Song Identification
Our identification algorithm uses Fourier Transform to extract dominant frequencies
for every 100 milliseconds of the song, called song fingerprints. We later use
these fingerprints of the original songs to identify user input by matching both of
their
fingerprints. Our database keeps the fingerprints in hashed form for the fastest
results.
You may use the song identification feature to find out the details of the songs
playing around you by recording a 5-seconds-length sample of
them using our web recorder. After recording, our song identification algorithm will
find the song so that you can discover the song, its artist and listen to it again.
Building a Social Network
You may create a profile for building a social network around you. You can add new
friends,
see the songs they liked, get new recommendation of friends that have similar taste
in
music
as you. You may also let them know your taste of music by sharing the songs you
liked.
Recommendation
We analyze users' previous
song search and extracts the songs, artists, and genres users
like. Using this information, we identify songs similar to their music
taste and other users with similar music taste to recommend.
With our recommendation system, you may get recommendation of songs and discover new
ones. As you use our
identification algorithm and add new friends, our recommendation algorithm finds the
songs you may like based
on both your and your friends' search history.
