WebRTC’s Future: is it H.265 or VP9?

There’s a codec war going on and it is about to get an upgrade to a newer version.

[If you are new around here, then you should know I've been writing about WebRTC lately. You can skim through the WebRTC post series or just read what WebRTC is all about.]

If you’ve been following WebRTC for a while, you should already know that the video codec issue is an open sour. Or more like an official war among WebRTC proponents.

The main themes around this fight can be summed in two main debates:

  • Paid versus free: H.264 requires a license to use and deploy, which prohibits a lot of use cases and adopters; VP8 is free, allowing everyone to use it (on paper at least)
  • Interoperable versus new: H.264 is commonly used with a huge ecosystem of existing vendors and deployed products, making it suitable for interoperability and adoption; VP8 is brand new, making its current footprint in chipsets, vendors and knowledge significantly smaller

While companies are fighting this round with no apparent winner so far, there’s another round brewing, waiting just around the corner: the successors of both of these codecs.

H.265 vs VP9

H.265

H.265 is the successor of H.264. It is led by the same group of companies who will be collecting royalties for its use, and is said to be twice as efficient in the bitrate it requires to get to the same video quality (but requires 3 times the CPU resources).

Approved out of the over, there are companies who are actively promoting it already. I’ve seen demos on YouTube from DoCoMo, Qualcomm, Ericsson and Cisco.

The main marketing pitch of companies? Video takes up most of the internet traffic. Being able to shrink it by 50% means better utilization of the network.

VP9

VP9 is Google’s successor to VP8. Its decoder can already be found in the Chrome browser (at least from what I could deduce on the web).

There’s not much information about it. The only thing I found in comparison to H.265 is this comment by Chester Moy on a Google+ thread:

According to a presentation by Google, VP9 is ~7% behind HEVC/h.265 in terms of quality/bitrate when they tested VP9 in Q4 of 2011 (They started developing VP9 in Q3 2011). Their goal is to become even better than HEVC.

While the details might wrong as to the exact differences, it can be easily deduced that Google is designing VP9 to be comparable in quality with H.265, sans royalty costs on patent licensing.

Which one?

The crossroad isn’t going to change. The same debate between paid and free, interoperable and new will stay with us.

Don’t expect any of these codecs to find its way to a phone near you before 2014, and even then, only on high end ones. It will take several years for this technology to make any difference.

Until then, we are stuck with our current transcoding problems between VP8 and H.264.

No…

Tags: , , , , , , , , , , , , , , , , , ,

Liked this post?

Share it!

Never miss a post!

Or just grab the RSS feed!

Comments

  1. Hi

    Thanks for another great post and update.

    Do you know if there is any support for H.265 in mobile chipsets. It seems like its going to be challenging to get a feasible WebRTC implementation on Mobile devices that supports real world use scenarios without that.

    Cheers, Graham

    • Tsahi Levent-Levi says:

      I am sure chipset vendors are working on such support. Qualcomm already “demoed” it. It will take time for it to get into commercial devices though – 1-2 years at the very least – it takes time to build such capabilities.
      First round of solutions will not be that interoperable, support from the ecosystem and server side will be lacking, etc.

      Wouldn’t hold my hands for it anytime soon.

  2. marcocom says:

    this story focuses on bit-rate, which is important, indeed, but is already sustainable. what VP9 brings with it is a much deeper feature-set than the H standard. transparency, streaming, and far smarter compression.

    as you can already see, in order to make a video codec perform like flash, you need to do what flash needed to do, use that CPU…and generate some heat.

    • Tsahi Levent-Levi says:

      VP9 has a better future than VP8 in chipset adoption.

      VP8 was out of the game for the years it took H.264 to get there. H.265 and VP9 are starting relatively at similar times, which means H.265 will probably get to chipsets first (due to its current ecosystem), but VP9 won’t be lagging behind for a lot longer.

      If Google nails down VP8 as the mandatory codec for WebRTC, you will see faster and wider adoption of VP9 as well by chipset vendors.

  3. Currently, there are two main problems with webm/VP8, first, it doesn’t have the same quality as mp4/H264, thus, it makes bigger files to have a comparable quality. Second, VP8 encoder is less efficient : with FFMPEG, in a multi-threaded environement, it can be set to use all the cores, but it doesn’t use them up to 100%, and videos take much longer to encode than with X264, which is really not convenient.

    • Tsahi Levent-Levi says:

      As far as I could discern, the quality differences are negligible and the main issue is availability of hardware coding for VP8. Oh… and there’s the patent royalties issue to deal with.

  4. Might want to be more accurate with the “H.264 needs a license to use and deploy”.

    If you are developing an app for or a user of Windows, iOS or OSX then you don’t need a license since Microsoft and Apple have paid for it already.

  5. David Schwartzman says:

Trackbacks

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code class="" title="" data-url=""> <del datetime=""> <em> <i> <q cite=""> <strike> <strong> <pre class="" title="" data-url=""> <span class="" title="" data-url="">