Youtube : All New Videos are transcoded into WebM

[HawkMan]

Must be the video you were watching. I'm using html5 webm on firefox for most of the youtube content I watch and the quality has much improved.

So this is what you call much improved ? I say this looks like significantly reduced.

Firefox can't include it because h.264 conflicts with most FOSS licences. Besides, why would anyone choose a patent encumbered with a precarious future, to a patent-free alternative?

yes they can. just like they can use any other platform specfici API that is patent encumbered, they can use the codecs that exist on the platform. i.e. the built in windows or OSX codecs, without problems. even on linux they coudl support them if the user install the required codecs himself, or if he uses a usable distro that understands that people can't live be FOSS alone.

[Udedenkz]

Think about this from a END USER perspective,

Windows Users,

- h264 is free and provided with OS

- DirectX Video Acceleration ensures super smooth and CPU-free video playback

Mac Users,

- h264 is free and provided with OS

- I think you guys have hardware acceleration as well, except for some GPUs and Adobe Flash Player or something like that.

Linux Users,

- You guys can complain, but then again you are irrelevant minority.

Must be the video you were watching. I'm using html5 webm on firefox for most of the youtube content I watch and the quality has much improved.

Firefox can't include it because h.264 conflicts with most FOSS licences. Besides, why would anyone choose a patent encumbered with a precarious future, to a patent-free alternative?

Another thing to seem to be forgetting is that there are far more FOSS browsers out there than proprietary ones. Even amongst the main browsers, Opera, Firefox, and Chrome all support WebM/Theora exclusively. Only IE, which can play WebM if the codec is installed, and Safari don't have WebM/Theora out of the box. Clearly the majority of browsers are going FOSS/Patent free codecs. And now with youtube as well, the most popular site on the web now supports a patent free alternative. H.264 is dead :D

Ye, their agenda is to free society from patent encumbered codecs, a noble one if I might say. Now we just need to fully replace mp3 with ogg theora; it's getting there :D

And GNU/Linux has built in support for WebM, Ogg/Theora etc, your point being? Firefox and Chrome use OpenGL for hardware accleration on OSX and GNU/Linux not DirectX.

It's very cool. I suggest everyone who uses Firefox 4 / Chrome enable WebM join the html5 trial at http://www.youtube.com/html5 . I've been using it for a while, and It's very sweet. No more laggy flash, or patent encumbered H.264 videos :D

I am using the screenshots provided here to compare. I tried HTML5 with IE9 and it really goddamn buggy compared to just using Adobe Flash Player 10.2

This is their issues, they will simply have to stick with Adobe Flash Player instead then. They refuse to support h264 because they are stubborn.

If you want h264 gone so badly, get everyone to stop buying Blu Ray movies. h264 will stay because it is the HD standard, hardware and software is forced to provide support for it because it is forced to play Blu.

Laggy flash means that you have a GPU which doesn't support hardware video acceleration. Or you are using a legacy operating system which came before the Windows XP era.

Lacking a good GPU and a good CPU will give you lagg with HD Adobe Flash video. IMO Adobe should acquire CoreAVC - their software h264 codecs are a failure.

[HawkMan]

CoreAVC is fast only because it heavily sacrifices quality though.

[ViperAFK]

Well at least it's still opt-in only. H264 remains standard and if you don't have flash player it'll still tell you that you need it for viewing Youtube videos. So while they do plan on offering WebM for all their videos, H264 isn't going away anytime soon. At least that's what I figured out reading the blog entry carefully and some of the user comments.

So if that's the case I'm not really worried.

I'm all for replacing Flash with HTML5, but technically they could keep the same codecs and serve it in HTML5... I just don't see what's the big interest for that crappy VP8 codec.

h.264 html5 wouldn't be very viable because firefox, opera, and chrome don't support it... (chrome did but removed, or is removing support can't remember which). Internet explorer can support webm if you install the webm codec:

http://blog.webmproject.org/2011/03/introducing-webm-in-internet-explorer-9.html

So for html5 webm seems the better choice when it comes to browser support.

[Udedenkz]

CoreAVC is fast only because it heavily sacrifices quality though.

Nope. It is just fast.

It is a lot faster when you turn off deblocking, but so is MPC:HC.

[Prince781]

If you use html and join the html test at youtube.com/html5 you can see for your self. Personally, the quality seems much improved over all vs flash.

Firefox can't include it because h.264 conflicts with most FOSS licences. Besides, why would anyone choose a patent encumbered with a precarious future, to a patent-free alternative?

Another thing to seem to be forgetting is that there are far more FOSS browsers out there than proprietary ones. Even amongst the main browsers, Opera, Firefox, and Chrome all support WebM/Theora exclusively. Only IE, which can play WebM if the codec is installed, and Safari don't have WebM/Theora out of the box. Clearly the majority of browsers are going FOSS/Patent free codecs. And now with youtube as well, the most popular site on the web now supports a patent free alternative. H.264 is dead :D

Flawed, your ignorant FOSS bias seems to have left you without a realistic assessment of the qualities of WebM vs Flash.

I actually do want WebM over Flash. Flash has to GO. However, I do realize that, at the moment, the H.264 codec does produce better quality than WebM. You know that too, Flawed. The only reason why you automatically assume all software that is not Free and Open Source to be low-quality is because it is, obviously, not open source. I'll admit, I'm a fan of WebM and wish it a long future on the web, but judging the quality of software (H.264 video) based on its openness or lack of openness in the community is sheer idiotic bias. I highly advise you to revisit your statements with a logical approach based solely on its quality, and not its openness. The fact is, if H.264 was admitted into the open-source standard, you would suddenly view it as high-quality too. With this being said, I hope that you may comment in a more logical way.

[Udedenkz]

Adobe Flash Player,

HTML5 + WebM,

[OuchOfDeath]

Chrome doesn't accelerate any part of the browser yet. Try Firefox or Opera.

[funkydude]

Chrome doesn't accelerate any part of the browser yet. Try Firefox or Opera.

Err, yes it does. Opera does NOT unless you're trying to re-commend alpha software to forum readers...

[Flawed]

HTML5 + WebM,

Doesn't seem to use that much on my system. Firefox usually consumes about 7-10% cpu on average, so an extra 8-11% is minimal really. Flash definitely uses more. By the way, I like that tune :D

[OuchOfDeath]

Err, yes it does. Opera does NOT unless you're trying to re-commend alpha software to forum readers...

It does not. And Opera has had hardware acceleration since 10.5: http://my.opera.com/core/blog/2009/02/04/vega (Vega has been in the stable release since 10.5)

[OuchOfDeath]

Adobe Flash Player,

HTML5 + WebM,

That's not a valid comparison for the reasons I've mentioned above. I think IE9 would be the best comparison to use as it has the best hardware acceleration currently, and I gave it a shot just now: https://tools.google.com/dlpage/webmmf

The problem is even with searching for webm videos in the youtube advanced search options, there's no real way to tell if it is indeed playing WebM aside from staring at the videos and comparing them to see if one's blurrier than the other. I don't have the time or the patience for that, but if someone wants to go ahead and do that feel free.

To the results: The HTML5 player in IE9 player was between 3-5% lighter on the cpu than the flash one. Whether it was WebM or h.264 I cannot say for certain, but according to Google's own instructions it should have been a WebM video.

[Udedenkz]

Chrome doesn't accelerate any part of the browser yet. Try Firefox or Opera.

Irrelevant.

I used two different browsers to have one in the HTML5 test and one out of the HTML5 test.

Additionally, Chrome does have hidden options to turn on hardware acceleration.

Both browser had hardware acceleration turned on, if that matters.

Doesn't seem to use that much on my system. Firefox usually consumes about 7-10% cpu on average, so an extra 8-11% is minimal really. Flash definitely uses more. By the way, I like that tune :D

Linux - no video hardware acceleration and less than 1% use it. No one cares.

EDIT: I did not mean that in an offensive fashion.

That's not a valid comparison for the reasons I've mentioned above. I think IE9 would be the best comparison to use as it has the best hardware acceleration currently, and I gave it a shot just now: https://tools.google.com/dlpage/webmmf

OK I will compare with IE9.

The problem is even with searching for webm videos in the youtube advanced search options, there's no real way to tell if it is indeed playing WebM aside from staring at the videos and comparing them to see if one's blurrier than the other. I don't have the time or the patience for that, but if someone wants to go ahead and do that feel free.

I hope you are joking. You can 100% tell whether or not it is HTML5 by a simply right click!!

Additionally, if is IE9 it is h264 HTML5 and if it is FF/Chrome/Opera is is WebM HTML5.

[OuchOfDeath]

Irrelevant.

I used two different browsers to have one in the HTML5 test and one out of the HTML5 test.

Additionally, Chrome does have hidden options to turn on hardware acceleration.

Both browser had hardware acceleration turned on, if that matters.

They are not the default option for a reason. They're beta and not fully implemented. Don't use them for any sort of comparison that wants to be taken seriously.

OK I will compare with IE9.

I hope you are joking. You can 100% tell whether or not it is HTML5 by a simply right click!!

Additionally, if is IE9 it is h264 HTML5 and if it is FF/Chrome/Opera is is WebM HTML5.

I hope you are joking. If you'd even try to read my post you'd realize that there is a WebM plugin for IE9. I'll state this again for emphasis. THERE IS WEBM FOR IE9.

See how my post now makes sense?

EDIT: Here's a quote from the WebM website: Also, playback performance and rendering is not yet fully optimized in the browsers that support WebM but improvements are forthcoming. The computational efficiencies of the VP8 codec are more accurately measured today using codec-level development tools in the SDKs.

[Udedenkz]

Now same video without the browser/plugin imposed overhead,

720p h264 obviously (Specs in signature, second rig)

They are not the default option for a reason. They're beta and not fully implemented. Don't use them for any sort of comparison that wants to be taken seriously.

I hope you are joking. If you'd even try to read my post you'd realize that there is a WebM plugin for IE9. I'll state this again for emphasis. THERE IS WEBM FOR IE9.

See how my post now makes sense?

EDIT: Here's a quote from the WebM website: Also, playback performance and rendering is not yet fully optimized in the browsers that support WebM but improvements are forthcoming. The computational efficiencies of the VP8 codec are more accurately measured today using codec-level development tools in the SDKs.

I am using Adobe Flash Player 10.2 in SRWare Iron 10. Chrome has no support for h264 any longer.

I am also using Firefox 6.0a for WebM, it is much faster than 4.0 to me.

Well obviously there is a plugin! Google doesn't silently install it though so with IE9 you will be using h264.

Also, you might want to argue that it is less BIASed to compare software-only decoders for h264 vs WebM. The de-facto fastest software decoder is CoreAVC 2.5... fastest free one probably comes with Media Player Classic Home Cinema. We can go comparing that route...

[funkydude]

It does not. And Opera has had hardware acceleration since 10.5: http://my.opera.com/core/blog/2009/02/04/vega (Vega has been in the stable release since 10.5)

We're talking about youtube videos. Opera doesn't accelerate anything of actual usefulness. Chrome does, you're simply wrong. Try some HTML5 speed demos: http://ie.microsoft.com/testdrive/

[Andre S. Veteran]

webm: http://img812.imageshack.us/img812/4671/1webm.png

h264: http://img862.imageshack.us/img862/1276/1mp4.png

webm: http://img847.imageshack.us/img847/1052/3webm.png

h264: http://img855.imageshack.us/img855/9206/3mp4.png

I'm not a "videophile" but they all look the same to me :unsure: well, not exactly the same, but about the same quality.

It looks like the source frame itself is quite blurry. Therefore both codecs would arrive at similarly mediocre results.

Quoting x264 devs is like quoting Microsoft on Linux.

Actually that x264 dev was right on the money: Patent attack launched on Google's open video codec

[Andre S. Veteran]

Meaning you used a different video decoder than the one the website/browser would use, possibly reducing the output quality of the h.264 video. The codec and encoding is only 2/3rds of the equation. and depending on what h.264 codec you have installed it may show the video at significantly lower quality.

Not really. Just like mp3s sound the same on any player or jpgs look the same on any image viewer, these codecs are standards and any decoder should produce the same output out of the same source. Unless flash/html5 does some hidden post-processing I'm not aware of, it's unlikely there are significant differences between playing the video on youtube and on ffmpeg.

[HawkMan]

Not really. Just like mp3s sound the same on any player or jpgs look the same on any image viewer, these codecs are standards and any decoder should produce the same output out of the same source. Unless flash/html5 does some hidden post-processing I'm not aware of, it's unlikely there are significant differences between playing the video on youtube and on ffmpeg.

It's not actually like that. nowadays MP3's do actually sounds mostly the same. but back in the days there was a big difference between the high and low quality decoders.

and unlike what the other guy in this thread is claiming so is the case with h.264 today, like CoreAVC. it's a really fast decoder that can decode on fairly low end hardware. but you'll never find anyone on a htpc site reccomending that you use CoreAVC. because it has lower PQ than other decoders. CoreAVC has of course gotten better than it used to be, but it is still a decoder made primarily for speed not quality.

[Miuku.]

Actually that x264 dev was right on the money

That patent dispute has since gone absolutely nowhere. MPEG-LA is, as always, completely impotent.

Reminds me greatly of Microsoft's strategy towards Linux, baseless threats and attempts at litigation whilst having absolutely nothing going for them - weeping in front of the inevitable.

[coth]

Back in years - I remember people were complaining that Microsoft abusing web standards and pushing their own. Isn't it what Google now doing?

Anyway. I don't like HTML5 video on YouTube as it lacks full screen feature. But i don't like Flash either, especially buggy 10.2.

[ichi]

Actually that x264 dev was right on the money: Patent attack launched on Google's open video codec

Did you read that article? What it basically says is this: "We don't have crap against WebM, so if anyone anywhere has any patent that can be used against them then please help us".

Nothing happened so far, though.

[OuchOfDeath]

Back in years - I remember people were complaining that Microsoft abusing web standards and pushing their own. Isn't it what Google now doing?

No, it's not. Microsoft pushed closed-source proprietary and to top it off non-standard behaviour in their web browser and the web. WebM is free, opensource, and of course standard as well, though it's hard to be non-standard when it comes to a spec that says any codec is allowed. The bottom line is no it's the opposite. Google is not abusing web standards but protecting them.

[coth]

Being open source doesn't mean to be automatically being standard. WebM is not listed among HTML5 codecs.

http://www.w3.org/TR/html5/video.html

And there is Theora in Ogg listed for those who like open source.

What Google is doing is simply pushing their own standards for own commercial interests.

[ichi]

Being open source doesn't mean to be automatically being standard. WebM is not listed among HTML5 codecs.

http://www.w3.org/TR/html5/video.html

All I see there is "The following list shows some examples of how to use the codecs= MIME parameter in the type attribute", which is about the parameter usage, not about codec support or endorsement.