What are some good libraries for extracting data from audio files?

Question

I recently started using the Shazam app on my iPhone. For those who do not know this, this application identifies songs by listening to a small segment of a song. I was amazed at its accuracy and speed, so I decided to work a little.

I found an article written by one of their developers here . In the article, the developer describes in detail the fingerprintng algorithm used in Shazam.

As a favorite project Id like to make my own fingerprint application, so I can get some experience in the audio program.

What are some audio libraries that help you extract things like frequency, amplitude and other characteristics of an audio clip or mp3 song over a duration?

I use .NET, but I am open to other languages. I am also well versed in both open source and paid libraries. As long as I can reliably extract audio characteristics, I will be happy.

See also:
How does Shazam work
Shazam Magazine

Answer 1

Try a look at NAudio. You may not have all the audio analyzes you are looking for, but it is quite extensible, and it would be nice to get started if you use .NET languages.

Answer 2

To start with audio features, you must first read this article .

Many laboratories have developed their own libraries for extracting audio functions. You can take a look at yafee, aubio, jaudio ....

Answer 3

The ffmpeg library supports many audio codecs, but it is very painful to interact with IMHO.

To extract audio properties, you should consider a decent library suitable for signal analysis. In particular, you will need the Fast Fourier Transform (FTT) transform, for example, to extract frequency data from your sound samples. The search yields a lot of results on this topic.

/ edit: for .NET, I'm sure there is a ffmpeg interface. You will also find signal analysis tools for .NET.