MySpace Begins Rollout of Video Monitoring Tech

C|Net is carrying an article looking into new technology MySpace is rolling out to combat user violation of copyright laws on their pages. Called 'Take Down, Stay Down', the service will attempt to ensure that once content is removed because of a complaint it can never be uploaded again. "Copyright owners have access to Take Down Stay Down free of charge, according to a release from MySpace. If the social-networking service receives a takedown notice regarding a copyrighted clip hosted through its MySpace Videos hosting service, MySpace's new feature will take a 'digital fingerprint' of the video and add it to a copyright filter that blocks the content from being uploaded again. '(It's) the ability to have a piece of content imprinted and put in a database so we can identify it,' said Vance Ikezoye, CEO of Audible Magic." The article goes on to discuss the problems YouTube is facing with the same issues, as well as recent investigations of this issue in the political arena.

Comments

Bogus take-downs.

[Anonymous Coward]

Are they going to have a mechanism to fix bogus take-downs? Will it be effective? Could someone with an SCO mentality take down half the site?

Re:Bogus take-downs.

[jZnat]

If MySpace is responding to DMCA takedown notices, anyone who fakes them would be violating the DMCA anyhow, but I guess the avenues for abuse of anything are beyond MySpace (which is why so many JavaScript-based worms hit that site).

The search for the Holy Grail

[janrinok]

I would expect that, no matter what technology is used, a simple trivial change to the size of the file will negate the fingerprint technology employed. This appears to be another search for the Holy Grail that will probably be as unsuccessful as previous attempts.

Re:The search for the Holy Grail

[Travoltus]

The workload this'll put on their computers will bring MySpace to its knees.

Re:The search for the Holy Grail

[femto]

Why bother? Just use another website.

Every website has a finite time in the sun, and RupertSpace is no exception. There will even be a day (soon?) when people say "Now what was that website called, TheirSpace or something?"

Who remembers AltaVista? It was as big as Google and RupertSpace in its day. Everyone thought it was the ultimate answer and unassailable when it "indexed the whole Internet". As altavista faded so will RupertSpace (and even Google eventually) and their restrictions on users will be irrelevant.

Re:The search for the Holy Grail

[owlnation]

Why bother? Just use another website.

Very true. It strikes me as a curious irony that anyone with a conscience would continue to use MySpace after the News Corp purchase anyway. You have young bands trying to express themselves free from the record industry parasites, and yet News Corp is indirectly linked to the RIAA through it's MPAA connections.

Makes no sense. The News Corp purchase should really have been the instant death of MySpace cool.

Re:The search for the Holy Grail

[jamar0303]

Yep- all the people at my school use Facebook now.

Re:The search for the Holy Grail

[Tuoqui]

Well if they are using MD5 SHA1 or SHA256 hashes for their 'digital fingerprinting' like they do for Computer Forensics dealing with files and/or entire drives.

All you would need to do to defeat their 'digital fingerprinting technology' would be to change a single bit in a file to something else then the odds of getting the same hash would be astronomical.

So yeah, I'm thinking its something along the lines of a publicity stunt by Myspace to look like they are actually doing something. Also they are required to obey the DMCA too which indicates that if someone issues a counter DMCA claim that they have to permit the content to be restored somehow and/or taken off of this 'blacklist'.

Re:The search for the Holy Grail

[J'raxis]

They do have to honor put-back requests, but there's a 10-14 day delay required by the law in question before taken down material can be put back. This gives people who want to have time-sensitive material removed from the site (e.g., content that could affect an election, or a protest, or somesuch, a couple days from now) a very big advantage.

Re:The search for the Holy Grail

[Jah-Wren Ryel]

Makes no sense. The News Corp purchase should really have been the instant death of MySpace cool. It makes plenty of sense when you take into account the fact the people who most need to care about myspace's MAFIAA connections are the least likely to know about them.

Young musicians have been signing recording contracts without doing their due diligence for decades, posting to myspace is 10000x more common and easy than that. If they can't do it for the big recording contract, they sure won't do it to use a "free" website.

Re:The search for the Holy Grail

[HAKdragon]

What percentage of the people on MySpace even know what NewsCorp is? I'd wager it's in the single digits, probably less than 5%. One has to remember that most people don't know about the evils of the MPAA, RIAA, NewsCorp, etc, assuming they've even heard of them.

Re:The search for the Holy Grail

[Anonymous McCartneyf]

I remember AltaVista. I miss it.

Re:The search for the Holy Grail

[Anonymous McCartneyf]

If MySpace takes down a file and every file matching it with this new software, and whoever put the first file up is successful with his put-back request, then will MySpace put everything that matches back as well? Or will everyone have to make a put-back request of his own (inc. people who are DMCA-illiterate) to get the matching files back?

Re:The search for the Holy Grail

[jZnat]

MySpace is already on its knees blowing the content industry (and due to all those "Technical Errors" so common to that hack of a website), so I guess they're going to fully collapse under the computation of video fingerprints, eh? I would hope so just so that something better than MySpace can fully take its place and house all of the world's emos in one centralised location without the need for other websites to offload them.

Re:The search for the Holy Grail

[Anonymous Coward]

I remember UUCP maps... I miss it.

Re:The search for the Holy Grail

[Anonymous Coward]

I have this discussion with my best friend from time to time -- his band has a popular myspace page, and he hates Rupert Murdoch almost as much as anyone else does. He also maintains a full web site hosted with a local, generally pretty decent, company.

There are several factors: his band pays for the hosted website traffic, but not for the myspace downloads; they are not technical wizards and while they're collectively pretty good at flash and design and pretty images, they aren't prepared to run their own forum, weed their own mailing list, or take the risk of being accused of sending UCE. Myspace lets them make announcements more cheaply and with less hassle, and all it really costs them is some internal inconsistency (we all have that though, don't we?) and a lot of time weeding out the spam from their myspace inbox (they get three myspace email pages worth of "please add me as a friend" spam-robots (and other, often commercial, bands) just about every day).

In short, myspace lets them send announcements to 12000 (!) people who have taken the trouble to add them as a "friend".

The contractual risks associated with doing this on myspace are negligible. (They worry more that the estate of a dead songwriter might come and demand royalties on one of their two or three cover tunes, or that the tax authorities catch up with them and send an accurate bill).

For the feeling of a bit "unclean", they just about have a forthcoming tour paid for in advance, thanks to the reach that myspace gives them. That's pretty damn good, although they still do an awful lot of work -- the fate of a DIY band -- and if it were me I'd consider selling out to the first management company that came along and offered to take care of all the annoying last minute details (venue problems, cancellations, merch logistics, and so on) for a fee. But then I can't play an instrument, so that's not really an issue.

hacking call to arms

[plierhead]

I guess I've never really understood the hacker ethic but reading this it sounds almost like some kind of challenge...

And then there's the issue of how well it will actually work, or if MySpace is just trying to save face. Cynical observers might be quick to ask how soon clever hackers will figure out a way to work around or even completely circumvent the "digital fingerprints." According to Ikezoye, it's not going to be easy because Audio Magic's patented technology is more complicated than simply generating a "hash value" for a file.

"A fingerprint is much more robust at identifying the content. Hashes identify files," he explained. If a Colbert Report clip were pulled at Viacom's request, for example, MySpace's filter would block all other forms of the file from MPEG to AVI, all various degrees of quality, and even video clips that contained only part of the content from the piece that had been taken down. "We simulate the human perception of the same content," Ikezoye said.

And to circumvent the filter, he added, a hacker would have to "screw up the content itself so it wasn't recognizable," to a degree where it wouldn't even be worth uploading in the first place.

But given hackers' long history of being able to get through just about anything, experts remain a bit skeptical.

Re:hacking call to arms

[Anonymous Coward]

I know how to get past it!

Post the video upside-down!

Re:hacking call to arms

[Animaether]

fwiw - no, it catches that.

Most video fingerprinting technology can deal with mirroring, rotating, shearing, compression, time stretching, channel swapping (RGB and YUV), and practically any other method you can think of. One thing you -can- is chop it up into fragments that are smaller than the watermarking window, and distribute those fragments across the canvas randomly. The problem with that (and, actually, almost all of the aforementioned methods) is that the video is unwatchable and compresses horribly. The latter you're stuck with, the former you could code a special plugin for that unscrambles the video. But at that point, it's no longer video for the masses.

I'm sure other posters have already pointed out that people will just upload/download on a different service, etc. so I won't go into that here.. nor the possible (unlikely, but possible) 'fair use' issues.

Re:hacking call to arms

[Anonymous Coward]

And to circumvent the filter, he added, a hacker would have to "screw up the content itself so it wasn't recognizable," to a degree where it wouldn't even be worth uploading in the first place.

Or just edit the file slightly with a hex editor

Re:hacking call to arms

[Lehk228]

could probably get around it by playing with the signal in CMYK, keep the black channel mostly undistorted while making everything else rainbow colors.

Re:hacking call to arms

[Blue Stone]

>If a Colbert Report clip were pulled at Viacom's request, for example, MySpace's filter would block all other forms of the file from MPEG to AVI, all various degrees of quality, and even video clips that contained only part of the content from the piece that had been taken down.

Translation: In Rupert'sSpace there is no Fair Use.

Re:hacking call to arms

[OECD]

...and even video clips that contained only part of the content from the piece that had been taken down.

There's your "unlikely, but possible" fair use issues.

Re:hacking call to arms

[Anonymous Coward]

And to circumvent the filter, he added, a hacker would have to "screw up the content itself so it wasn't recognizable," to a degree where it wouldn't even be worth uploading in the first place.

In other words, watch it on Vista.

Re:hacking call to arms

[Anonymous McCartneyf]

"Solarized" videos all over MySpace? Sounds like fun!

Re:hacking call to arms

[Animaether]

that's the possible - the question is how likely it is still to be. Keep in mind that the fingerprint isn't the -only- bit to check against. There's things like duration that they can take into account. For example, if you make a video of 4 minutes of a Colbert Report section and just tack a "ROFL" frame before and a "watch my other videos!" thing at the end - that's not fair use. If it's a 10 second clip inside a 4 minute video you make - it probably is fair use. Obviously, there's plenty of room for "I'm not sure if it is, or isn't". At that point, video services could investigate for themselves, let the copyright holder investigate / whatever. That's well outside the scope of fingerprint technologies and what they can or cannot do.

To answer the obvious question - if it's not reasonably certain, then content should never be taken down automatically.. not even as a precaution. if nothing else, write to the user first, ask them if *they* believe it is fair use or not, as to give them the opportunity to take it down themselves if they figure it quite probably isn't. If they do, then they can indicate that, and the video site can take it up with the copyright / license / whatever holder and have them either agree or not. It's in all parties' best interest not to mistake it.. i.e. if it is fair use, and the copyright holder believes otherwise, then things like mediators or even lawyers come into play.. they cost money.. money they'd prefer not to spend if it's obviously they're in the wrong. Same applies the other way around, of course.

Re:hacking call to arms

[OECD]

At that point, video services could investigate for themselves,

What makes you think they start now? Heck, it took a user rebellion for Digg to grow a pair and refuse to take down a fraking number.

...let the copyright holder investigate / whatever.

Right. And we can trust the copyright holder to diligently consider the fair use value of things like a parody campaign against the Colbert Report. They're just going to fingerprint everything and have a robot send out notices, just like they've always done. Your 'fair use' will have to be backed up by lawyers, or it ain't going to happen.

Not so simple....

[janrinok]

OK, if it is able to identify 'content' - and does not use a human to provide this function but achieves it entirely in software - then it must take a series of snapshots of the video and use some form of key (or hash?) for each snapshot. That is not a hash for the entire file but a series of hashes which can provide a unique fingerprint. The processing power required to do this, and to subsequently search submissions in an attempt to find a matching hash will be immense. Pattern recognition is improving all the time but it is still nowhere near able to recognise content(i.e. girl dancing in bedroom, skateboarder falling arse over tit etc) with sufficient accuracy to enable PR to be used. If this problem HAS been solved, I would expect the military and other scientific fields to be investing in it far more than a web video hosting site.

Re:Not so simple....

[plierhead]

If this problem HAS been solved, I would expect the military and other scientific fields to be investing in it far more than a web video hosting site.

The military would pay some amount of money for it, and it would take 2 years to cut a deal with them. Google and MySpace on the other hand will lose untold wealth in just a few months - if they don't come up with a way to get the lawyers off their backs.

These smart guys have the right technology at the right time. They're cashing out big time.

Re:Not so simple....

[janrinok]

Perhaps in the US, although I am not convinced, but there is a sizeable investment going on around the world for this sort of technology, particularly for weaponeering. For example, imagine being able to identify someone who is about to fire a rifle as apposed to someone carrying tools around his garden. Or all those cameras in the UK. Being able to identify a crime taking place automatically would be a huge benefit to If the problem of PR has been solved I don't think that your government or mine would have any qualms about using the technology in the name of 'national security' or whatever. Money might change hands at a later date but it wouldn't stop the technology being taken and used.

Re:Not so simple....

[mikael]

They already have. Do a google search for:

image classification
image retrieval
texture classification
texture retrieval

Re:Not so simple....

[Andrew Kismet]

That's a completely different technology - you're talking about object identification. They're just using pattern-recognition (ie. hashing frames and comparing the hashes) and dedicating a CPU or two on the video server to it. It just means videos will go through a validation procedure prior to being fully uploaded. They'll probably just look at whatever keywords you provide, see if they match any in a blacklist (ie. Family Guy, Fox, Simpsons, whatever copyrighted videos), and then use the same flagging system on top of that they've always used.

Re:Not so simple....

[Anonymous McCartneyf]

That could lead to some interesting keywords. I see "pr0n" and "warez" a lot already. We might get "51mp50nz" videos if keyword checks are done before hashes are.

Re:Not so simple....

[Anonymous Coward]

Remember myspace's goal. Most crucially it's NOT to avoid showing videos for which they've recieved DMCA takedown notices. It's to gain the appearence of a reasonable attempt at taking down those videos. All they want is to cover their asses in court. This doesn't mean that the technology has to work, it just needs to appear to work well enough, or be arguably the best that they could reasonably be expected to deploy, and their asses are covered. Basically, it's a all a show, the fact that we can easilly change certain low order bits of an mpeg, and, since it's in frequency domain, the quality would only drop minimally, or add a millisecond of dead time at the end doesn't matter to them.

Image recognition is actually fairly easy. A few gabor filters run over the image, take the maximums and minimums, store their geometric information, and you're good to go. This is feasable, even at reasonable framerates actually, but the amount of processing power required is ludacrous for a large website. They most likely can't spend 5 minutes on a high end machine processing a 5 minute video for upload, that would be crazy. And one would still have to establish just HOW similar counts as the same video. I find it highly improbable that they're planning to roll out this type of softare. It's hard, and tweeking it is harder. They'll probably just md5sum the file and be done with it, or if they want to get fancy, md5sum once per 10 second interval so they can pick out pieces that are identicle (this should work with mpeg on the raw file format, just round to and lock in on the full update frames). Then they just have a big library of forbidden md5sums and when someone uploads they hash and check. Given that this library is only a few million entries or so it can easilly be duplicated on every server. DMCA takedowns give a reasonable amount of time, so if they have any sort of automatic update system, or they load the machines off network attached storage, whatever, they don't even have to deal with syncronizing updates to this library, just use a flat file in the network attached storage, or if they're system is REALLY hoarked cycle reboot every system one at a time automatically every couple of days (I.E. one or two machines rebooting at all times).

Hmmm....

[evilgrug]

The bold claims made in this article make me even more skeptical. If the technology could really identify 'content' in the way they describe -- handling different forms, resolutions, lengths partial clips, watermarks and other changes --- with reasonable time and processing constraints, it would be a lot more valuable in other fields than as a form of DRM protection. At the very least I would be wondering if it would only be monitoring the audio track of the accompanying video to determine its matches.

It reminds me of the claims made by various "smart" porn blockers that "know" naked flesh from regular skin tones and photos -- generally it's nothing but baseless hype, or it's going to find a lot of false positives.

Re:Hmmm....

[Kjella]

Well, here's my idea. It's right off the top of my head, so it's probably patented:

To create the fingerprint:
1. Transform the frames to contrast edges. I forget what the filter is called but it'll basicly give you an outline of all the edges, with very little info
2. Split the frame into squares, calculate a brightness value from the edge frame
3. Throw all of this into a database

To check the fingerprint:
1. Decode a few keyframes
2. Calculate same info for those frames
3. Use a quick sort on brightness to find candidate frames. Don't require 100% match, make it so that a frame with a few watermarked squares will still score
4. Sanity check for file/location in other clip
5. Second pass either taking more frames, matching constrast lines or something, plenty opportunities here

Example, if they've blocked "mymovie.avi" with an NBC logo, and you're trying to match it against "othermovie.mpg" with a Fox logo:
1. Pick frames 154, 437, 1023, 2022 from Fox clip
2. Do the math
3. Find candidate matches
4. If they so *roughly* match frames 184, 467, 1053, 2052 in mymovie.avi you have a probable match with a 30 frame offset. In practise, just require an approximate match because of different commercial cut-offs
5. Verify match more closely

Sounds to me like a not too heavy job. It's very hard to fuck with because contrast lines are basicly the content of the picture. The same clip of a person walking will be approximately the same no matter how badly you fuck with it. Maybe you need to normalize it a bit if people try to upload high/low contrast vid to get around it, but that's going to look like shit anyway.

Re:Hmmm....

[slart42]

I dunno, but I think someone could come up with a reasonable scheme which could catch many cases of uploading the same content.

My cell provider offers this service where I call a number, leave it connected for 30 seconds, and it will reply with an SMS with the artist and name of the song which is currently playing. It works extremely well, even in crowded bars with lots of noises, and it requires only a 30 second clip. It's quite impressive.
But then again, nobody actively tries to fool this thing to gain some advantage.

Re:Hmmm....

[tepples]

My cell provider offers this service where I call a number, leave it connected for 30 seconds, and it will reply with an SMS with the artist and name of the song which is currently playing. Does it work with live performances of music or only with published, UPC'd recordings?

Re:Hmmm....

[slart42]

Does it work with live performances of music or only with published, UPC'd recordings? It is only supposed to work with published recordings. But I once tested it on a cover version of a johnny cash song, because i wanted to find out who was performing, and it gave me the original song instead. So it might actually detect some live perfomances.

Re:Hmmm....

[lawpoop]

A few questions: Couldn't this method be defeated by applying various photoshop effects to each frame? Is it a question of how much you change?

Also, wouldn't the video editors change the frame rate and do some other transformations in the video? What about various encoding schemes?

Basically, could you transform it enough to repost it and defeat the matching algorithms?

Re:Hmmm....

[Kjella]

Everything is a question of how much you change. But the constrast lines are the lines in the image, you can't change those easily without creating ugly artifacts. There's a few other tricks you can do like alter the aspect ratio/crop or adding bars to throw the "grid" off, but I think that could be compensated for. Changing the framerate would help in that simple example but I imagine you'd translate it to time offset instead of frame offset. Encoding schemes wouldn't matter unless they introduce enough artifacts in the result to matter. Ultimately it's also a matter of how crappy it would look to the enduser. If you can't get anything but a noisy, artifacted, framed image posted them it's probably not that interesting anymore, even if it's technically still copyrighted.

Re:Hmmm....

[Anonymous McCartneyf]

So the music recognition system thought a cover band was Johnny Cash.
I hope MySpace's system doesn't do the video equivalent. I hope it doesn't treat parodies of commercial copyrighted work the same way as actual commercial copyrighted work, and take all the parodies down with the less-than-legitimate videos...

Re:Hmmm....

[jZnat]

Audio fingerprinting is a lot more advanced right now than video fingerprinting; it's much easier to do it with audio.

Re:Hmmm....

[slart42]

Audio fingerprinting is a lot more advanced right now than video fingerprinting; it's much easier to do it with audio.

Well, only fingerprinting the audio track would still suffice to get rid of pirated south park episodes.

Re:Hmmm....

[Anonymous McCartneyf]

Yes, but it would also get rid of parodies using those audio tracks. That may not be a big issue with South Park, but I bet it would crimp the style of Star Trek and anime fans.

here's a question

[Loconut1389]

What if someone requests a takedown of someones content because the offending person uploaded a copy of their original content that is already (legally) on youtube- wouldn't this filter remove both copies?

Re:here's a question

[NPN_Transistor]

What if someone requests a takedown of someones content because the offending person uploaded a copy of their original content that is already (legally) on youtube- wouldn't this filter remove both copies?

I don't think so. Although MySpace and YouTube are both announcing that they are going to use similar copyright-protection systems, they are not part of the same system, so if a video is taken down on MySpace, it won't necessarily be taken down on YouTube.

Re:here's a question

[Loconut1389]

I substituted subjects unintentionally- forgive my mistake. I meant that if someone posted a copy of the video on the same service, what happens when the rightful owner requests a takedown.

Re:here's a question

[Anonymous Coward]

Somehow I doubt that this will be used for when the average joe complains about his video on someone else's page as opposed to when some media company complains that something they don't want on myspace at all appears there.

Re:here's a question

[Loconut1389]

I didn't say average joe- what about a big media company wanting to legally post a video on the system, and someone else posts the same thing so they ask for that to be taken down, wouldn't the filter catch themselves?

Implementation

[Anonymous Coward]

I Imagine in order to properly integrate this they'll be coding in some quit strong fuzzy logic in order to stop people from circumventing the system to easily

Oh noes!!!!

[Anonymous Coward]

Oh noes!!! whining bullshit about freedom of sp34ch will be along in 5, 4, 3...

Thats just about everyone

[twakar]

As I see it, anyone who has ever written something.. ie an essay, a love letter, a dmca takedown notice, or someone who has taken a video, or a photo is automatically a copyright owner. Given that, I guess we'll have to see how this goes because we should all have access to this "Take Down, Stay Down" crap... never mind the legitimacy of the complaint.
Will there be some kind of registration for commercial copyright owners? This is how it looks to me:

• We're all copyright owners of some sort, so we all should have access
• No mechanism in place to determine the validity of the complaint, or even who the complainant is.
• Fair Use out the window again
• Too much possibility for white noise
• Infringement isn't the same from country to country, what maybe illegal in country, may be legal in another

I guess the new ISP monitoring tools in place from an earlier article will be able to trackdown rogue posters.
I've really had enough of this crap. Commercial copyright owners will never learn that any exposure, that is non-commercial, or motivated by profit is good for your content. Serve it up and people will pay (;|;)

Re:Thats just about everyone

[tepples]

Will there be some kind of registration for commercial copyright owners? The United States Copyright Office provides a service for copyright owners to register works. It involves donating two copies to the Library of Congress, filling out a form, and enclosing 45 USD.

Re:Thats just about everyone

[ScrewMaster]

This is a particularly interesting quotation from Thomas Jefferson:

It has been pretended by some, (and in England especially,) that inventors have a natural and exclusive right to their inventions, and not merely for their own lives, but inheritable to their heirs. But while it is a moot question whether the origin of any kind of property is derived from nature at all, it would be singular to admit a natural and even an hereditary right to inventors. It is agreed by those who have seriously considered the subject, that no individual has, of natural right, a separate property in an acre of land, for instance. By an universal law, indeed, whatever, whether fixed or movable, belongs to all men equally and in common, is the property for the moment of him who occupies it; but when he relinquishes the occupation, the property goes with it. Stable ownership is the gift of social law, and is given late in the progress of society. It would be curious then, if an idea, the fugitive fermentation of an individual brain, could, of natural right, be claimed in exclusive and stable property. If nature has made any one thing less susceptible than all others of exclusive property, it is the action of the thinking power called an idea, which an individual may exclusively possess as long as he keeps it to himself; but the moment it is divulged, it forces itself into the possession of every one, and the receiver cannot dispossess himself of it. Its peculiar character, too, is that no one possesses the less, because every other possesses the whole of it. He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me. That ideas should freely spread from one to another over the globe, for the moral and mutual instruction of man, and improvement of his condition, seems to have been peculiarly and benevolently designed by nature, when she made them, like fire, expansible over all space, without lessening their density in any point, and like the air in which we breathe, move, and have our physical being, incapable of confinement or exclusive appropriation. Inventions then cannot, in nature, be a subject of property. Society may give an exclusive right to the profits arising from them, as an encouragement to men to pursue ideas which may produce utility, but this may or may not be done, according to the will and convenience of the society, without claim or complaint from any body. Accordingly, it is a fact, as far as I am informed, that England was, until we copied her, the only country on earth which ever, by a general law, gave a legal right to the exclusive use of an idea. In some other countries it is sometimes done, in a great case, and by a special and personal act, but, generally speaking, other nations have thought that these monopolies produce more embarrassment than advantage to society; and it may be observed that the nations which refuse monopolies of invention, are as fruitful as England in new and useful devices.

--- Thomas Jefferson

To encourage their invention, copyright and patent law were developed in most Western countries. These laws were devoted to the delicate task of getting mental creations into the world where they could be used - and could enter the minds of others - while assuring their inventors compensation for the value of their use. And, as previously stated, the systems of both law and practice which grew up around that task were based on physical expression.

Since it is now possible to convey ideas from one mind to another without ever making them physical, we are now claiming to own ideas themselves and not merely their expression. And since it is likewise now possible to create useful tools that never take physical form, we have taken to patenting abstractions, sequences of virtual events, and mathematical formulae - the most unreal estate imaginable.

In certain areas, this leaves rights of ownership in such an ambiguous condition that property again adheres

Re:Thats just about everyone

[Z0mb1eman]

We're all copyright owners of some sort, so we all should have access Don't be silly. Only those who can afford expensive lawyers and multi-million dollar lawsuits are copyright owners. The rest of us are either consumers or pirates.

Lemme get this straight...

[The Master Control P]

And to circumvent the filter, he added, a hacker would have to "screw up the content itself so it wasn't recognizable," to a degree where it wouldn't even be worth uploading in the first place.

AudibleMagic claims to have invented an algorithm that can recognize the same video in different forms better than the human brain can, across any format? That must mean the video is decoded into it's component frames before hashing, since that's how our brains get it. And better yet, this "Hash-Every-Frame" routine (which is apparently better than any other by leaps and bounds) will run on a site the size of MySpace without a BlueGene/L driving it? Uh-Huh. Say, NewsCorp, I've got this old bridge...

Hypothesis: AM's claim is bullshit.
Test: Everyone try uploading the same video, but add static and drop random frames from the start/end.
Outcomes: If hypothesis is true, AM and the Copyright Mafia look incredibly stupid. Again. If hypothesis is false, they handed us a free DDoS to push MySpace off the 'Net with by consuming all their processor time with hash checking.

Conclusion: Regardless of outcome, hackers win. Once again, DRM and everyone associated with it are Lolcows, unable to stop others from milking their stupidity for our amusement.

Re:Lemme get this straight...

[prencher]

Yes. Because obviously they'll be running the hashing on the same servers serving the website. Oh, wait...

Re:Lemme get this straight...

[jbengt]

Audible Magic software is one of the tools used by the RIAA "expert" in that recent RIAA suit where the RIAA expert was knocked down:
http://yro.slashdot.org/article.pl?sid=07/04/28/20 2206:/
http://www.ilrweb.com/viewILRPDF.asp?filename=umg_ lindor_070223JacobsonDepositionTranscript:/
Nothing like relying on trade secrets and black box algorithms to make you sure that you're taking down the right files and leaving up the clean ones.

Re:Lemme get this straight...

[PeterBrett]

AudibleMagic claims to have invented an algorithm that can recognize the same video in different forms better than the human brain can, across any format? That must mean the video is decoded into it's component frames before hashing, since that's how our brains get it. And better yet, this "Hash-Every-Frame" routine (which is apparently better than any other by leaps and bounds) will run on a site the size of MySpace without a BlueGene/L driving it? Uh-Huh. Say, NewsCorp, I've got this old bridge...

I reckon it wouldn't be too difficult to implement, actually. First, there's no need to examine every frame; one in every 50 would probably be enough. Then, the procedure is probably as follows:

1. Convolve the image with a Laplacian of a Gaussian to identify interest points ("blobs"), say at two different scales. Compression artefacts mean that there probably wouldn't be any point in trying to identify features smaller than say 1/8 of the height of the frame, and that will probably give you enough points anyway.
2. Around the features identified, use a low-pass filter to identify edges, and create a histogram of edge directions in the vicinity of the feature, scaled by normalised distance from the centre of the feature.
3. Normalize each histogram by orientation and magnitude. You now have a set of values (usually around 128) which identify that feature in a way which is robust to rotation, scaling, changes in brightness/contrast and changes in colour (this is called a Scale Invariant Feature Transform, or SIFT, descriptor). Take say 3 or 4 principle features per frame examined, and put them in a database.
4. Any video which has more than 5 frames in a row which match a series of 5 frames in the database is a positive match. This corresponds to ~10 s of video. It would be possible to use algorithms developed in bioinformatics to optimise your database for that kind of search.

The false positive rate from such a system would be negligible, and robustness and accuracy could easily be increased by recording more features from each frame. This is a mature technology, and it's been designed for much harder computer intelligence tasks than the one described.

digital fingerprint?

[deftcoder]

So basically, now they also make a CRC/md5 checksum of the video and store it into the database?

Genius.

Considering this should have taken all of 20 minutes to implement site-wide, I don't see why this is really news-worthy! It'd be easily bypassed anyways...

confused but IANAL.

[joe 155]

So they are trying to take down copyrighted content? Doesn't that mean that they can no longer claim protection from DMCA safe-harbor provisions?

So if someone puts some of my content up without my consent can I sue them? What if it was uploaded by an anonymous person using Tor as a browser registered using fake information for the purpose of suing?...

This seems like a bad idea.

Re:confused but IANAL.

[vertinox]

So they are trying to take down copyrighted content? Doesn't that mean that they can no longer claim protection from DMCA safe-harbor provisions?

IANAL either, but from my understanding you are still protected under safe harbor if you proactively remove copyrighted content or questionable content. As in... Youtube is still protected by safe harbor if they remove a video which was too obscene or violated its ToS without going through DMCA hoops. You don't have to have a common carrier status to be protected from safe harbor I believe.

Re:confused but IANAL.

[Kjella]

Not really. Clearly any site has the right to reject or remove content which they feel is in violation of their ToS. I doubt any "failure to detect" by an automated filter will ever change that. 512 C) 1) A) states that you mustn't have any knowledge and C) states that you need to take it down. It comes down to B):

"(B) does not receive a financial benefit directly attributable to the infringing activity, in a case in which the service provider has the right and ability to control such activity"

Advertising is financial benefit, directly attributable to page hits on infringing content. They have the right and ability to control it, because they could put reviewers in place to approve all content before posting. If you took that literally, basicly all "free" hosting would drop dead in an instant since it's all ad-supported. Regardless of whether an automated filter or human editors are in place, you still have the ability so it's not really relevant.

This filtering is their way of trying to fulfill this one, saying "We've done this to the best of our ability, you can't hold us liable for failing to meet this requirement". I doubt the courts would hold anyone to perfection and say "Because you've tried, you're liable for everything you missed". This is the paragraph that makes it dangerous for them not to - the courts could say "You have the ability, you could have filtered but chose not to because that's making you money." and disqualify them from protection.

Underlying technology...

[Anonymous Coward]

I don't know if Audible Magic's come up with anything special for video, but for audio, the underlying technology is at least partly based on MFCC. You can find some more details here and here .

Also, they have to apply this filter only when content is added to a page, not every time it is played, so its a little less computationally intensive than some people have suggested.

It may be more expensive to search all existing fingerprints every time a new take-down request comes in.

A lot of us are still waiting?

[Mathinker]

> As altavista faded so will ....

You are certainly correct, but ....

I'd guess that almost everyone on Slashdot is waiting for their personal vendetta objects to fade, be they Microsoft, Apple, the open source movement, Google, MySpace, the music industry cartel, the movie industry cartel, and a list of thousands more, including for some people (probably not mainstream Slashdotters) large organized nations and religions....

Much as I would like to fantasize to the contrary, even if you believe in future shock / technological singularity / etc. I don't see how it directly applies to the fading of organizations like MySpace or others listed above, considering that their existence and power have more to do with organizing or influencing people to act in a cooperative fashion than with the power of newer and newer technology.

It could (and probably will) take a lot longer than you think....

Cranky because your job got outsourced?

[FatSean]

Guess you shoulda tried harder in school, eh?

Copyright owners have access

[MobyDisk]

Copyright owners have access to Take Down Stay Down free of charge, according to a release from MySpace So, as a copyright owner, where do I sign-up?

Strike me down..

[nurb432]

And i will only become even more powerful.

Its time to get their heads out of the sand, the entire concept of 'content rights' is out the window.

You know...

[Cheezymadman]

Instead of asking "Will this work?" or "How does this work?" or even "How can we get around it?", shouldn't we all be asking ourselves the really important question?

Who gives a shit what MySpace does anymore?

Learn from the past

[mattpointblank]

All I can say is: why won't people learn from history?

Almost 10 years ago now there was a little app some of you may remember called Napster. It offered mp3 downloads that, at the time, could take half an hour or more to complete. But it was worth it, because you couldn't get the music anywhere else (for free, anyway). Napster got closed down, but everyone just moved their collections over to Kazaa, Limewire, BearShare, etc etc. A few years later, the music industry catches up and realises that users are resilient and know what they want. This the iTunes Music Store (and its rivals) were born.

Now we're in a faster internet age, the same is happening with video. People want on-demand content. If someone tells me about a funny Colbert clip, I'm not going to check the TV guide for a repeat showing. I'll stick it into YouTube and watch it there. YouTube delete the Colbert clips? I'll watch it on DailyMation. Repeat ad infinitum.

Myspace can block out videos but people will find a way, and continue to find a way until the networks realise that in 2007, for the first time, the audience is starting to control the media.

Re:Learn from the past

[krbvroc1]

If someone tells me about a funny Colbert clip, I'm not going to check the TV guide for a repeat showing Not to mention, that there is generally no 'repeats' for shows like Colbert Report or The Daily Show. If you missed it and didn't get it on a DVR, you're out of luck.

Re:Learn from the past

[Anonymous Coward]

If someone tells me about a funny Colbert clip, I'm not going to check the TV guide for a repeat showing. I'll stick it into YouTube and watch it there. YouTube delete the Colbert clips? I'll watch it on DailyMation. Repeat ad infinitum.

Myspace can block out videos but people will find a way, and continue to find a way until the networks realise that in 2007, for the first time, the audience is starting to control the media.

Yeah, it would be so cool if Comedy Central realized that and placed the clips on the web themselves.

Oh, wait...

Re:Learn from the past

[tedd169]

You forgot to mention tape recorders and VCR's.. Before napster, these inventions were to end all paid programming just like ufo's took all the dinosaurs away... I saw that on quality programming brought to you by... There is always a way to do something. And since television can't mesmorize you/me with half a$$ off budget shows, we will make them spend their hard earned money to battle the consumers paying for it. Is this like the food chain. hmmmmm... I'm not a business major, but how do you make money doing that??? Oh... I see. Another reason why we are behind.

Re:Learn from the past

[Anonymous McCartneyf]

You know, iTunes sells The Colbert Report. You can get any given episode there if you have iTunes. Or are you insisting on free?

Re:Learn from the past

[sakasune]

there is generally no 'repeats' for shows like Colbert Report or The Daily Show

Yeah, except for the 3 times the next day (or sometimes the next week) that they replay it. They also throw up repeats when they take a week vacation, but that's not good if you're looking for something specific. I'm just wrangling...

Double edged sword

[Coopjust]

How is a computer supposed to tell, with exact certainty, whether a video matches content pulled for content violation? There are so many problems with this idea: -If you make it so it only picks up exact copies, people will resize or make some other small video adjustment. -If you make it aggressive enough so it can pick up on those modifications, it will probably hit a bunch of false positives. -If it is possible to make it smart enough to pick up on videos and not get false positives, it will hammer the servers. -What about the ease of abusing the DCMA? What happens when an owner of a copyrighted video is blocked by a fake request? What is the unblock mechanism? -If Myspace manages to do all of the above without issues, then you've driven all of the worthwhile content off of the site and alienated the userbase...

Re:Double edged sword

[tepples]

What happens when an owner of a copyrighted video is blocked by a fake request? What is the unblock mechanism? Falsely accused uploader files a counter-notice. Video goes back up after two weeks. Copyright owners making repeated false accusations are sued for perjury.

Audible Magic

[jbengt]

Audible Magic software is one of the tools used by the RIAA "expert" in that recent RIAA suit where the RIAA expert was knocked down:
http://yro.slashdot.org/article.pl?sid=07/04/28/20 2206
http://www.ilrweb.com/viewILRPDF.asp?filename=umg_ lindor_070223JacobsonDepositionTranscript
Nothing like relying on trade secrets and black box algorithms to make you sure that you're taking down the right files and leaving up the clean ones.

Easy to Thward

[Anonymous Coward]

And in less than one hour, I wrote a program to alter the content just enough.. to thwart their stupid little program.

Myspace's priorities

[British]

So they roll out this technology for the relatively unimportant issue of video piracy, yet they let the spambots roam free? Myspace's biggest problem, IMO is the gazillions of fake accounts being made every minute. It ruins the site. No, they are more interested in video piracy, jeesh.

Re:Myspace's priorities

[lordSaurontheGreat]

Yes, they are more interested in video piracy than spambots. If they let video piracy continue, something even worse than spambots comes: LAWYERBOTS!!! (queue the ImperialMarch.avi file, please...)

It will become unusable

[rabel]

Oh please, this is moronic. I'm with the posters here that are calling bullshit on the claims that they can fingerprint all content even when it's been modified.

But if it's true then there will be so much content marked for takedown that the site itself will be completely unusable. If nothing else people will mark every video as in violation just out of spite.

Not making it impossible, just impractical

[Alien54]

For the vast majority of folks who are not fluent with video editors, this might be enough. It might be good enough to make sure that most folks can't upload. Sure, altering the speed by 1 to 5% might screw up the digital signature. But then there was that case of a classical piano artist whose husband had been caught passing off other performances are her own.

It might come down to another technology arms race. Or it might come down to an uneasy truce, if the lawyers can keep their paws off of things.... (not likely)

Oy.

[GeorgiaCodeMonkey]

Oh, wonderful. How long before Uri Geller starts going apeshit on Myspace?

I think these media companies are crazy

[tkrotchko]

Actually, what I'm wondering is what this really accomplishes in the long run.

If everybody is showing a 5 minute clip from some TV show that Big Media Conglomerate (BMC) owns the copyright on, and assuming this is foolproof, does BMC think "Oh gee, when those dirty filthy kids can't 'steal' my content, they'll be sure to pay me to use it?".

This is a classic case of:

1) Stop everybody from listening/watching my content
2) ?????
3) Profit!

I was flipping through the XBox 360 the other day, and I realized they're trying to get people to pay $2 for Colbert and Southpark, things that you can watch for free. And I wondered... is that really a growth market? I get the feeling the media conglomerates want us to use a business model that the average consumer has no interest in: pay per view for everything.

What's really ironic is that when they show it for free on TV, they do their best to make sure everybody watches it and there is no restrictions on recording and sharing it with your friends. But when they want to charge for exactly the same content, they do their best to try to make it hard to watch and use the content.

I'm trying to wrap my head around their strategy, and I can't make head nor tail of it.

Re:I think these media companies are crazy

[Anonymous McCartneyf]

There are two benefits to buying The Colbert Report off iTunes or XBOX:
1. No advertising!
2. Permanency. Less of an advantage.
It's not as easy to record programs permanently as it used to be. DVD recorders for TVs are scarce, VCRs and their tapes are rare, and standard DVRs tend to delete things after a certain amount of time--or, for the one I used to use, an uncertain amount of time. This, of course, is the MPAA's fault.
South Park has reached syndication. It's on over-the-air stations now--truly free, at least as in beer. (Not sure how edited.)
The Colbert Report is still on a pay-channel, though; we just ignore the cable or satellite costs like we ignore our ISP costs when we figure the cost of the content, since under normal circumstances we'll be paying those anyway.

Re:I think these media companies are crazy

[jZnat]

Unless you "steal cable", your TV subscription is most certainly anything but free. Of course, if you've got cable internet, you probably had to get basic cable since it would end up costing the same regardless on whether or not you got the basic cable subscription in the first place. In that case, it's basically free.

Also, the TV shows you download on Xbox Live are ad-free and whatnot. It's pretty much the iTunes Store, but by Microsoft instead of Apple.

I think this is how Audible Magic works

[Anonymous Coward]

1) Hash frames like everyone else said
2) Sample-- don't just check every frame
  or
2) Use Key frames selected by human
3) Once you have identified a target video simply reproduce this video in different codecs and repeat the process. This will not allow fidelity to be increased as expected.