|Google Gearing Up Antipiracy Tools for YouTube "very soon"|
|Google, keen to head off a media industry backlash over its video Web site YouTube, is to offer anti-piracy technologies to help all copyright holders filter unauthorised video sharing. |
'We are definitely committed to (offering copyright protection technologies),' Google Chief Executive Officer Eric Schmidt said in an interview. 'It is one of the company's highest priorities,' he said.
'We just reviewed that (issue) about an hour ago,' Schmidt told Reuters when asked what Google was doing to make anti-piracy technologies widely available to video owners. 'It is going to roll out very soon ... It is not far away.'
Google Gearing Up Antipiracy Tools for YouTube [news.zdnet.com]
I wonder if their "tool" is as simple as a button on each video to report it as a copyright violation, to flag it for further inspection. Now that would keep them busy day and night.
Seriously I don't see how they can do it in an automated way but we'll see soon I guess.
|Seriously I don't see how they can do it in an automated way but we'll see soon I guess. |
|The company, which Google bought in November last year for $1.65bn, was currently working on "audio fingerprinting" technologies to identify copyrighted material. |
It's gonna have to be some pretty darn clever technology to make it work, but I guess if anyone can do it, with their arsenal of talent, they can.
Sinners morphing into saints.
Google to the copyright recue!
Kinda like your friendly slasher selling band aids.
I always think of Youtube as a viral video site, nothing much to do with "user generated" content as much as "user submitted" content.
sure lots of the videos are generated for popularity on youtube, but is that crap worth paying for? THe good stuff is already well known outside the "viral" views of youtube.
If it is automated, you can bet your bottom dollar that there will be SCADS of false positives, resulting in people's videos being whacked left, right and center. You can also bet that the costs for "un-whacking" your video will go through the ceiling, so their content will drop like a rock once word gets around. Look for changes to their Ts and Cs that will indemnify them from this.
Google + YouTube = :(
Ah audio fingerprints... perhaps for future videos, broadcasters will add a tone.
(and then within a few weeks utilities will come out to remove the tone)
How many sites does google own that aren't "officially" associated with the google juggernaught we know today?
I mean, even google can't dominate every sector they touch.
edit: The worlds best content will always come from the people, I applaud google for trying to ensure that the little guy will retain full rights to his stuff... but it seems to be a conflict of interest to want to use that same content for company gain too on sites like Youtube. Watching with interest.
[edited by: Kurgano at 6:21 am (utc) on Feb. 23, 2007]
If it's automated it would have to be to do with the titles of the videos, not the actual video content. There's no way that even Google could compare every video submitted with copies of every copyright film and TV series ever made.
-- I always think of Youtube as a viral video site, nothing much to do with "user generated" content as much as "user submitted" content.--
The user generated stuff is just a smokescreen. If they removed all the original material, the number of pageviews would barely change because almost everyone is there to watch the copyrighted stuff.
YouTube doesn't make anyone famous, it generates its traffic through pirated versions of shows that are already famous.
I'm not sure they can recognise videos - yet - but audiofingerprinting, in principle, is quite straightforward; if you consider audio as a kind of two-dimensional signal; it can be visually represented as a wave form, so dupe recognition should be technically within reach.
Video ... another ball game; the '3D' equivalent, I guess? I'm sure there's ways of watermarking, but how that will survive reproduction and transmission by mobile phone is anyone's guess!
But - and this is what matters - the stuff everone is so rattled about, copyright a-v items, has an audiotrack that can easily be traced.
Yes, they could go silent. I don't think so!
Either way, I don't think it'll kill youtube; from the little I've seen of it (not my cuppa tea!), the 'core' of it is DIY home video stuff. People who want to lift broadcast items just help themselves, don't they?
And increasingly, companies will sign deals to allow at least some stuff; you'll see!
Like it or lump it, it's here to stay, and it will pay for itself in spades.
[edited by: Quadrille at 2:41 am (utc) on Feb. 24, 2007]
--Either way, I don't think it'll kill youtube; from the little I've seen of it (not my cuppa tea!), the 'core' of it is DIY home video stuff.--
The front page is full of DIY home video stuff, but don't let that fool you. YouTube choose what goes on the front, and don't want to flaunt the fact that almost all the material their visitors watch is pirated.
The core of YouTube is copyright material distributed for free, and in a manner that's so easy and safe that it will make practically anyone a pirate. Even people who avoided stuff like Kazaa because they didn't like the idea or didn't understand how it worked are sucked in to YouTube's piracy service, it's just too darn easy to watch almost anything you want to: TV series, movies, music videos etc.
--I'm not sure they can recognise videos - yet - but audiofingerprinting, in principle, is quite straightforward; if you consider audio as a kind of two-dimensional signal; it can be visually represented as a wave form, so dupe recognition should be technically within reach.--
It's not the technology that's the biggest problem, it's the sheer scale of what they're expected to do: remove all copyright material.
Even if just one large organisation such as the BBC asked for its copyright material to be removed, their total video and audio output over the past 50 years is enormous.
There's no way Google or anyone else could compare every single file on YouTube to every programme the BBC has ever made, whether it's done by comparing audio or video.
Add in more major broadcasters and you have a task that will probably be impossible for many decades to come.
The only way they can police this is by text searches and, maybe, by concentrating on a very small number of copyright films and TV series. Whatever they do, the vast majority of copyright material cannot be removed in an automated manner.
You may be right; I'm no techie, and I have no knowledge of the requirements of an audio search engine.
But if Google can search the Internet and spot duplicate content pretty effectively, why could not an 'G-audio Search Engine' search the web - with YouTube as comparator - for duplicate audio fingerprints?
I can fully understand that setting up such a scheme would cost a few dollars, but on the face of it, once set up, it would be just one more SE, just configured to look only at certain file types.
Provided copyright owners give Google samples of what they want removed, I'm sure it's do-able; I'd be interested to hear about cache sizes and technical objections - but it seems to me to be 'do-able', provided organizations who want stuff removed are willing to provide the basis to do that.
Once there's an 'industry standard' audio finger-printing method, then it's just a matter of time. And whatever it costs, it'll be cheaper than a scheme to 'Report, Review and Remove' manually.
And make no mistake; now that its so easy to make copies of other people's material, the problem will not go away, and if the owners think attacking YouTube solves the problem, they are insane. But of course they are not. The majority are playing brinkmanship, like the Belgian Newspapers, for cash.
We've had all this 'Web 2.0' BS for a few years now - But THIS is really the threshold of the new web; policing users who can do things that were unanticipated a few years ago.
Napster caught the music industry on the hop; so far, neither audio or video producers seem to have learned anything; music industry share prices reflect their sad faith in copyright protection software; if they don't get real, the video owners will be left behind as well.
Realistically, the ONLY way forward is by working with Google (and other video sharing sites); and if they don't pick up that message, they're doomed. Betcha.
If they can put a tone on it to recognise it. Even if they got over whole TV producers etc. having to make it for ever single bit of video, it would be simple to add a one over the top with a program, thus screwing it up.
Plus. If they got real tight on piracy... People would use something else. Just ask Kazaa and Napster.
Ha, apparently these tools work and are in effect.
Two videos which I had marked private and showed only to select people got DMCA whacked. I have no clue how they found them, I had no useful description or filenames.
I still don't understand how they do it though.
Tell us more!
I'm not surprised though; Google knew all along that there's no way they could depend on the surrounding text - six years of dealing with spammers has taught them something!
I'm not getting at you; just a comment on how what they learn in SEO they have applied in online video.
But how did it occur, and what actually happened?
Not being a snot, but anything I had using copyrighted materials (such as a photo slideshow with an unlicensed soundtrack), that I only wanted to share to one or two people, would be hosted on my own server, and would have some kinds of access restrictions. These could be as simple as a private, unlinked URI, or a passworded directory.
I would never use YouTube or any other hoster for something like that.