Slashdot Mirror
Audio Fingerprinting Via Cell Phones
aruil writes: "MSNBC has a story reporting on yet another audio fingerprinting application. Next year, Royal Philips Electronics will begin selling licenses to allow users to identify songs using their cell phones. Similar technology has already been open-sourced in FreeAmp, which uses the Relatable engine."
It seems kind of hopeless for all these companies and the audio fingerprinting. About a year and a half ago some company came out with a similar idea, except is was a separate device and you would have to wait until you got home and connect it to your computer. And it only worked with 2 radio stations. Obviously this idea was found to be a dud rather quickly. So now you can get the same information with a cell phone, this wouldn't by any chance be easier than actually calling the radio station and asking them what's playing? The only really effective method I've seen to do this are those giant billboards by the freeway with the screen that tells you what song is playing.
And what about the distortion this will cause in the audio? It's not like FM radio is great quality, but embedding some sort of watermark/fingerprint that only requires 3 seconds of playback has got to have some sort of negative impact on the sound.
Anyway, what the hell is up with all these trolls tonight? Time to start browsing at +2 now I guess.
The future isn't what it used to be.
no watermark/fingerprint/anything in the music - just an algorithm that samples it, creates a small chunk of data describing it and looks that up against a database for matches. So no loss of quality involved, and if the algorithms as good as they say it'll cope with quite badly broken up music (AM radio? Noisy club?) as input.
This could be a dastardly test-bed for keeping track of pepople using their voiceprints, since that technology, linked up to a global GPS satellite network, might just be "the end of privacy" as people are tracked down as soon as they speak on any telephone on the planet. Quick, call in Drew Barrymore, Lucy Liu and Cameron Diaz to foil them :)
Matthew @ Bytemark Hosting
Bitzi (the "another" link in the article's "yet another" statement) isn't primarily an audio fingerprinting application. It's a file metadata catalog, audio fingerprints being just one sort of metadata collected. File metadata is keyed by a "bitprint" composed of two cryptographic hashes. The code for generating bitprints and contributing metadata to the catalog is in the public domain and the catalog itself is available for free reuse and redistribution under a dmoz-like license. Disclaimer: I work for Bitzi.
I'm impressed if the service really does accurately identify songs after only 3 seconds, and any 3 seconds of a song. Presumably you'd need to fingerprint second or two chunks of every song to have this capability. This is quite different from what I understand Relatable does, which is to fingerprint the first 30 seconds of a song, meaning a song can only be identified after 30 seconds, and only the first 30 seconds.
Fast Search & Transfer has developed a "whistle sreach". Just whistle a few notes from the song you want, and their searchengine finds the songs mathcing.
I have tried it and can confirm it works really well.
Story from newscientist.com here. (cache)
Also a article from GEMENI here.
Is there any plans to combine the Bitzi database with FreeDB?
The first mental image I got from the head line was a picture of someone rolling a cell phone across a black ink pad... :)
And, correct me if I am wrong but this kinda says "digital watermark":
"The fingerprint might contain small mistakes. The technology is so robust that it can handle that," said Jaap Haitsma, a Philips research scientist.
Haven't we gone thru this already?
Seriously. The only thing missing was an SMDI challenge and the RIAA. Even though Microsoft is involved in this, I'm quite sure the RIAA will (pardon the pun) chime in Very Soon Now(TM).
Ok, quit possibly I am missing the point, but read this:
As well, a legitimate online music services running on the Napster model could use the technology to stop copyright-protected material from being shared.
Legitimate, Napster and stop sharing all in the same sentence?
Eh? The whole point of Napster was to share (leave out the legit or not) songs.
{Pitr voice}
Someone please to be explainink this to me.
{/end voice}
Moose.
.
Have you read the moderator guidelines? Well, have you, PUNK? (and I want a Karma: Gnarly option)
Can this actually tell two different Britney Spears songs apart? Even most humans can't do that...
At the recent ISMIR conference in Indiana, I saw a demonstration of a very similar thing from the Franhoffer (sp) people - it was pretty cool. Just to clarify some of the questions posed by above comments:
1) No tampering is done to the audio - ie there is no watermark, it is "just" signal processing.
2) The system I saw could take any 3 or 4 seconds, so that means that a fingerprint was calculated over the whole song. This means they must have some clever algorithms to make sure that the hash is calculated using the same time slices (or something...)
3) The song has to be in the database. So that means that the fingerprint has already been calculated, and probably had metadata assigned to it either by hand or using mp3 id3 tags (this is a guess). The fingerprint size is about 16KB per song, which seems pretty reasonable.
4) The technique only works on a per-recording basis, so even the same performer doing a slightly different version or another recording won't match if that recording was not already fingerprinted.
5) The version I saw was standard PC software, using audio input through the microphone.
My suspicion is that this technology is more likely to be of value to copyright holders looking to automatically identify violations (eg public airings, radio stations not paying royalties) than it will be useful to joe sixpack (or even people like us...)
in the "Name that tune" section of bar quizzes.
Boffoonery - downloadable Comedy Benefit for Bletchley Park
Including metadata in a digital stream is trivial, and radio should go digital soon enough. (There is not enough room for all the would-be radio stations to broadcast in the scant airwave real estate the way it is parceled now, but digital broadcasting would make more room for reluctant web-based radio stations.) Soon, the only non-digital (and therefore non meta-tagged) music will be stored in the brain. As others have stated, technology that can only identify playback isn't really that useful, but a program that could identify a song by a few hummed or whistled notes would be Really Cool.
As a musician, I value authenticity and would love to have an app that would prevent my brain from tricking me into thinking I had composed a remembered melody. Also, many people would love to know what that song stuck in their head is. However, is this enough consumer interest to sustain such a product? Furthermore, is there a chance that copyright lawyers will get out of control with this kind of power?
Remember that the way digital cell phones work is that they analyze the sound input, and send it not as audio data, but as coefficients to a human-voice synthesis DSP to save bandwidth.
Err, what?
You're talking about GSM encoding?
"The quality of the algorithm is good enough for reliable speaker recognition; even music often survives transcoding in recognizable form (given the bandwidth limitations of 8 kHz sampling rate)."
What's this about human voice synthesis DSP? Sounds a bit black-helicopters-faked-the-moon-landing to me. (or else I'm totally out of it - it happens! my cellphone here's still analog)
-- "Ignorance more frequently begets confidence than does knowledge." (Charles Darwin)
Gone are the methods of avoiding detection used to date.
Even if this detection has no way to discern between the original and a cover of the song, I can see the RIAA and major labels nailing a bunch of people, and using this system as proof.
Yes, but even though Bitzi is a general file catalogueing project, taking information from other, more specific sources can help (when their license allows it, of course)
M Y4OA2X.46IPPFIFT353PXN2BWBZEMYBF3ASZXTCWJN43RY
o ck&id=c011130f
For music files, this would of course be FreeDB. For movies, a good choice would be IMDB etc.
For instance, if I look up one of my Nightwish songs, Bitzi gives me the following information:
http://bitzi.com/lookup/TTGZBRZLZ2HLXDHSQYBTEJD33
the data from FreeDB is more accurate, giving the album the song belongs to, the playtime etc.
http://www.freedb.org/freedb_search_fmt.php?cat=r