in presentations such as the one that V-NOVA ran earlier several times before, with codecs such as Dirac, Snow, VP8 and HEVC in the spotlight. Many of those do represent a considerable advance in picture quality per unit bitrate, but often present larger (sometimes very much larger) CPU loads to decompress.
Perseus was shown being decompressed on run-of-the-mill hardware such as phones, tablets, as well as a large 4K display. One of the most interesting things about this (and the reason I mention it here) is that there is serious talk of providing Perseus content using existing set-top boxes. One of the company staff has been heavily involved in the design of set-top boxes and it is apparently possible that the hardware in them, which is by and large designed to assist with the decompression of H.264 and similar block-based DCT codecs, could be re-purposed, via a firmware update, to accommodate Perseus. This is an unanticipated development and represents some very good thinking if it turns out to be practical. Software and hardware reference implementations for both encoders and decoders already exist, although none is currently public.
V-NOVA is proposing Perseus as a distribution codec and making great pains to promote it as a way to distribute 4K material at HD bitrates. One of the biggest problems with this situation is that broadcast industry businessmen are likely to be very eager to believe that they can have 4K for HD bitrates, in the sense of economising on expensive broadcast bandwidth, and it's therefore more important than ever to demand and carefully examine both objective and subjective performance metrics. I mention subjective tests because the easiest objective analyses, such as plain signal-to-noise ratio through the codec, can produce results that turn out to be misleading when a subjective, observational trial is done with live human beings.
Either way, if the claimed performance figures are accurate, this is potentially quite important. The company voiced no objection at the idea of Perseus eventually becoming an acquisition codec, although it's historically been true that many codecs have sweet spots regarding compression efficiency. It's actually quite unusual that H.264 and its derivatives scale reasonably well to both low-bandwidth distribution and high-bandwidth acquisition. V-NOVA has apparently existed for five years or so and has kept quiet about its activities, on the basis that a new codec technology is unlikely to gain any traction without industry support. They now have the involvement of Hitachi, with camera and broadcast technology, as well as other common names in the world of content distribution.
What we need now is good information on how well Perseus actually works as a codec. V-NOVA is not the first organisation to announce a compression technology and several others have failed to gain traction. Yet if it works as well as it is claimed to work, it certainly deserves to succeed.