News for hackers. Stuff that matters.
Re:Google to Open-source VP8 for HTML5 Video by tjogin on Monday April 12, 2010 @07:21PM (#1260690)
How would it benefit Google to re-encode every video on YouTube into VP8?
Re:Google to Open-source VP8 for HTML5 Video by aphyr on Monday April 12, 2010 @07:57PM (#1260747)
(My earlier reply was to the original version of your comment.)

It will benefit Google in two ways:

First, it insulates them from encoder licensing fees going forward, and from decoder license fees if MPEG-LA decides to back off their free-for-web-streaming arrangement. It also means they're free to bundle the codec with Chrome.

Second, if you believe the publishing hype, it could save them up to 50% bandwidth. I doubt it's that effective, but at Youtube's scale, processing is cheap and bandwidth is expensive; every 1% helps. If they can reduce the number of storage nodes, bytes transferred, and by extension, open conns/sec, that translates to very real savings.

Re:Google to Open-source VP8 for HTML5 Video by aphyr on Monday April 12, 2010 @07:27PM (#1260702)
They won't recode from h264 and drop the originals immediately. There are too many hardware h264 devices out there. My guess (given how their transition from h263 went) is that youtube will add it as another encoding target and use it for their web formats as soon as flash supports it. Mobile will remain h264 in 3gp, but that's only one out of six formats.
Re:Google to Open-source VP8 for HTML5 Video by FooBarWidget on Tuesday April 13, 2010 @03:44AM (#1261354)
Why is everybody talking about recoding? Can't they just keep the current videos as they are and encode new things in a different codec? As far as I know all the old H263 videos on Youtube have never been recoded. And why would they? Recoding H263 to H264 would just degrade quality.
Re:Google to Open-source VP8 for HTML5 Video by nailer on Tuesday April 13, 2010 @07:43AM (#1261641)
I understand what you're saying, and I'd love to be educated if you know more about the topic, but aren't they both perceptual encoders? Ie, wouldn't the parts of each frame they're identifying as unimportant be quite similar?

(I'm quite certain I'm wrong on this point, but I'm interested in learning why)

Re:Google to Open-source VP8 for HTML5 Video by sp332 on Tuesday April 13, 2010 @01:28PM (#1262698)
Each encode removes information from the video. No matter how nice the encoders are, the video quality will degrade every time.
Re:Google to Open-source VP8 for HTML5 Video by nailer on Wednesday April 14, 2010 @09:24AM (#1264862)
I understand the sentence, but what I'm trying to understand is why the information removed is (significantly) different - since they're both perceptual codecs, there should be a large overlap in what is removed from the output.

Eg, take a frame which has a solid chunk of blue. Codec 1 removes some of data to make things more consistent and then compresses it. Codec 2 looks at the output of codec 1, sees the consistently large blue area, doesn't see any info worth removing, and compresses it.

Re:Google to Open-source VP8 for HTML5 Video by FooBarWidget on Thursday April 15, 2010 @04:27AM (#1267295)
I don't know all the maths behind it, but try saving a JPG, close it, open it, save it again, repeat a few times. You'll see quality degrade every time you close it. My guess is that the lossy compressors try their best to preserve the previous compression artifacts, but in doing so they introduce more compression artifacts.
Re:Google to Open-source VP8 for HTML5 Video by sp332 on Tuesday April 13, 2010 @01:26PM (#1262692)
Google stores the original uploaded videos, just for purposes of re-encoding later.
Re:Google to Open-source VP8 for HTML5 Video by easp on Monday April 12, 2010 @09:21PM (#1260920)
Well, what's the licensing cost for MPEG4? It's probably per user/minute streamed, rather than a one time cost per minute of video encoded.
Re:Google to Open-source VP8 for HTML5 Video by ZeroGravitas on Tuesday April 13, 2010 @04:34AM (#1261407)
You pay per encoder and per decoder, both of which would adversely affect Google with their encoding farms and popular browser if the charge wasn't capped at around $5 million (I think it's gone up to just under 6 million this year) to stop the big players doing something silly like getting together and developing their own royalty-free codec.

Unfortunately for that cunning plan, Google clearly sees more benefit from a codec that's compatible with the open web and open source than they do from licence discounts.

Re:Google to Open-source VP8 for HTML5 Video by kierank on Tuesday April 13, 2010 @01:47AM (#1261249)
No charges for free web streams until 2015.
Re:Google to Open-source VP8 for HTML5 Video by nailer on Wednesday April 14, 2010 @09:25AM (#1264866)
That's terrible. Imaging 'charges for images on free web streams free for the next four years'. How did that ever become acceptable?