WebGL
No IE support then. I didn't think browsers on mobile supported it either, I know my iPad doesn't (I checked).
Mozilla and Los Angeles, California–based graphics software company Otoy have jointly announced ORBX.js, a new JavaScript library which the companies say can deliver full 1080p, 60fps digital video in a browser window using only web standards–based technologies. "It is a remarkable achievement to see a high performance video …
"Conceptually it is a security nightmare. That was always the given reason by Microsoft for not implementing it."
Ok, I'll bite. Can you explain why decoding one thing on your GPU is worse than any other? I would think that in particular to Microsoft, they never implemented it because DirectX is a nightmare in itself, not too mention it would be similar to incorporating GNU software into their OS. I'm probably wrong, but I think OpenGL was one of the things intentionally crippled (in the ASM) if the CPU was detected as non-Intel...or maybe I'm thinking of something else.
When I try to run WebGL games currently, I have to use Chrome. Firefox just doesn't run some, and doesn't run others as fast. Also Chrome can run some C apps via an extension, which really could be a security nightmare, but it does work. However, being that Firefox is my preferred browser, I do hope the other players start incorporating WebGL to it's full spec. I figure if my GPU is going to render it, then rendering it faster is a good thing.
Ok, I'll bite. Can you explain why decoding one thing on your GPU is worse than any other?
Last time I looked, WebGL basically just exposed the GPU APIs to the world. Can anyone think of any known problems associated with allowing the foetid outpourings of teh intahtoobes unfettered access to the deepest machinations of one's machine? No? Ah, well that's alright then.
To be safe, WebGL really needs vigorous sanitisation of graphics calls built in, but that would mean that things wot run off the web have an inherent performance disadvantage against things wot run locally and is not to be countenanced apparently.
Last time this was raised, the response basically came down to "well it's up to the GPU lads to make sure that their shit is bulletproof". Nothing like sticking one's head in the sand to avoid potential problems, is there?
@TeeCee
If we weren't seeing all browsers being exploited as they are now, I probably wouldn't think twice about forgoing increased browser speed for security, but we are seeing it. Also, outside of WebGL, where are the advancements in this area for hardware? If you want to state that it can be a step in the wrong direction, I wouldn't argue that, but at least it is a step in some direction. I can't think of any advancement that starts off bulletproof, but if bulletproof was a requirement to advancing, I don't think there wouldn't be much advancement. Shit happens right now, and shit will continue to happen, so why not make the best of it? New territory, new problems, new solutions.
Of course, all of these type of talks are being overshadowed by the ever growing monster which is Javascript, so maybe the world won't ever miss something like WebGL or anything like it at all.
Apple has a working implementation of WebGL under iOS — it's enabled for iAds (which are vetted) and can be enabled across the system on jailbroken devices and/or in individual web views through undocumented API calls. Guesses for it not being on by default range from it being insufficiently secure for the main browser (ie, the Microsoft argument) to Apple not wanting to lose app store revenue (ie, the anti-Apple argument).
So, anyway, if killer WebGL apps come along then Apple needn't allow its OS to be left behind.
"Otoy's codec offers a different approach. Instead of DRM, the ORBX codec embeds a unique watermark in each video stream that's keyed to the individual viewer. But if a copy of their specific encode ever shows up on a BitTorrent server, the feds will know exactly where to come knocking."
Surely this unique key would therefore have to be based around their MAC address / IP which can easily be spoofed as if its just a software serial then running several VM would easily get around it
Depends how it's done. A little symbol or number in the corner is easily removed, but what about.. say, a 1% modulation in frame brightness, spreading a 128bit ID over 128 minutes of runtime? It's be far too slow to notice, and endure almost any form of transform. Reencode, flip, rotate, mirror, show the thing on a TV and film it, add some noise, it'll still be there.
Of course you can do this with any codec, but it'd be easier with one where you can adjust brightness in the compressed stream rather than have to encode for every user individually.
DRM, in the sense of making you unable to do what you want, is ALWAYS a less good experience than the pirate version.
Watermarking has its own issues, but is much better as it can be made invisible to the viewer (in the same sense of "acceptably small" which video compression relies on) not to get in the way. By knowing their download is marked, less people are going to share the copy they paid for with others, which is the main goal of a commercial operation.
Of course, mashing up watermarks by using several copies is possible, but potentially hard to do in a way that stops any of the donor's being identified.
Hopefully the content makers will realise that you can't stop piracy, but you can make the paying option cheap enough and good enough (from the customer's point of view) to make the risks of pirating enough to stop all but the most hardened freetard.
Just now I would not hold my breath, as the 'big content' industries have shown themselves to be very dumb in this respect so far.
"By knowing their download is marked, less people are going to share the copy they paid for with others"
So it might stop a few people from handing out copies to friends at work. It will only slow casual copying.
You can pay with a gift card, download in chunks from a coffee shop wifi, use a fake name.
So now you can post it all over the net and Hollywood thinks it was downloaded by I.P. Nightly who lives at Starbucks.
Or they could use infected computers to download a file.
You only need one person with a copy that is not linked to them to uploaded it to the net.
"Hopefully the content makers will realise that you can't stop piracy, but you can make the paying option cheap enough and good enough (from the customer's point of view) to make the risks of pirating enough to stop all but the most hardened freetard."
You do realize that the whole point of this watermarked vid idea is to allow the customer to have a completely open, copy anywhere, backup as many times, view on whatever experience and is only meant to stop mass-sharing of the content (e.g. torrents, et al)? There are likely ways around it, such as if the watermark is some digital bits in the stream, doing a screen capture instead of pulling the raw data (or simply filtering out the bits or replacing them with other acceptable ones if it works like a software key...). The previous comment of embedding it as random one-off noise in the film, such as brightness, is a smarter idea, depending on the resiliency of being able to snatch the ID from a suitably short enough clip (there were comments of mashups to produce the whole length). Now, the download with a gift card from a coffee shop would need to be addressed, and short of a DNA sample and world-wide registrar, can still be worked around (stolen credit card numbers, etc). So no, as long as there's ways of digitally sharing data, there will be the possibility for piracy. It's just a matter of the level of acceptable mitigation.
"You do realize that the whole point of this watermarked vid idea is to allow the customer to have a completely open"
And what are the chances that content companies would go for that? When we have record company execs saying that copying a track from a CD to an iPod is theft and they want to sell you the same thing over and over.
"It's just a matter of the level of acceptable mitigation."
But this will have almost zero effect on "mass-sharing". It might reduce the number of people giving a copy to friends at work, but for mass-sharing you only need one cracked copy the same as any copy protection system.
Well the point is, no matter what you do regarding DRM and watermarking. It will _always_ affect the legitimate user and never affect the pirate
An invisible watermark does not affect me as a legitimate user. If a whole bunch of material I have purchased shows up pirated and they can identify the source as me, then that can well affect me as a pirate. The files are distinguishable. As well as knowing that they were originally sold to me, they can theoretically find the first time that file was uploaded and from where.
This kind of watermarking is already done in cinemas. I read an interview recently with the producer of "Cloud Atlas" who said that with digital projections, he could identify exactly not only the cinema but the particular screening where a digicam recording of the movie was made. This doesn't prevent the copying, but could be used as evidence in the unlikely event that the perpetrator was caught. The soundtrack is similarly watermarked.
If nobody notices the watermarks in the digital projections in cinemas, then why should it be noticeable on a computer screen?
Not subtle enough and it turns consumers away, too subtle and a simple reencoding can corrupt or strip out the watermark info.
It's an interesting idea, and certainly better than the full DRM options, but it's still security by obscurity, which just doesn't last (see AACS and others).
@suricou raven
Can't see that happening for a few reasons, firstly because they would need the entire file to match against and it would be too easy to defeat by splicing copies together at the cutscenes, and secondly because if this takes off it's going to be served from a worldwide CDN and continually adding different noise to each copy as it's served is going to get very expensive very quickly.
It's more likely implemented as a subtle marker containing the entire code at shortish intervals, which would also allow identification of short clips, like someone spltting the content up into 10 min youtube videos for example.
It's also going to be quickly bypassed by the pirates, in much the same way that Time Coded review copies are similarly ID encoded but are all over the download sites regardless.
Some more info about the codec would be nice but I seriously doubt it's competative with H265 if they're comparing it to H264 and it was designed with javascript decomression in mind, I suspect this is going to have less of an impact than the video codec Google opensourced.
Amazing, if they can do this, then why would you even need the bloat of a browser client-side? You could push sophisticated client/server type apps to the dumbest of dumb terminals then, maybe just a screen with some input devices...I don't know how this will lead to a world takeover, but I stand ready to welcome our new overlords.
Not sure this would work. Compression will be lossy and you might well have to look into the future to find out how to compress the stuff. Expect jumping mousepointers and blurry text, not to mention the 2 x 150ms round-trip delay inserted into your cybernetic loop.
The only problem will be that pirates, the real ones who actually make money at this sort of thing, will then use trojans to steal your tagged content and distribute that so then the industry goes after you. If the pirates have removed their tools you'll be out of luck. Granted real pirates are probably harder to come by than the average sharer but it does make the pirate market more profitable once the file sharers are gone.
They will pirate anyway, so it is not helping the content providers and, while it is unlikely to result in the innocent/dumb user being prosecuted, at least they may get their PC fixed and save the rest of the Internet one less zombie.
What helps the case for content providers is the end user gets the pirated-like freedom to view anywhere/anytime but is paying for the privileged, rather than paying to jump through hoops. More likely to get and keep customers that way.
I think they just mean that, in the style of Citrix, OnLive, VNC or a host of others, you could use their streaming to stream a moving picture of a computer program rather than moving pictures of actors. That program would probably be hosted as a virtual machine on the originating server. Which gives them a neat extra buzzword vaguely to attach to their software.
Just another tube that the matrix can stick into your brain. Entertainment 24 hours a day. I have to laugh. People always have these ideas about the noble uses of technology and it always ends up being just another PR delivery and advertising device. People are connected alright - but to what ?
It's just another way to turn everything that is so cool about the Internet ecosystem into c and b grade media streams interspersed with ads. AND encourage humanity to stay asleep at the wheel; continuing to under utilise the amazing tech it could otherwise be saving its own kind (and entire planet's future) with.
Although this idea sounds very consumer-friendly, the potential exists for hackers to reverse engineer the process by examining multiple copies of the video obtained via different accounts. After that they can decode it and distribute the watermark-free version on bit-torrent. Or if they were particularly malevolent, watermark the video again to frame the innocent before distributing it.
Aren't there already apps for that? And I imagine the quality won't be great if you're streamng a video. And since it's a desktop, how do you control it if it's a streaming video? (There's probably a way involving mouse pointer tracking, but it seems really complex and kind of silly)
That's often asked by people who have not seen what you can do in HTML4 if you ignore IE and a lot of other things too. With some simple active server pages you can replace most of your desktop and office with HTML4.
With 'proper' standardised sound and video in there all it needs is a large advertising budget to tell people what it can do and you will see why certain companies want to patent the bleeding obvious all the time - it not to protect their IP, its because they are the modern Luddites.
but what runs the browser if not an operating system?
There seems to be a modern obsession with trying to convince people that your web apps will replace their whole operating system.
Firstly, people don't know what an operating system is, so its not hard to convince them that it is what it isn't.
Secondly, everyone subconsciously hates Microsoft, so why don't they just fucking say it: "our new browser lets you ditch Microsoft"?????? That would be a much stronger selling point than this disingenuous 'look at our new OS' crap.
Most modern AV file formats have lots of stuff in them that can be "adjusted" to put a recurring very long pseudo random ID code in it, either video or audio (or both).
And if this is yet another reason to not use Internet Explorer I'm not exactly going to be shedding a tear about that either.
Who would have guess an interpreted language like ecmascript would be fast enough to do this?
"By comparison, the new library is based on Otoy's proprietary ORBX codec, which was built from scratch to be easier to implement in a browser environment."
I'm a little confused on this one. Aren't Mozilla the ones who are trumpeting everywhere how they are keeping the Internet FREE for all of us? Now all of a sudden they are promoting a proprietary codec as their next big thing? Or "proprietary" is fine - as long as *they* are the ones who put it forward in the first place? I'm not saying that we might not have to live with proprietary codecs for some reason or another - but where does the above fit with what Mozilla supposedly stood for all these years? Who's going to rake in the revenues from content providers in licensing fees for using this codec to generate material? And who's going to get a cut from it for promoting it on their platform? Oh, I forgot - Mozilla already gets a cut from Google for selling us to them and towing their line.
I saw the first iteration of this (Java driven video) around 1999. No, it wasn't high resolution anything, but it WAS video that ran in ver. 4 browsers without using a plug-in and that alone was miraculous.
At that time, Quicktime was the only real player in video-over-Interent, and it flat out sucked. Windows had its AV player but it sucked even more. Flash was still just fancy animation. So Quicktime was your only real choice.
Being free of those 2 piss poor players was truly and eye opener. I believe, eventually, that Java trick was later incorporated into Flash. (not completely sure. many moons and pints ago)
Good to see the concept making a comeback. Hopefully we won't have to wait for HTML 5 in order for it make into the wild.
Except that Javascript libraries are compiled onto every chipset out there - as opposed to being tied to a toolchain that hasn't been compiled to less-used CPUs.
Ask PowerPC-based Mac owners who were left out in the cold when Adobe dumped that CPU's support after Flash 10.0, then the Beeb updated iplayer to need 10.2.