Thoughts on Mozilla's New Stance on HTML5 Video
The debate over what the HTML5 video standard should be has been going on since around the latter years of Web 1.0. These debates have historically centered around three codecs: the patent-encumbered but extremely ubiquitous H.264, and the two royalty-free codecs Ogg Theora, and VP8. The full details of the debate are too numerous to place in this post, but all that is needed to know is that up until just recently, the HTML5 <video> camp has been divided in two: those that support H.264 (such as Microsoft and Apple), and those that support VP8 (such as Mozilla and Opera); Ogg Theora has mostly faded into irrelevancy at this point. Google used to remain neutral in this debate, however they have recently sided with VP8 (which they bought and opened to the public in 2010), believing that the open web shouldn't have to pay a fee to utilize these technologies.
A lot of us thought that this stalemate would never break, but after a surprising post from Mozilla's Andreas Gal that Firefox may soon begin supporting H.264 in its browsers, there has been something of a firestorm in light of this recent revelation. Many people thought this day would never come, since Mozilla has historically remained adamant with its stance regarding HTML5 video. However, if Mozilla does cave in and support H.264 playback on its browsers, this could mean that the long debate will finally be over, and that H.264 will become the new standard for HTML5 video.
I tried to give this piece of news a little more thought, and these were what I got:
- H.264's dominance may already be secured. Many critics of the H.264 codec have been hopeful that the appeal of VP8's royalty-free license would gain strong traction across the industry, and that the need for the patent-encumbered H.264 would eventually erode. What we are seeing, however, is the opposite. The install-base of H.264-capable devices continues to grow, thanks in no small part to the explosion in mobile. Nearly every device we own, nearly every software we use, and nearly every video service we rely on utilizes as aspect of H.264, and it pretty much works just fine the way it is. In other words, from a technical standpoint, there isn't really much work left to be done when it comes to adoption; H.264 is surviving real-world use conditions just fine, and pretty much everyone who watches digital video is using the technology anyway. At this point, appointing H.264 as the new standard looks like the path of least resistance.
- VP8 isn't living up to its promise as an equivalent, if not better video experience than H.264. Even prior to the acquisition by Google, many have already pointed to On2's VP8 as the closest rival to H.264. It's quality was better than Theora's (and let's face it: it never really was that good), and its technology was rigid enough for commercial use (I believe Skype uses an earlier version of On2's codec). When Google decided to open up VP8, a lot of people believed that it had enough technical leverage to surpass H.264; that didn't happen, unfortunately. It turns out that as far as quality goes, VP8, for the most part, still doesn't beat H.264. However, what I think is an even more crucial weakness in VP8 goes even lower-level than that: hardware support. Many chipsets come with a dedicated chip for processing H.264 video. What this means is that these chips are able to bring high-quality video without stressing the processor to much. The problem is that VP8 is still not supported by these dedicated chips. I, for one, cannot name a single chip that can playback VP8 video using hardware acceleration. That means that in order to play VP8-encoded video today, you would have to force all that immense data through the CPU, which could lead to greater heat generation, and larger power drain. On my laptop, playing a VP8 video would cause my fan to kick in and generate a lot of noise, almost as if I was playing Battlefield 3 or some other intense video game. And don't even get me started on mobile, where CPUs are considerable weaker. Herein, I think, lies the inherent weakness of VP8. Not only is it not widely adopted, but the software and hardware support thus far isn't good enough to bring it toe-to-toe with H.264.
- Mozilla could not keep its ideological stance forever. I don't really blame Mozilla for finally caving in like this. Its stance, in light of the things I pointed out above, seems nearly impossible to keep over the long term. At this point it probably costs a lot more to change everything (hardware, software and content) to support VP8 than it is to simply pay the licensing fee. Mozilla probably realized that under these circumstances their arguments were too weak to have any practical benefits. We also cannot discount that Mozilla, despite being a non-profit organization, still needs money in order to operate, and Firefox's continued relevancy is crucial to keeping the cash flowing in for Mozilla. Firefox needs to be a compelling product, and without native video playback, the browser would already be at a major disadvantage, especially in mobile, where the ability to playback smooth video is especially important. This may have been a business decision more so on than it was an ideological decision, but Mozilla must find ways to keep Firefox competitive, or they risk facing irrelevancy.
These were just a few thoughts that I had in light of their recent announcement. Please know that I am in no way being overly critical of VP8 or Mozilla for that matter. Personally, I would love to have a royalty-free codec for the web, one which we can freely use without any worry of money, licensing, or the like. The problem is that that prospect becomes more difficult to pull off every passing day. The advances in today's tech and web are not helping VP8's appeal in any way, and we all, sooner or later, must stop being stubborn and resistive, and instead actually work towards something progressive. It seems more and more likely that we may have to abandon VP8 as an option, and that H.264 may be the way to go moving forward. I think Mozilla's recent decision reflects that consensus, and this story may not be the last we will hear in a while. So, do you have any other thoughts on this matter?