You Can Now Hum Songs Into Google Search

The Hum to Search feature uses machine learning to transform 10 seconds of humming into a unique "fingerprint," which it can match against an existing recording

Google has built a machine-learning feature into its Search bar that allows users to find out which song is stuck in their head — simply by singing, humming, or whistling the tune. The feature was rolled out in 2020, but has recently been added to YouTube, which is also a Google product.

When you hum your tune, Search identifies the pitches and rhythmic durations of the excerpt, and transforms this pattern into a sequence of numbers — which acts as the song's "fingerprint." This sequence is then compared to hundreds and thousands of other sequences yielded by music that has already been recorded. If there are close matches, Search will show you a number of options that might match the song that's stuck in your head.

To use the feature, open the Google Search app on your phone and tap the microphone icon, then select "search for a song."

Hum to Search is available on both Android and iOS.

"We launched Now Playing on the Pixel 2 in 2017, using deep neural networks to bring low-power recognition of music to mobile devices," wrote Google Senior Project Manager Krishna Kumar.

"In 2018, we brought the same technology to the SoundSearch feature in the Google app and expanded the reach to a catalog of millions of songs," he continued. "This new experience takes it a step further, because now we can recognize songs without the lyrics or original song."

"All we need is a hum."