Slashdot Mirror
MySpace to Use Audio Fingerprinting
dptalia writes "MacWorld reports that MySpace is going to start implementing audio fingerprinting to prevent copyrighted material from appearing on their site. The new technology will be used to review all uploads and prevent 'inappropriate' material from ever seeing the light of day."
Add a little noise here, stretch and squeeze some sounds there, change some frequencies over in that corner, and pretty soon you'll either have false-negatives or the potential of false-positives.
If it defeats the filters expect such tools to become widely used within a few months.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
when their songs are (wrongly either because it's their original song and it's copyrighted by them or because of a technical glitch) forbidden from being uploaded.
Please, for the good of Humanity, vote Obama.
Now I can definitely pat myself on the back for not having a myspace account. Another freedom of ours eroded away - the ability to infringe copyright, ignoring the consequences. It's kind of like the seat belt thing. I don't have to wear it (if I'm stupid enough) but no one else gets hurts in the process.
I takes forever for a video feed to start with a 6Mb/s downstream. WHat going to happen when they start analyzing that data on there end also?
Here's to hoping MySpace bloats their site out of existence.
It's true I tell you, feller at work's next door neighbour read it in the paper.
Ah, yes, because every MP3 file playing a certain specific song is identical in length/checksum ... [/sarcasm]
By reading this signature you agree to not disagree with the post you just read.
...asuming it works as claimed. Myspace is chock-full of copyright violations of still pictures and videos and the typical Myspace user seems to not have a clue.
Or has GOOGLE crossed the evil line?
Hmmm... Google? Did I miss something?
So say we all
I am absolutely certain that this audio-fingerprinting software is aware of the concept of fair use and has embedded logic to handle cases where fair use is employed.
Ok. I'm having troubles writing that without losing my face.
Not Buzzword 2.0 compliant. Please speak english.
The 'article' is woefully low on information, apart from a mention of Gracenote MusicID being used. From Gracenote's own page (Its on mobile music recognition, but I assume the principle is the same):
How it Works
1. When music fans hear a song they want to identify, they tap a command on the phone keypad to start the audio recognition process, and then hold the phone up to the music source.
2. The phone captures a few seconds of the audio and extracts a waveform fingerprint of the snippet. The snippet can be from any section of the song, even the last few seconds.
3. The fingerprint is sent to the Mobile MusicID recognition service from the service provider that may be located anywhere in the world.
4. The Mobile MusicID recognition server compares the fingerprint to its database of reference fingerprints and responds with the exact match.
5. The artist, song title and related information, as well as content like album art and download links are relayed to the fan.
I can explain it for you, but I can't understand it for you.
http://www.gracenote.com/music/corporate/press/art icle.html/date=2006103000
Wincopy
I wonder how well this will actually work. Audio Fingerprinting is designed to be insensitive to most 'naturally occuring' music distortions such as encoding artifacts, noise and changes in equalization, but I don't know of any audio fingerprinting system that will work well when faced with people who are actively trying to evade detection. It won't be too difficult for a properly motivated MySpace user to find a set of filters that can be applied to any song that will allow the song to get a unique fingerprint, without actually changing how the song sounds. A quick trip through Audacity to apply a micro-pitch change, a little equalization, and perhaps a slight tempo change will probably do the trick. Of course, the folks over at Gracenote are pretty smart and may be able to adapt to evasions, but this will no doubt lead to even more sophisticated evasions. In the end I don't think it is possible to create a fingerprinting system that will be able to deal with people who are actively evading the system. In the end, the evaders will win.
But... we're talking about MySpace, not YouTube?
I think this is where the confusion comes in...
Sweet! I'd better make some noise music quick, I'll have the rights on white noise on MySpace in no time!
THIS IS THE INTERNET. PLEASE PICK UP YOUR SERIOUS BUSINESS SUIT AT THE FRONT COUNTER.
It's good. Good idea. Should cut out much of the copyrighted content posted on MySpace.
/. What about false positives? How can they be accurate? Why do the RIAA think the world revolves around them (at the inconvenience of others)?
Oh wait, this is
You know, there is a difference between trolling and pointing out the flaws in your reasoning. Just saying.
Funny, I never realized that about those two songs.
But beyond that example, all music is derivative. There are a number of chord progressions that lots of music uses. Standard blues riffs are the basis of countless songs. I highly doubt that at this time there's a way for software to accurately recognize songs. Similar songs? Sure, but there's no law against creating music that sounds similar to other music. The false positives will make it difficult to upload damn near anything.
Others have already commented on the Google thing, I'll just chalk it up to being early in the morning.
Am I the only one getting crazy here or the only one NOT getting crazy? As a good slashdotter I didn't read the article (well, now I read and it's pretty short, by the way), but I found reference to Youtube or Google nowhere! Are we using the same internet?
So say we all
... a death sentence to the Weird Al's MySpace page to me.
You do realize that this is MySpace trying to police what is uploaded to their (free) service? I imagine there will also be mechanisms in place to have things that get through removed. So what would courts have to do with this?
That thinks this is a GREAT development? No more music playing when I visit someone else's page...finally! Why people think they must play a song when I visit their page is beyond me. It was popular back in '95 or so, when the web first became popular, but then common sense broke out and everyone stopped doing it...until MySpace came along. (And don't get me started on the awful designs people use - backgrounds that make the text impossible to read and slow to scroll, etc...)
Will there be a court to deside when it is wrong?
Why would there be? We're not talking about prosecuting people (yet...), just about filtering copyright materials that legally people shouldn't be uploading anyway.
Why would a court be involved?
It's official. Most of you are morons.
posting an off-site link to illegal content
ripping and saving under a different file type
stop using MySpace and moving on to the next big hype It's their money let em waste it how they want. They should know by now its only a matter of time before whatever solution they use will be defeated.
Infiltrated dot Net
Is there any chance at all this will work? Myspace can't even handle basic uptime issues, let alone complex audio fingerprinting technology. I'm not even 100% sure they have a test environment yet (for many years they didn't). Half the time you go to the site at least one part of it is completely broken. Will we start getting messages from Tom, "sorry guys, every song thinks that it's hey ya. We'll fix it. In the meantime don't email us. We know.".
If an officer ever threatens to taze you, say you have a pacemaker.
It says NOTHING about how this will be implemented. For all we know, they are not cutting the human out of this process. It's very, very possible that they'll be using fingerprinting to flag potential copyright violations, and have a human review it before deciding to reject an upload.
Besides this, audio fingerprinting is not a binary process: "is a known track / is not a known track". Many implementations return a "confidence" value, or even the top 3 "best guesses". Thus, it could automate the process for, say, hits where the system is 98% sure that it's a copyright violation or above, but notify a human for hits that are between 80% and 98%.
I really don't see a problem with utilizing statistical techniques to determine whether or not a song is likely copied off of a CD. There's nothing unethical about identifying music. I'm really surprised at the number of negative posts here...
One of the classic signs of getting old is when you can't tell the difference among the things kids are into these days (MySpace, YouTube, you know, one of those social thingies; Iron Maiden, Megadeath, whatever, one of those noisy bands).
I'm not knocking you, I don't care to make such distinctions myself, but it's still funny to witness.
Aphex Twin already holds copyright for all kinds of noises.
I'll never be able to sleep soundly again!
Gentoo Linux - another day, another USE flag.
...and not on, say, responding to emails from my friend. He's being stalked on the Internet by a ruthless AOL-using lunatic who has conjured death threats against his ex-girlfriend while faking his identity, using publicly available information from Yahoo & Facebook & a personal blog. Whoever this is has convinced the police in three different states that my friend is writing these death threats (and thereby gotten him questioned and an investigation ongoing) and has convinced my friend's ex-girlfriend to file a restraining order.
It's a terrible mess, and Myspace is dragging their feet. Average age of each new profile created by the stalker: weeks to months, even though they're contacted promptly by email each time. Amount of uniquely identifying information provided about the stalker to police by Myspace: zero.
This is the kind of situation you would expect a very large, very public Web site to have some defined policy for. As far as I can tell my friend has not been able to find this policy. (Or to find a working phone number for these folks.)
But instead of worrying about little details like freakish Internet stalkers who pose a significant danger of actually hurting someone, they're working on bottom-line things like "not getting sued by recording industry".
Color me surprised?
(Poor "my friend", though. At least all I have to worry about is problem sets.)
We recently had heard in the office over one of the Yellow Machine that's made by Anthology Solutions.
I think his point was that since it took a court to decide whether the two songs were legally different, will such a recourse be available in the even of something similiar happening - the MySpace audio fingerprint detecting a new composition as copyrighted, for example. I suspect there won't be.
http://myspace.com/StepUpUK Mind you if that horrible attempt is /dev/nulled I've no compliants.
is http://musicbrainz.org/. It's an open source music fingerprinting project that can, for example, take a hard disk full of untagged MP3s, and tag them all up fairly accurately. It's free, and it works. I *think* it uses the same engine as the UK Shazam! mobile phone service where you could ring a premium rate phone number in a club, hold your mobile up near the speaker for 15 seconds, and it'd text you back the track/artist details seconds later.
Presumably it'd be trivial for Myspace to run this in the background on the boxes where they keep user audio content...
Maybe it could block a few of the jackass wanabee videos by mistaking pain screams for Celine Dion.
Not that I care anyway. When they are trying to remove themself from the gene pool, at least, they're not playing M-rated videogames.
Found it strange that macworld reported this story?
Apart from a few dusty 78rpm shellacs, pretty much all music recordings are copyrighted, so MySpace is going to find itself pretty barren.
Sorry, Murdoch's NewsCorp never had a "don't be evil" policy!
Under the terms of the DMCA, ISPs cannot be held liable for copyright infringments by their users if they follow certain procedures, particularly the "take-down" procedures whereby a copyright holder can request the removal of allegedly infringing material.
This provision applied a "common-carrier-like" regime to ISPs, treating them as conduits rather than as publishers or editors. Once you start reviewing uploads for potential infringment, doesn't this undermine the conduit model and open the ISP to potential claims of contributory infringement? Once ISPs begin reviewing content, they become editors and not simply conduits. I would think this is a dangerous road for an ISP to tred.
same internet, different tubes.
At some point, somewhere, the entire internet will be found to be illegal.
We can only hope that myspace puts as much effort into this feature as they did into all the other great and well-designed features on their site. BWAAAAHAHAHAHAHA
..hehehaCut it. That rebellious act stops right here and right now. Either you're with slashdot or you're with myspace. I'm seriously suspicious if you live in a basement even.
This myspace story is here only to make fun of myspacers and you blatantly abused the goodwill of the poster by not ridiculing them. Immediate remedy is required here, do something fast, even poor joke about myspace and you can keep your license to nerd.
Preserve old classics: copy your collection onto all hard drives.
The funny thing about the Google/MySpace confusion is that I noticed recently that some of Google's services will now integrate with MySpace (and will flat out mention MySpace, as in post to blog/MySpace buttons). And then there's that Google being the search provider for MySpace.
I really wouldn't be surprised if one day Google takes over the day to day running of MySpace as a service provider for Fox/News International.
At some point, somewhere, the entire internet will be found to be illegal.
Just asking...
We have officially invested more brainpower, money and technology into correcting and restricting technology compared to simply using it to its own best expression.
Why can't someone somewhere do something like this for identity theft? Oh yeah I forgot, because WE'RE ALL FUCKING SHEEP BEING LEAD AROUND BY CORPORATE ATTORNEYS !!!
My bad.
To expand on that a bit, Murdoch has a "DO be evil" policy.
Show me where it's codified into law, that I am able to take any copyrighted piece and place it onto an open network for all to take free of charge?
I've wondered about this, ever since buying a Neuros mp3/ogg player, which theoretically could identify for you which song from the radio you were recording.
It seems like the most naive approaches would be far too brittle: a bit-by-bit comparison or an MD5 sum, for example, would be thrown off by just eliminating or adding one audio sample in the song.
Even something like spectral analysis would be subject to errors: unless the reference copy they kept of a spectral analysis was produced using exactly the same start- and end- sample in the recording, as was used when checking a newly submitted recording, even the spectra might mismatch.
So how do they do it without getting a ton of false positives and/or false negatives? (Or do they not even manage to avoid those errors?)
I agree. The programmers working on integrating music fingerprinting technology into their website really should be more involved in making sure your friend doesn't get harrassed online.
Just like all of the theoretical physicists out there should be working on a cure for cancer. Dumbass.
Don't blame me; I'm never given mod points.
Until they fix their media player this doesn't really matter - you can't play any track that's been added in the last six months or so where the uploader hasn't marked it as available for download. So about 80% of all tunes on various MySpace artists' pages are now completely unlistenable anyway.
The recourse is for the alleged infringer to just settle, as all possible songs are likely to be substantially similar to something that's already copyrighted.
It's plausible deniability for Myspace/Newscorp - so they can tell all the other media companies they're doing their best to remove infringing material.
It doesn't have to actually work. In fact it's better for MyCorp if it doesn't work, because then they get to keep using all the copyrigthed stuff that attracts so many users to their site.
I guess to clarify without sounding like I know anything about the company's organization:
The problem I identified is managerial, not technical. The people involved in hiring programmers to do fancy RIAA CYA business should also hire more customer service representatives to handle "CYA from federal law" business. It just makes sense. Even PayPal had to do this eventually.
I apologize for sounding like I was blaming somebody (a computational audio expert) for doing their job. I meant to suggest that somebody else entirely (a CEO, or perhaps a controlling company) was not doing their job.
We recently had heard in the office over one of the Yellow Machine that's made by Anthology Solutions.
You're missing the forest for the trees:
Trees = court judgement.
Forest = human intervention was required to make an accurate judgement call.
I am not a crackpot.
I can name that DRM in 3 notes, Jim.
Wise men say, "Forgiveness is divine, but never pay full price for late pizza."
God I hope they aren't using Musicbrainz.... they'll rape their bandwidth, and then they'll not give any money back to that project. The new audio fingerprinting is so accurate there it's not funny.
Finally! A slashdotter who understands U.S. copyright laws.
You never really know how close to the edge you can go until you fall off.
that's incorrect. i added 2 songs to my spaceverb profile about 3 weeks ago, neither is marked for dl, both play just fine. feel free to check it out for yourself http://www.myspace.com/spaceverb
-.no
what of the folks who sample songs for their own use... wouldn't this always come up as copyrighted?
"i stand on the edge of destruction" -shai hulud
MacWorld reports that MySpace is going to start implementing audio fingerprinting to prevent copyrighted material from appearing on their site.
Weird Al had recently posted for free download a couple of his songs from his new album. Ironic that one of those free downloads was his song "Don't Download This Song". But those are HIS songs, and I assume that by posting them he had made sure with his producers and whatnot that they agreed to it as well. Will MySpace now prevent Weird Al from posting his own songs of his and his producers' own free will, because they also happened to be available for purchase on CD at your local music store? Will there be a way for people to legitimately post their own copyrighted works, or are legit artists now going to have to find a different site to post their own works that they wish to post?
Depending on specifics of the algorithm, it may be very hard to defeat it if you still want the music to be recognizable by the listeners. I am familiar with the audio fingerprinting algorithm from another company. The false positives are not a problem. The hash space is huge thus collisions are very rare. The false negatives can be a problem, but if they can weed out even 95% of attempts to upload copyrighted music, their life is going to be much, much easier. And if you distort the music enough to defeat the fingerprinting, then maybe you just have created a new masterpiece (c) you :-).
Does MySpace actaully permit users to upload MP3s to their servers? I've seen dozens of MySpace users upload their MP3s to our webserver and then MySpace leeches the bandwidth from us instead of MySpace servers. Am I seeing this wrong or are the users loading their MP3s to our servers to avoid MySpace TOS problems? I am not a MySpace user.
signature pending slashdot approval
There is a serious limiting factor that has to be accounted for - listeners. When you uploading music you want your listeners to enjoy it, right? If you add too much noise or other distortions, the music will sound like an old scratched vinyl. I worked with this audio fingerprinting system. The hash space is huge. The false positives are very rare. The rate of false negatives can be controlled by tweaking parameters. Then MySpace doesn't need 100% recognition. Even if they prevent just 95% of copyrighted uploads, it would create too much hassle to upload copyrighted music on purpose. Also, myspace would get a lot better legal standing by demonstrating that it's actively fighting the copyrighted uploads.
...how good this fingerprinting is - will it detect the Beastie Boys / Dan the Automator mashup I've posted on my own Myspace page, which is two copyrighted tracks combined?
Game dev and music blog
Current:
"Gracenote provides businesses with critical embedded software and metadata that ENABLE consumers to better MANAGE, ENJOY, and DISCOVER digital media."
Updated:
"Gracenote provides businesses with critical embedded software and metadata that PREVENT consumers from MANAGING, ENJOYING, and DISCOVERING digital media."
I bet it uses one of the Self-Organizing Maps variants to identify the sound snip. You may be able to fool this algorithm by injecting sound frequences in your upload that are too high to hear for people, but is still represented digitally. Of course this depends on *how* you inject the sounds as not all microphones would pick it up, but if you just edited it in a studio or patch program, no problem.
MySpace has created whole industries around it and this will be another. People will setup sites using Adsense as a revenue source. The sites will input a sound clip and output the same clip in a way that will bypass the audio fingerprinting.
Ninjas don't carry tic tacs
Which is clearly not true; anything subject to copyright is copyrighted simply by the act of creation, so unless "audio fingerprinting" can somehow identify that a work is a original creative work legally subject to copyright, it won't "prevent copyrighted material from appearing" anywhere. Even the slightly more detailed Gracenote press release (or perhaps the MySpace policy is referred to) seems confused about this.
It will prevent material which (according the matching algorithm used), "matches" material that is found in the "Gracenote Global Media Database". It supposedly will block "unauthorized copyrighted material", though the article isn't clear about any method to verify that the use is "unauthorized".
Are you actually arguing that a service provider needs the permission of the person uploading a file to copy that file onto their disk? Don't you think that maybe the act of actually sending them the file implies that you're granting them a right (assuming you have the right to do so) to receive it? Are you really that dumb?
Don't blame me; I'm never given mod points.
I ask for something concrete and I get "But you can cut pictures out of a magazine for your scrapbook!!!1"
When that BMW I cut out of the Car and Driver can actually drive me to work, then I'll be impressed.
Stop using myspace.
Myspace can't stop the barrage of stalkers and child predators, but they will be able to find and remove copyrighted files? Sounds like Murdoch's priorities need a bit of realignment.
Let's say I write and master a song, and then share it with the world at large under the CC license.
Now let's say someone has that song in their library. So, they fire up their jukebox software, which happens to use GraceNote to ID songs. They run a scan on their entire library, and 20 seconds of the song gets sent off to GraceNote to compare it to their database.
Gracenote is a for-profit company using a proprietary system to ID songs. They charge their customers (not usually the end user, but the companies that integrate their system) for access to that database, and don't provide unobfuscated access to that information. If GraceNote actually stores the 20 seconds of that CC song that has been sent to them, this could be construed to be copyright infringement, as the use of that song goes against the CC license agreement.
However, copyright is not carte blanche one way or the other. It could be argued that 20 seconds of a song used for reference purposes does not infringe copyright; nobody gets to hear those 20 seconds, and it is not enough of a sample of the song to argue that a person has copied a sizeable portion of the song for infringing use (unless the song is under a minute long). So, even though the CC license becomes void due to how GraceNote is using the data, copyright law itself allows for such quoting of original material. This means that in most cases, what GraceNote does is perfectly legal.
However, keeping a complete database of song lyrics and album art requires permission of the composer or artist who created the work in the first place. If they put THIS information in their database for a CC work, they would indeed be breaking copyright law as well as the CC license.
I hope they've thought this through.
Gracenote almost certainly doesn't serve album art that's licensed under CC_noncommercial or any album art that they're not licensed to use. They certainly claim that they have licenses from the publishers for all of the lyrics they serve, and unless someone can show that they've got CC lyrics they're not entitled to serve I see no reason not to assume they're working in good faith.
As for music identification, it's not entirely clear to me that actual music snippets are sent to Gracenote to be identified; it seems to me (and I haven't seen a complete description of the exact details to be sure) that a fingerprint is generated for the song and then compared to their database. If the fingerprint computation is done of the client machine, there's no doubt whatsoever that this isn't a copyright violation, since nothing's being copied.
I think I'm not missing the point at all, though, as the AC I was replying to seemed to be implying that it would be illegal for MySpace to take the music he uploaded to them and then to do these calculations on it because he hadn;t licensed the music to them, which is patently absurd.
Don't blame me; I'm never given mod points.
Send them so much chaff to Fingerpint that no one can upload any audio whatsoever. At that point the site, hell maybe all such sites, will be worthless. Let's make the whole fucking Internet deaf and blind. Maybe then the greedy bastards will get the point.
Apparently, the *AA think information wants to be useless. So, let's give them more useless information than they can handle.
Maybe nothing is being copied, but is the "fingerprint" a deriviative work? Of course to take that fingerprint, one has to copy it from the disk into RAM to process it..... Kind of in the same way a cd player copies the data off the cd to convert it to analogue....
You call that a inconvenience or a relief you can't visit a website without (A) protecting your ears or (B) turning the speakers down? I find it relaxing to know I can visit a website without smashing my brains out...
Although; it is because the RIAA wants to set a standard how they think music (artistic works) should be treated; they don't think about the middle-man or the end-user; they just care about what gets back to them and supposedly the creator; who is getting paid a cent from the cuts they take...
--- I am known for the ones who want to find me on the net. Is that a privacy risk or a privilege? One might wonder..
That's one HUGE music library!
You are right, but myspace will be using an external database to determine if the music is copyrighted. The audio print tells them something like "this record is equivalent to db entry #123458783 with probability of 97.4%". If that entry happens to be marked as public domain, the upload can proceed.