Google Search helps you find that earworm by humming

Sometimes when you're trying to look for a song, you can only remember the melody or the cords. Google wants to help you find the track you're looking for with a new feature coming to Search. It will let you hum, whistle, or sing a melody to identify the song for you. All you have to do is tap the mic icon on the Google Search widget or head to the Google app and ask, "What's this song?" or click the "Search a song" button. 

The feature is available in English on iOS and in over 20 languages on Android, with plans to add more languages in the future.

Google uses machine learning to help identify the song matches, and no, you don't need perfect pitch to use the feature. The app will show you the most likely options. And then it'll also show you information on the song, artist, accompanying music videos, lyrics, links to listen to the song on your music app of choice, analysis of the track, and even other recordings of the song available.

Google briefly explains in a blog post how it can identify a song with just your humming:

An easy way to explain it is that a song's melody is like its fingerprint: They each have their own unique identity. We've built machine learning models that can match your hum, whistle or singing to the right "fingerprint."

When you hum a melody into Search, our machine learning models transform the audio into a number-based sequence representing the song's melody. Our models are trained to identify songs based on a variety of sources, including humans singing, whistling or humming, as well as studio recordings. The algorithms also take away all the other details, like accompanying instruments and the voice's timbre and tone. What we're left with is the song's number-based sequence, or the fingerprint.

We compare these sequences to thousands of songs from around the world and identify potential matches in real time.