next »

[nop_90]

http://www.webmproject.org/about/
also are adopting matroska format.

Taking on h264

[isthisthingon]

I believe Steve Jobs swore to destroy Ogg by constructing discovering software patents that kill it and all other non-Apple approved standards ASAP 

[nop_90]

A very smart tactical move by google.
Again like the android, it does not matter that it is a POS, but that it is cutting into Apple's profits, market share.
From all accounts VP8 codec is a POS compared to h264, but high profile videos on you tube will be in VP8 format, and videos ipad users see will be low res h264.
So google shells out like a few million for android, and a few million for VP8.

The Steve Jobs gets all paranoid, and sues.
http://www.vanityfair.com/online/daily/2010/03/steve-jobs-and-rupert-murdoch-lets-sue-the-internet.html
Whether he wins or loses, it does not matter, Google has tied up apple, and for a few million is cutting in apple.
Also again it makes Steve Jobs look like the insane lunatic which he really is .

You see, it is a minority attack.
http://chess-training.blogspot.com/2007/08/strategic-planning-minority-attack.htm
The object of the minority attack is not to win, but just to isolate the other guys pawns, and then pick them off 

Steve Jobs meanwhile is hiding in his bunker, ranting and raving, while his apple fanbois goose step around claiming to be the superior race.
Well we all know how that one will end up

[nutballs]

Meh 264 is so entrenched at this point, and it is apparently superior, there is no way that it will change anything any time soon in a general sense. However, they may have a very specific plan for it, since it may excel at a specific delivery system. Like small screen phones for example, orsomthing like that.

Who knows, but a little weird to me.

[kurdt]

For once I like what google is doing. I really dislike current video situation. The situation is really nightmarish for anyone who's planning on doing a website that offers video about something. Firefox won't support h264 and Safari will support only h264. So with VP8 we at least have Chrome, Opera and Firefox. That leaves Internet Explorer and Safari. Internet Explorer holds about 30-40% share at the moment. So IE gets behind VP8, we will have a winner in the web but there's the problem with mobile devices still.

When you start thinking about it, I have to hand it to Apple, they really outsmarted us this time. Almost all mobile devices play only h264 and not Theora or VP8. And after 2015 you'll have to pay minimum $2500 per year if you have more than 100000 eyeballs per year and $10000 if you exceed 1000000. And if you believe predictions, mobile devices will take over as primary net usage devices by 2012.

Oh what doom & gloom.

[Phaëton]

The race is on to deliver the internet experience.. however its dealt out..
Perpetuating the tower of babel, forcing the cream to rise to the top..

http://www.brisbanetimes.com.au/digital-life/hometech/google-premiers-web-tv-20100521-vvvq.html


resistance is futile.

[kurdt]

The race is on to deliver the internet experience.. however its dealt out..
Perpetuating the tower of babel, forcing the cream to rise to the top..

http://www.brisbanetimes.com.au/digital-life/hometech/google-premiers-web-tv-20100521-vvvq.html

resistance is futile.

First I thought this was Onion.com news Everybody including Apple has tried this and it didn't work. Why? There's few very good reasons. First of all, every who has used AppleTV knows how much it sucks to write with remote. It just doesn't work. And when the whole basic premise is the "here's a search box, type what you want" it'll get painful pretty fast. And to use Android as remote.. yeah, you gotta buy Android phone first so big majority of the people are not going to have one. I mean this is just screams WE WANT TO SHOW YOU ADS ON YOUR TELEVISION.

Second problem is channels. I mean what channels? You mean YouTube channels? Those totally useless things where somebody puts their own videos and quality ranges from pure shit to fairly funny? Yeah, that's what I want to watch from TV. So many TVs already have YouTube built-in that you can watch as many cat videos as you want already. Why hasn't WebTV already happened? Well, for one there's no webshows that can compete with quality TV series. Though all fan related podcasts are actually better quality in many times than what you can see on TV. But it's still series what people watch from TVs and news. Unfortunately it's going to take A LONG time before WebTV can pay same amount that what regular premium cable companies can per season.

This is just another way of Google to show more ads. I'm getting more and more disgussed how all Google does nowadays is to get more adviews and how they try to potray themselves as some sort of innovative company bringing public service. Google is a snake oil salesman that preys on clueless general public with free services that are not free and they harvest their personal data without telling and asking explicit permission. All this with excuse "you don't have to use our services if you don't want".

[kurdt]

If you are thinking about using WebM, you might want to read these articles:
http://x264dev.multimedia.cx/?p=377
http://www.appleinsider.com/articles/10/05/20/steve_jobs_says_no_to_googles_vp8_webm_codec.html

First one is x264 dev's look into VP8 and he demonstrates how VP8 is actually really shitty codec that doesn't hold up to even VC-1. Second one despite the site is actually very refreshing view at the whole h264 closed patent ridden boohoo thing going on. And it explains quite clearly why WebM/VP8 is a crappy choice for any site except those who are going to serve videos only to people who use computers like laptop or desktop. VP8/WebM is not suitable for mobile or high-definition encoding. For me that's pretty much it for VP8/WebM, I find it completely useless for any purpose right now.

Personally I'm starting to get behind h264. If you think about those license fees, it's kinda stupid to cry about paying $10000 for +1000000 views. If you can't get $10000 out of 1000000 million views, then there's something wrong with your business model. Serving +1000000 videos eats so much bandwidth you just have to have some sort of income. In my opinion it's time for people to wake the fuck up and realize that Twitter-like "ok let's do this and we'll figure out how to make money someday" is just not working and you can't build a business if you don't know how you are going to make money. This used to be quite clear for everybody in the real-world but for some reason in internet everybody expects there's no costs at all.

[isthisthingon]

"Jobs' problem with VP8 is not ideological"   

Nuff said.  Gotta love the Apple Insider.  And people complain about the utter irrelevance of PCWorld 

[kurdt]

"Jobs' problem with VP8 is not ideological"   

Nuff said.  Gotta love the Apple Insider.  And people complain about the utter irrelevance of PCWorld 

Well if you look at VP8 and just look at the codec, it's pretty clear that it wouldn't make any sense for Apple to support it. It doesn't work for mobile or high-definition encoding. So if Apple doesn't want to support Flash because it drains the battery, why would they support a codec that doesn't offer any improvement over Flash? If you want to do whatever you want no matter how it will kill your device and ruin the user-experience, get an Android device.

Like I said, you are doing yourself a disfavor if you don't read the article and just cling on the fact that it's AppleInsider who wrote it. It's good stuff and brings back little reality into this whole web video battle.

[nop_90]

http://www.perkiset.org/forum/empty-t2843.0.html;topicseen
I accidentally split topics.

One thing that people forgot to mention including myself is the container format.
http://en.wikipedia.org/wiki/Comparison_of_container_formats

For webm they will be using the matroska container format.
The idea behind matroska is very interesting, it uses ebml which is kinda like a binary xml.
So you could use ebml for other things. I had forgotten ebml, maybe i will make a python pickle with it
So it is very flexable. But with very little overhead.
Matroska is designed to use streaming.

Also all codecs right now are using DCT to do the actual encoding. DCT is a relative of FFT and caused blockiness, which h264 etc overcome with filters.
For still images, the compression that is the best use wavelets, ie jpeg2000 etc.

I have not looked at wavelet codecs recently, but there are 2 that are actually functional, snow and dirac, dirac is sponsored by BBC.
Biggest problem right now with dirac for example, is that it requires too much computation power to decompress at realtime.
But 5-10 years from now, that will not be a problem.
With matroska if google decides to use codec XYZ, matroska can handle it, with MP4, you are fuked plain and simple.

[kurdt]

For webm they will be using the matroska container format.
The idea behind matroska is very interesting, it uses ebml which is kinda like a binary xml.
So you could use ebml for other things. I had forgotten ebml, maybe i will make a python pickle with it
So it is very flexable. But with very little overhead.
Matroska is designed to use streaming.

You can actually get mp4 to stream but it's a hack. Matroska is indeed a great container choice.

Also all codecs right now are using DCT to do the actual encoding. DCT is a relative of FFT and caused blockiness, which h264 etc overcome with filters.
For still images, the compression that is the best use wavelets, ie jpeg2000 etc.

I have not looked at wavelet codecs recently, but there are 2 that are actually functional, snow and dirac, dirac is sponsored by BBC.
Biggest problem right now with dirac for example, is that it requires too much computation power to decompress at realtime.
But 5-10 years from now, that will not be a problem.
With matroska if google decides to use codec XYZ, matroska can handle it, with MP4, you are fuked plain and simple.

I think biggest problem with Dirac is that it looks like shit compared to h264. Now why would Google's decision matter in any way? As long as there's hardware support for h264 and no hardware support for codec XYZ in common devices, h264 wins. You can of course argue all day long about which is the best codec technically but none of that matters when your job as service provider is to make sure your service can be enjoyed as big audience as possible.

You can argue that in 5-10 years we will have next gen devices and hardware in our hands but that doesn't in anyway help it because you still have A LOT of people who can only watch h264 encoded stuff. It's like mpeg2, even it's shit compared to current gen codecs you still have to support it because so much of video encoder software exports only to mpeg2.

[nop_90]

You can argue that in 5-10 years we will have next gen devices and hardware in our hands but that doesn't in anyway help it because you still have A LOT of people who can only watch h264 encoded stuff. It's like mpeg2, even it's shit compared to current gen codecs you still have to support it because so much of video encoder software exports only to mpeg2.

There is strategy and tactics.
Apple has good tactics but poor strategy.
Google and M$ do not have as good tactics as Apple, but they have good strategy.

Intel had excellent strategy when they switched from memory chips to processors.
For example when intel was selling thier first processor, cost to make was like $20 per chip, but they sold it for 1/5 that price and signed long range tactics.
This was a huge gamble, but strategically very smart. They counted on that production costs will drop, and within a year they where making a profit.

For ~$125 million, google gets a chance to hurt apple.
Android is the same BS, it does not matter if it is a POS, just as long as it cuts into apple sales.

The big thing of the future is streaming of video (or this is a pretty good assumption).
Infrastructure for greater bandwidth will take longer to achieve. It is much more expensive to put up more cell towers etc.
While computing power of the client will get cheaper, so if codec XYZ can stream at 5k/s vs h264 at 20k/s even if it takes 10x the computing power who ever has codec XYZ wins.

Strategy is like playing chess. A human player can beat a computer, because he understands strategy. As in certain key squares control the board.
The human, unlike the computer can not see ahead. If is impossible

Fuking apple,google,M$ etc are all opportunistic,self serving bastards with their own agenda to take over the world
But it is like watching grand masters playing chess. And you have to appreciate a good move

[kurdt]

Yeah, I agree with you. The problem with VP8 is that it doesn't work so well for low bitrate encoding as well, it's strictly web video which means your basic YouTube. Actually when you read more about VP8, it starts to look really useless. In every use case there's better alternatives available. And don't forget, using h264 doesn't cost you a dime unless you have A LOT of eyeballs. The future is in the mobile devices so only way for Google to actually make a dent is to do a killer hardware decoder for VP8 but wait, it still doesn't help the fact that VP8 sucks at low bitrate encoding. Now what's going to happen if there's two video sites where other one doesn't eat your battery and provide great quality video and other one eats your battery and provides bad quality? With all the bots going around, soon there won't be a video that's not in every video hub in the internet once submitted to one.

In my opinion Google is the biggest hot air seller right now in the market and when competition rises, companies are happy to dump Google as soon as possible. I can't be the only one who sees how lame Google products are. I mean just look at what Google has been doing. They are pissing in every direction buying this and that. They bought a shitty codec which was suppose to be patent-free. They probably studied it and realized that it's probably not patent-free and it was a bad investment. Then they thought that hey, what if we could sell this as patent-free codec to the masses and appear that we have a solution to this whole <video> tag battle. This would give us some good PR and we paid only few million for it. Sounds good, let's do that. If somebody finds a patent, then we can just say "our codec was given to you for free, you had every choice to use something else". In reality what you got is a time bomb just like GIF and Google knows this. They specifically say in their license that they do not take any responsibility if you get sued for patent infringement.

If any company starts to use VP8, I wouldn't want to do business with them. They obviously don't read the fine print, take stupid chances or are stupid enough not to verify the claims in the marketing pitch.

[isthisthingon]

  

Search this page for 0p3n 50urc3, no wait, just search the page for the word 0pen and see something very interesting.  It doesn't appear anywhere except in the title of this post.  In one sense you might think well of course it doesn't ITTO, hardly anyone really cares about that aspect of things - the freedom "angle."  But nop you posted a great article about how Apple has become like China which hit on something I think very relevant here.  The argument has to be between pragmatism and pragmatism.  It can never be about personal freedoms, liberty and license lock-in. 

kurdt claiming that the extremely cheap rate for the use of h264 perfectly misses the point, much like relying on the hope that Microsoft will not pull the trigger on any of their ridiculous patents that claim, for example, the rights to the invention of breathing with your lungs while using an OS.

It's the support of patented and proprietary technologies that needs to be avoided - if possible.  By avoiding this you have already won the largest part of the battle, imo.  After that comes the sea of subjectivity regarding user experience etc.

But I realize I'm in the extreme minority on this one.