How Do Music Recognition Apps Work?

Most people remember that moment when they first used the Chocolate phone or discovered the Shazam app. When it could successfully tell you what song was playing, it got people really excited. This is probably because people didn’t know how it works. It felt like black magic. However, this technology is actually pretty simple although it is very interesting.

According to MoneyPug, a site used to find mobile phone deals, since the LG Chocolate, music recognition has come a long way. If you have seen the show Silicon Valley, you might remember that Pied Piper began as a music recognition app using a very effective compression algorithm. Music recognition has been around a lot longer than that, Shazam actually began in 1999. It has been used over 50 billion times to identify songs wince then. How does it work? What makes it such a popular app?

How Music Recognition Works

Music recognition apps begin with a large database of information about every song they can find. You use the interface to find exactly what song is playing when you are curious by storing song data, its unique patterns and notes, making an audio fingerprint of the song based on just a few seconds, and creating an interface for the user to interact with. The data is stored as a spectrogram, which is a graph that represents sounds. The x-axis of the graph symbolizes time and the y-axis represents frequencies. Finally the amplitude is represented by color. As it turns out, any sequence of sound can be converted into a spectrogram. There are assigned sets of coordinates attached to it any time.

Music notes are turned into numbers, and the spectrogram is stored into the database as the data number. Every song is stripped to just a few data points. To hash out a sequence of peaks is taking a set of inputs and run them through the algorithm that assigns an integer output, the distance between the two highest peaks and put the two frequencies together. This makes them easily identifiable, storable, and searchable. At the end of it all, the song will have a unique 32-bit number that becomes its ID within the database.

The Breakthrough

Of course it was a realization that we can identify sounds with data. The people working on this technology found out that getting rid of a song’s lower energy parts decreased the spectrogram and makes the app less susceptible to dull sounds and background noise. This technology has continued to change, and different apps take on various angles to weed out their competition and become a worthy competitor to the dominant app Shazam.

Music Recognition Apps

Soundhound is a music recognition app that is similar to Shazam. It has features like voice control as well as a tab of charts that helps you find features on its music player. With this you can play samples, tag songs, read lyrics, and find the album the song is on. Another feature is that you can sing the song to the app to identify the song. Musixmatch is a music recognition application that focuses more on lyrics. The app helps you find music by the lyrics instead of the song data. It will allow you to save lyrics offline.

Of course Shazam is the big dog on campus when it comes to music recognition apps. It is by far the most known and used application using this technology. This is probably because it remains by the best app in the industry. It has a simple interface that users like and a panel that helps you discover new music, read music news, and find new releases. You can hear previews or add songs to your favorites. It gives you access to tons of information about the song and artist. With Shazam, you can listen to samples, share songs with your friends, and find the music online on Spotify or Apple Music. People love Shazam because it was designed for music lovers.

Music recognition uses technologies that could end up being used for other purposes, but for now it is used to help people find out what song is playing to listen to it later. Whether you are interested in the technology of it, most people are intrigued by Shazam and use it to find songs they really like

Al Hilal: