Loading...
Logo
Processing Forum
Hello Guys

For an Exhibition I'm working on a Quiz with Processing. Participants have to identify them selfs, but not by name, RFID or barcode. No, even crazier: by playing a song from a record.

I want to install a recordplayer (turntable, phono) and every participant get's a record at the entrance. He puts the record on the turntable and my code should compare and recognize the played song with a Library of about 100 (prerecorded) songs.

By looking arround I found this blogpost from Roy van Rijn. http://www.redcode.nl/blog/2010/06/creating-shazam-in-java/

He describs how it works very nice and clear, but since I'm verry new to processing an Java I need a lot of help. That's why I'm looking for people also interessted in this.

Simon


Replies(5)

That's precisely the kind of question that doesn't fit in the Programming Questions section as it is defined (to ask questions about own Processing code shown here), it totally belongs to General Discussions (as it defines a general, hard problem and ask how to make it). I moved it.

Beside, the question is interesting, and you provide at least a link perhaps answering partially the question. I will read the article with interest. Thanks.
Argh, a typical Newbie mistake. Please forgive me, but now I know where to put such posts in the future. Thank you for the moving.
The reading was interesting. Funnily, it links to a page explaining the algorithm, linking to another page showing it as explained by the Shazam engineers... That's second-hand information! But seriously, each one is useful.

I think that part of the code shown in the Java implementation can be handled by Minim or another Processing audio library. Thinking of audio input, FFT and such.
The remainder of the code is perhaps the difficult part, but seems doable. Will need some time to sample the songs used to build the database, though.
It's cool, isn't it. He did a great job and especially, if you know what happend after his weekend project ... Especially if you watch his Parley on the devoxx conference, where he describs how he got the idea http://www.redcode.nl/blog/2012/03/devoxx-2011-talk-freely-available/

But back to my topic, to be honest I already stucked at the beginning. I could capture sound with an example sketch, but how can I put it in an array, to simply get the values ... Would so apriciate, if someone tries it too.
As written, I think Minim can do that, but I have little experience with it.
Look at the example sketches coming with Processing, I think there is one showing the sound data as a graph, so it probably extracts numerical data from it. Also take a look at the Minim home page, their JavaDoc, etc.