Google recently launched a “hum to search” feature. Now you no longer need to live with that song tune stuck in your mind for days, you can just hum the tune to google and it will tell what you have been thinking of. Google will give you a list of songs that match the tune stuck in your head, you can then listen and find out. It is a cool and much-needed feature addition though.
Google says that the feature works by using its machine learning models to “transform the audio into a number-based sequence representing the song’s melody,” which it can then compare to existing songs. The company says that it trains these models on “a variety of sources, including humans, singing, whistling or humming, as well as studio recordings,” stripping away things like the instruments and vocal quality to focus just on that numeric sequence. Consequently, the hum to search feature should work whether you’re tone-deaf or have perfect pitch.
While there is understandable skepticism about this feature’s ability to provide the correct song title based solely on someone recreating the track through the use of whistling or humming, this tool gives users a chance to reconnect with the song they once thought was long gone.
The feature allows you to compare these sequences to thousands of songs from around the world and identify potential matches in real-time. For example, if you listen to Tones and I’s “Dance Monkey,” you’ll recognize the song whether it was sung, whistled, or hummed. Similarly, our machine learning models recognize the melody of the studio-recorded version of the song, which we can use to match it with a person’s hummed audio.