There’s a lot of interest in ultra-high definition (UHD) video, and the two codecs that drive it, HEVC and VP9. Over the past few months, a new UHD codec called Daala has also come to the fore. I wanted to take this opportunity to update the status of HEVC and VP9 and introduce you to Daala.
By way of background, HEVC/H.265 is the royalty-encumbered, standards-based successor to H.264, while VP9 is the free, open-source codec from Google. In case you hadn’t heard, MPEG LA has announced its proposed royalties for HEVC, which includes a $0.20/unit charge on encoders and decoders, with the first 100,000 units excepted. The maximum yearly charge is $25 million, a substantial boost from H.264’s $6.5 million. For content producers, however, there will never be content-related royalties, even for video distributed via pay-per-view or subscription.
So if you’re a content producer, your next two questions are likely “how does the quality compare?” and “where will the two codecs play?” Even though I’ve tested multiple HEVC codecs, none of the encoders that I typically work with support the VP9 codec. I could use FFMPEG, but command lines aren’t my thing, particularly when attempting to produce files for quality comparisons.
So, I checked for objective third-party comparisons and found several, all of which find that HEVC offers superior quality to VP9, though by varying degrees. One study, headlined by Dan Grois, a senior member of the IEEE, was presented at the 30th Picture Coding Symposium in December 2013. Grois and his group compared the quality of HEVC, x264, and VP9, and found that, “according to the experimental results, the coding efficiency of VP9 was shown to be inferior to both H.264/MPEG — AVC and H.265/ MPEG — HEVC with an average bitrate overhead at the same objective quality of 8.4 percent and 79.4 percent, respectively.”
Another study produced by Maxim P. Sharabayko, a postgraduate student at Tomsk Polytechnic University in Russia, compared the quality of x.264, HEVC, VP9, and Daala. Regarding the first three, Sharabayko found “while [HEVC] provides 31 percent better compression rates in keyframe-only mode and about 40 percent improvement in intercoding mode compared to [x264], VP9 is only 18 percent better than [x264] in both modes.”
In terms of where the two codecs play, VP9 unsurprisingly has the early advantage in browser compatibility, with support in Google Chrome, Opera, and Mozilla Firefox. On the desktop, both codecs play in the VLC Player, while only HEVC playback is supported in Rovi’s DivX. Judging solely by press releases, hardware support is coming for both codecs, though HEVC appears to outnumber VP9 by about 2:1. Still no word from the Adobe (Flash), Google (Android), Microsoft (IE), or Apple (Safari or iOS) camps regarding either codec, which basically means that both are dead in the water for any website not named YouTube.
The most intriguing development in the UHD space is Daala, which is the code name for a new codec being developed collaboratively by Mozilla, Xiph.Org, and other contributors. Heading up the development is Xiph.org founder Monty Montgomery, who joined Mozilla to work on the codec in late 2013. Montgomery was the mastermind behind the Ogg Theora codec, which was the leading open source codec until it was supplanted by Google’s VP8.
Where both HEVC and VP9 leverage improvements to existing compression techniques, Daala explores new techniques to achieve both better compression and a technology that doesn’t infringe upon any IP developed for these existing codecs. As mentioned above, Sharabayko tested Daala along with the others, and found that “experimental results obviously show that Daala video encoder is still rather far from being competitive.” Montgomery expects to have “something solid enough with which to begin standardization work in late 2015.” Given the relatively nascent state of the UHD market, that might just be soon enough.