News for hackers. Stuff that matters.
Re:Google's dropping H.264 in Chrome is not a step backward ... by grayrest on Thursday January 13, 2011 @12:23PM (#2100530)
All the outcry over this I've read is basically a complaint that you can't ship one codec for HTML5 video, which you've never been able to do. That's what the whole argument over the video tag has always been about.

The only difference this makes is that this cements the split instead of everybody expecting Firefox and Opera to give up and adopt H.264. If you were willing to ship just H.264 and flash fallback for Firefox/Opera, why wouldn't you be willing to ship H.264 and flash fallback for Chrome?

Re:Google's dropping H.264 in Chrome is not a step backward ... by MiguelHudnandez on Thursday January 13, 2011 @12:45PM (#2100596)
On desktops, falling back to flash for H.264 decoding is a pretty good strategy.

On mobiles, though, Google's change has a bigger effect. Flash is abysmal (if existent) on Android devices--and now there is no clear solution that doesn't involve serving multiple encodings.

Before this change, you could be comfortable that your H.264 HTML5 video would play on the majority of smartphones; the flash fallback was only necessary for desktop browsers.

Re:Google's dropping H.264 in Chrome is not a step backward ... by eordano on Thursday January 13, 2011 @01:26PM (#2100743)
There is no Chrome for smartphones, and so far no one has said anything about dropping H.264 on Android (although you can see it happening as soon as WebM extensions for ARM start being more popular).

(edited: typo)

Re:Google's dropping H.264 in Chrome is not a step backward ... by MiguelHudnandez on Thursday January 13, 2011 @07:13PM (#2101906)
You are correct, but why would you imagine Google would treat Android differently? Even though they haven't announced any changes specific to Android, I expect Google's left and right hands to work together... eventually.
Re:Google's dropping H.264 in Chrome is not a step backward ... by beej71 on Friday January 14, 2011 @05:05AM (#2103075)
That's a fine argument, but it's just not the one in the article.
Re:Google's dropping H.264 in Chrome is not a step backward ... by DrHankPym on Thursday January 13, 2011 @08:22PM (#2102108)
Because Android is open and developed with the cellular carriers and manufacturers. If they want to add a H.264 chip, why would Google stop them?
Re:Google's dropping H.264 in Chrome is not a step backward ... by ZeroGravitas on Friday January 14, 2011 @05:41AM (#2103133)
They'll almost certainly have an H.264 chip but the standard Android browser (and Firefox and Opera Mobile) might not take advantage of it when faced with an HTML5 video tag.

Maybe we'll see Apple and Microsoft release browsers for Android!

Re:Google's dropping H.264 in Chrome is not a step backward ... by Osiris on Thursday January 13, 2011 @01:16PM (#2100711)
From what I understand, the main reason for Mozilla and Opera not wanting to include H.264 is that there would be licensing fees for including a decoder with the browser. Correct me if I'm wrong, but it could potentially cost both Mozilla and Opera a significant amount in licensing fees to include H.264 support.

Apple doesn't care because it's already a licensee, the same for Microsoft (including in Win7) and Google (encoding for YouTube).

Re:Google's dropping H.264 in Chrome is not a step backward ... by bzbarsky on Thursday January 13, 2011 @02:36PM (#2101034)
The fee would be about $5 million a year. Arguably, both could afford it (it would be about 5-10% of revenue, but doable).

The reason Mozilla is not wanting to include H.264 is that:

1) It can't do that in the open source version, and doesn't want to have separate open-source and closed-source releases. 2) It feels H.264 would be bad for the web, since it requires paying to create content in H.264.

Reason #2 is somewhat more important than reason #1, since #1 could be worked around by using system media frameworks, etc.

Re:Google's dropping H.264 in Chrome is not a step backward ... by Locke1689 on Thursday January 13, 2011 @03:11PM (#2101155)
Both of those things are misleading or wrong. There's no reason why they would have to make it closed source, especially if they use x264. H264 patents don't require everyone to pay to create content, you should read the MPEG licensing agreement.
Re:Google's dropping H.264 in Chrome is not a step backward ... by bzbarsky on Thursday January 13, 2011 @04:45PM (#2101444)
> There's no reason why they would have to make it closed > source, especially if they use x264.

That's not what the lawyers said. In the US, at least; things may be different in your jurisdiction.

Note that this is the route Chrome used to take: Chromium did not support H.264, while the Chrome binaries, built from some source that was not public, did.

> H264 patents don't require everyone to pay to create > content

They do require you to pay to produce an encoder. Of course someone can then distribute said encoder for free if they want, so the end user doesn't technically _have_ to pay. But in practice they do.

And if you use an unlicensed encoder, then you're infringing on the patents. So is anyone who views any videos that were produced with the unlicensed encoder, even if their decoder is licenced. And some other restrictions to do with the fact that encoders are typically not licensed for arbitrary use (e.g. the encoder in a typical digital camera doesn't give you the right to put the resulting video on a website that you make money from, via Google Adwords, say). You should in fact read the licensing agreements involved, or at least the excerpts in http://bemasc.net/wordpress/2010/02/02/no-you-cant-do-that-w...

Re:Google's dropping H.264 in Chrome is not a step backward ... by blasdel on Friday January 14, 2011 @09:43AM (#2103601)
You are spectacularly incorrect about how Chrome and Chromium are built — the only difference between the two regarding <video> is which flags the bundled copy of ffmpeg's libavcodec was built with.

There is no secret source code. When you build Chromium, you can just swap in a symlink to your system's copy of libavcodec, and then it magically supports whatever codecs that one was built with.

Re:Google's dropping H.264 in Chrome is not a step backward ... by Locke1689 on Thursday January 13, 2011 @05:56PM (#2101700)
That's not what the lawyers said. In the US, at least; things may be different in your jurisdiction.

http://mailman.videolan.org/pipermail/x264-devel/2010-July/0...

Even if you decide to use a closed-source decoder, just because the decoder has to be closed source doesn't necessarily mean the entire browser has to be closed. What exactly was the lawyers' reasoning behind this?

http://www.mpegla.com/Lists/MPEG%20LA%20News%20List/Attachme...

They do require you to pay to produce an encoder. Of course someone can then distribute said encoder for free if they want, so the end user doesn't technically _have_ to pay. But in practice they do. And if you use an unlicensed encoder, then you're infringing on the patents. So is anyone who views any videos that were produced with the unlicensed encoder, even if their decoder is licenced.

What are you talking about? The page you linked clearly says that non-commercial content is free. I didn't say that no one has to pay to create content, I said not everyone has to pay to create content.

Re:Google's dropping H.264 in Chrome is not a step backward ... by bzbarsky on Thursday January 13, 2011 @11:44PM (#2102587)
> doesn't necessarily mean the entire browser has to be closed.

Right. I did say that #1 can be worked around. If you you you approach, you still get the Chromium/Chrome model, where you can't release the actual source to the binaries you ship, and worse yet others can't easily fork your "open-source" app while keeping the functionality or redistribute your binaries.

> The page you linked clearly says that non-commercial content is free.

"commercial use" in that context includes putting it up on a website that has ads. So yes, not everyone has to pay to create content. If you find someone who will give you a free encoder for which they paid the licensing fees, and you make sure that either you don't put your content on the web or if you do that there are no ads on the page and you don't charge for the videos, then you can get away with paying nothing.

Quick; name me the places on the web where there are free videos and no ads. I'd love to know where this is!

Re:Google's dropping H.264 in Chrome is not a step backward ... by Locke1689 on Friday January 14, 2011 @02:28AM (#2102865)
"commercial use" in that context includes putting it up on a website that has ads. So yes, not everyone has to pay to create content. If you find someone who will give you a free encoder for which they paid the licensing fees, and you make sure that either you don't put your content on the web or if you do that there are no ads on the page and you don't charge for the videos, then you can get away with paying nothing. Quick; name me the places on the web where there are free videos and no ads. I'd love to know where this is!

You completely misinterpreted that. First, the author pulled ads out of his ass -- there's no legal statement to support that theory. Second, the statement from MPEG-LA that I linked to trumps that article -- as long as you're not charging for H264 encoded content you have nothing to fear.

Re:Google's dropping H.264 in Chrome is not a step backward ... by bzbarsky on Friday January 14, 2011 @09:55AM (#2103659)
You're misunderstanding the statement. That seems to be pretty common.

H.264 has royalties that you need to pay to produce a licensed encoder.

Then it has _separate_ royalties you have to pay to take content encoded with a licensed encoder and send it to someone else.

Furthermore, whether an encoder is "licensed" depends on what you do with its output; the details are in the license that came with your encoder.

The statement you linked to is that the royalties on sending to someone else don't have to be paid as long as you're giving it to them for free and sending it over the internet _and_ your encoder is licensed for whatever else is going on (e.g. if you're making money in the process then your encoder has to be licensed for that).

I'm saying that to produce the content in the first place you have to pay money to the MPEG-LA. If you don't, then anyone you send it to has to pay money to the MPEG-LA. Here "has to" just means "they can sue you and will win if you don't, since the licensing terms don't allow you to do that". They might, of course, not sue you. That's up to them.

Further, most encoders out there are not licensed to produce content you make money from (note that this is NOT the same as charging users money). So you'd have to buy your camera, then pay the MPEG-LA for an encoder license that allows commercial use. This is a one-time payment. Then you could stream the resulting encoded media as much as you want and not have to pay for each download that happens. The per-download fees are what they waived.

Re:Google's dropping H.264 in Chrome is not a step backward ... by varjag on Friday January 14, 2011 @05:35AM (#2103120)
You are left here on benevolence of MPEG-LA consortium. If they ever change their mind, you'll have to shell out.
Re:Google's dropping H.264 in Chrome is not a step backward ... by blasdel on Friday January 14, 2011 @09:53AM (#2103649)
You are again spectacularly incorrect, this time about the relationship between the distribution of open source code and patent law.

It is perfectly 100% legal to distribute uncompiled source code without a license to any related patents. After all, it's a description of the patented machine and not the machine itself, much like an ideal patent application. Where you run into trouble is distributing the compiled binaries, the actual machines covered by patents.

You can get into no trouble at all over source tarballs of ffmpeg, but Debian is deathly afraid of hosting usefully-compiled debs of it in their repositories for a reason.

Re:Google's dropping H.264 in Chrome is not a step backward ... by bzbarsky on Friday January 14, 2011 @10:15AM (#2103742)
I didn't say anything about distributing source code. To a first approximation users download binaries, not source code (things like Gentoo and MacPorts are edge cases; very few users use either one, really) .

What I said is that the old Chrome model, wherein the official binaries ship H.264 and have a paid-for license, if adopted by Firefox, would have meant two things, as I understand:

1) If someone wants to take the Mozilla binaries and redistribute them (possibly with some simple changes like a different default theme or a default preferences file that whitelists some intranet sites or whatever) they have to pay for an H.264 license. Things like copying Firefox from your computer to your friend's computer would need you to have an H.264 encoder license, since you're creating a new encoder (now it happens that the license fee for that level of distribution is 0, so you're actually ok; just don't put that modified package on a public web site).

2) If someone wants to fork the Mozilla codebase and build a browser with equivalent functionality and distribute it, they have to pay for an H.264 license.

#2 there is precisely what GPLv3 is intended to prevent, by the way. Mozilla is licensed under GPLv2, so this would be within the letter of the license, but arguably not the spirit. Certainly we felt that it would not be within the spirit.