PDA

View Full Version : help with .NET project for music recognition


randyisarealboy
11-30-2008, 04:47 AM
Okay,

Just becuase I happened to be on the forums... this question I'd normally post on an MSDN forum somewhere but .. since I'm browsing and the question is burning; I thought I'd give it a shot.

First off, if there are any .NET programmers out there that also watch HAK5 - does anyone know of an SDK that allows for pattern/sound matching?

What I mean is something similar to the products you see now on the IPhone - where you hold your phone up to your radio, TV, or whatever, press record for 20 seconds and then give it a moment to think ... while it does that it searches a database of songs and then returns to you the name of the song that you heard.

.. Is there an SDK for that? And if not, is anyone aware of any open source project that is similar to that that could be ported to the .NET framework?

I'm actually not interested in building another song recognition program but I've an idea for the same technique applied in a different application.

Any suggestions? Anyone at all?

Bani-Banan
12-06-2008, 02:14 PM
I don't believe there is an audio-recognition "SDK".

Because computers are dumb, it's very tough to implement audio-recognition, and a database.

I doubt your question will get answered here, and on the MSDN forums.

I will be watching this thread for more info.

Good luck!

GoNZooo
12-08-2008, 11:56 PM
Theoretically, neural networks should be able to do great sound recognition. I don't know how a NN would do in practice, most of the time, but pattern recognition is what NNs are perfect for.

Pick up some books, read them and understand them and make this. ;)

randyisarealboy
12-30-2008, 07:32 AM
and no money, so that means my vacation is stuck in this apartment coding...

for those curious, I haven't gotten very far, but I have got a decent thread going on codeprojects:

http://www.codeproject.com/script/Forums/View.aspx?fid=1649&msg=2858413

that's my thread on the Guitar Tuner. I'm currently dumbfounded on how to interpret the FFT magnitude results - if I'm sending in the wrong kind of data, or if my mic on my PC just plain sucks.

Or if it's the coder.

*shrug*