Content Recognition Software new weapon in war over Piracy

[Polaris573]

Last week, Audible Magic in Lost Gatos, Calif,. demonstrated their new content recognition software by downloading a video clip from Youtube. The system was able to identify the clip as a scene 49 minutes and 37 seconds into "Kill Bill: Vol. 2." even though the clip had been recorded on a camcorder and dubbed in Chinese.

MySpace said last week that it would use Audible Magic's system to identify copyrighted material on its pages. However, YouTube is still one major hold-out resisting the new system.

The system works by comparing a file to a database when it is uploaded to a website. Then it either allows the file to be posted or blocked depending on if the file is licensed for use on that website.

View at TechPowerUp Main Site

 

[Scavar]

That's pretty sweet. I can see how it might get hijacked though....YouTube is kind of going down the tubes anyways....

Is Google Video going to use it?

 

[Polaris573]

That's pretty sweet. I can see how it might get hijacked though....YouTube is kind of going down the tubes anyways....

Is Google Video going to use it?

Google Video is a holdout too since YouTube is owned by Google. However, I imagine they will be forced to use it eventually.

I just summarized some of the important points of the article, the whole thing is two pages long but it is a fairly interesting read for those who want to check it out.

 

[Scavar]

Yeah I am kind of skimming through it now. It seems like an interesting piece of software. They will probably end up using it at some point.

 

[Completely Bonkers]

Although, from a software standpoint, this is a breakthrough... it is also an intractable problem. If you have ONE clip to test against ONE source (altered) then it is doable. But to test thousands of clips against thousands of possible sources is just not doable.

Can you imagine:

1./ The computational power needed? (Including kWh of burned un-green carbon fuels)

2./ The amount of HDD space to hold "legit" versions as the testing bank

3./ The TIME is would take to test a new clip against ALL copyrighted material?


ANSWERS

1./ If would require a HUGE array of computers at huge cost and burn up politically unacceptable resources

2./ The space requirement, cost, and license issues of holding full and complete bank of all copyrighted video material

3./ An upload is done one day... and 6 months later it has been tested against every possible combination of copyright material to finally be "authorised" for upload

AWEFUL state of affairs

 

[HookeyStreet]

LOL sounds like bullsh*t to me! How can 'software' detect that kind of information? It knew that the video was a scene from Kill Bill that was recorded on a camcorder??...lmao!! So in no kind of way would that file be digitally signed like how an original DVD would be ( a bit like extracting an audio CD using windows media player and it automatically gives you the song names) , so what data would the said program use as a reference to find out what 'movie/file' it is from the develpoers database??

I smell a big bit of BS that just been made up to 'scare' the pirates!

 

[zekrahminator]

Great idea, but very easy to bypass if it only compares how long a clip is...you can always slow-mo something/ cut it short.

 

[Scavar]

They really shouldn't call it piracy. Pirates are badass. If they had called it ninjaing, it would be even more rampant.

 

[Thermopylae_480]

LOL sounds like bullsh*t to me!

Uhh. NY Times doesn't typically concoct BS stories like some of the online news sources some people frequent. If you get discovered concocting a story you loose a rather lucrative and respected job. So, imo it is probably not made up, but true. Of course I'm sure the conspiracy theorists will scream otherwise.

 

[PVTCaboose1337]

Move google and youtube to a huge server farm on sealand

 

[zekrahminator]

Or better yet, that island the pirate bay is buying .

 

[Thermopylae_480]

Or better yet, that island the pirate bay is buying .

Lol, that's Sealand .

 

[zekrahminator]

...Oh. I knew that . Sealand just sounded more like Shamoo's house or a jet-ski company, but then again I don't spend enough time outdoors.

 

[Thermopylae_480]

Lol, Seadoo, and Seaworld .

 

[Batou1986]

lol they can have fun building that database wonder if it works for porn i could use some software for that

 

[AshenSugar]

thermo i think he ment the idea that this could be put into wide use was bs or that the fact that it was a "random choice" was bs, and the info the reporter was given could be bs, or "tweaked facts" as some people call them, u know, they set things up a certen way and report them a certen way, like ms's spining vista as better then xp


i got an idea, a good one, move server farms to old unused oil rigs, refit them and use them as fileserver platforms, imean honestly they could do the job, would be easy to cool(u got VERY cold water under you after all.....hehe) and honestly probbly wouldnt be any priceyer then building some new warehouses to house your server farms, biggist caust would be for communication gear/cabling, and im sure somebody would give them a good deal on that, or google could just buy the company that does that stuff and run their own lines

 

[WarEagleAU]

Kind of cool, but I doubt the industry is really loosing this much money based on a crappy you tube video.