YouTube tests a search feature where users hum to identify songs

YouTube is testing a new feature on its Android app that allows users to identify a song by humming, singing, or recording it. This search-by-song capability, currently available to a limited number of users, identifies the song and directs the user to relevant YouTube videos. The technology used is similar to a feature launched by parent company Google in 2020, which uses machine learning models to match a user's hum to a song's signature melody.

The new feature is seen as a significant upgrade from other music recognition apps like Apple's Shazam, SoundHound, and MusixMatch. If rolled out more widely, it could prove beneficial for many users, given YouTube's popularity as a platform for looking up songs. The main difference from Google's feature is that YouTube's version requires users to hum for only three seconds, compared to Google's 10-15 seconds.

Key takeaways

YouTube is testing a new feature on Android devices that allows users to identify a song by humming, singing, or recording a song.
The feature is currently only available to a small portion of Android users and directs users to relevant YouTube videos featuring the searched song.
This technology was first launched by Google, YouTube's parent company, in 2020, but required users to hum for 10-15 seconds to identify a song.
The new YouTube feature uses the same machine learning models as Google's feature, which matches a person's hum to a song's "fingerprint" or signature melody.

YouTube tests a search feature where users hum to identify songs | TechCrunch

Key takeaways

Discussion (0)