11 Comments

Summary:

There are many strategic reasons why Google might have dropped support for H.264, but lowering costs for YouTube isn’t one of them. The decision will only add cost and complexity for the video site, which now just has yet another video format to support.

on2videothumb2

Google’s decision to kill support for H.264 in its Chrome browser has many scratching their heads and wondering why it would do so. One popular theory going around the blogosphere Wednesday was that Google is switching to its open source WebM/VP8 format as a way to lower costs for YouTube.

The idea is that by consolidating its video file formats, Google could save massive amounts of money on infrastructure costs. (There’s a somewhat related theory that Google will use YouTube as a way to increase adoption of VP8 by moving all of its video assets to the open source format and “forcing” consumers to use supporting browsers.) But despite the move by Chrome, YouTube isn’t going WebM-only anytime soon, which means that supporting the new format will raise its costs, not lower them.

We’ve already discussed how Google’s decision to kick H.264 from its browser will raise costs for regular video publishers, but the same argument holds true for YouTube. Adding a new video format means re-encoding existing files, as well as adding a whole new group of files to store on its servers. YouTube has already begun this process, and nearly all its files are already WebM-compliant.

But in addition to creating a whole new video profile for all of the bajillions of videos it already hosts, YouTube would have to add the VP8 profile to all existing files uploaded. Since YouTube users upload an average of 35 hours of video each minute, that means adding considerable cost on the infrastructure side.

Those costs could be offset, it is argued, if YouTube rationalized the file formats it supports — that is, got rid of H.264 files — which would then free up all the infrastructure and resources currently being taken up by encoding and story those files. There’s just one problem: YouTube would never do that, nor would any rational online video company.

Let’s be clear: YouTube is in the media business, which means that it wants to be in front of the widest audience possible. This is the reason that YouTube has invested considerable time, energy and, most importantly, money to develop mobile applications and mobile web experiences, as well as getting its videos embedded on connected TVs and other devices in the living room. So YouTube would not simply “turn off” its H.264 streams and hope for the best. It would need to keep supporting those legacy files and encodes, simply because it would need to continue serving those files to users who want to see them, in whatever format they are capable of viewing them in.

Perhaps more important than just its desktop footprint, though, is Google’s interest in serving videos to mobile phones and other connected devices. Today, due to H.264 hardware acceleration built into those devices, the YouTube needs to be able to serve H.264-encoded files. That is, if YouTube shut down H.264 streams, it wouldn’t just be cutting off unfortunate desktop users that opted for Safari or hadn’t upgraded to the most recent version of Firefox, it would be shutting itself off from a wide range of devices — including its own Android mobile handsets and Google TV devices — that don’t currently support VP8.

In short, there are many strategic reasons why Google Chrome might have dropped support for H.264, but lowering costs for YouTube isn’t one of them. The decision only gave YouTube, just like everyone else, more complexity to worry about.

Related content on GigaOM Pro: (subscription required)

You’re subscribed! If you like, you can update your settings

  1. What is Google going to do after 2015 when the current h.264 web licensing agreement expires? Let’s say the licensing agency picks $4/video and sends Google a bill for $1B/yr? You’d be crazy to take a risk like that.

    If h.264 wants webm to go away, supply a permanent free license for web use.

    1. Frankly I think licensing fears over H.264 are overblown, but see your point. Still, I think rhetoric based on infrastructure costs are even more overblown!

      1. I see two main reasons for doing this:
        1) to have a real option when 2015 when comes around. If the MPEG-LA is unreasonable Google will 100% pull the plug on h.264. The existence of a credible webm option will keep MPEG-LA in check. Do you really want some of your major competitors (Apple, MS) picking the royalty number you are going to have to pay for h.264?

        2) Firefox is screwed. Firefox is tri-licensed. One of those licenses is the GPL. From the GPL v2, “b) You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.” Firefox’s license prevents them for ever incorporating h.264. Same for Chromium.

        So if MPEG-LA wanted to get rid of this problem, they could provide a royalty free license for web use until 2030 when their patents expire and be content with shaking down the hardware manufacturers. Instead they are trying to ingrain h.264 as a widely used standard. Then I fully expect to see robber baron level royalty requirements after 2015 unless there is something around to keep them in check.

    2. Jon, you missed the big news a few months ago that MPEGLA changed the 2015 deadline on the current agreement of no royalty fees and made it a permanent moratorium. Here’s some info from a quick search: http://arstechnica.com/media/news/2010/08/mpeg-la-counters-google-webm-with-permanent-royalty-moratorium.ars

      I am personally starting to think that the real reason why Google made this change to Chrome is to pressure MPEGLA to drop their royalties for h.264 software in general (or at least browsers), and possibly change their stance of forbidding open source implementations, in order that all browsers will eventually be able to include h.264 support without violating anything and without cost.

      The WebM trump card has worked well in obtaining concessions from MPEGLA in the past (see the arstechnica article above), so why not try to use it again? It’s like Google pulling MPEGLA aside and saying “make it free and available to all or we will never let you be THE web video standard.”

      1. Just to clarify my post above: MPEGLA made permanent the moratorium on charging royalty fees for SERVING or DISTRIBUTING h.264 video FOR FREE on the web, but they still charge royalties for any software that implements h.264 and for websites that charge for viewing the videos. This is why I think that there are still more concessions that Google can squeeze from them.

      2. “effectively making free-to-view H.264 encoded video royalty-free indefinitely”. I thought that was being interpreted as “free” as in carrying no advertising. The whole point of YouTube is to carry advertising.

    3. OSS supporters claim FUD with valid patent concerns, the bait-and-switch argument is just as much FUD. It’s not worth consideration.

      1. MPEG-LA owns a patent troll.

        http://www.techdirt.com/articles/20100430/0232599255.shtml

        My trust in patent trolls is total – they will definitely do what ever they can to extract the most money possible out of the market place.

  2. Mgeg-la can only raise costs by 10% every 5 years.

    See
    http://fosspatents.blogspot.com/2010/06/mpeg-las-avch264-licensing-terms.html
    Lots of good info.

    1. We don’t know what the rates for encoding a web video for internet play are going to be. They aren’t collection royalties on this currently but they have said they will start in 2015. The rates for broadcasters encoding with h.264 in non-internet areas are quite high.

      That’s the number that will kill Google. Even a dime would make them owe hundreds of millions of dollars each year.

      The $5-6.5M cap only applies to codec engines distributed.

  3. It’s the web, stupid
    http://robux4.blogspot.com/2011/01/open-web-conspiracy.html

    Still don’t understand why so few people got this simple fact yet. It’s simply avoid using patented tech on the web, ie the way it should be. No more, no less.

Comments have been disabled for this post