YouTube Is Testing a Hum-to-Search Feature for Song Identification

YouTube announced a new experiment on its Android app. The new feature that is being tested enables users to identify songs by humming, singing, or recording a brief snippet of music they are listening to.

YouTube to Launch an Ad-Free Video Player for Education
It was created specifically for educational organizations that often use YouTube as a platform for learning. Educational institutions will now be able to embed videos on websites without distracting students from the online learning experience.

Users taking part in the experiment can switch to the new song search option on the app and hum or sing a song for at least three seconds. Leveraging machine learning models, YouTube then identifies the tune and directs users to relevant YouTube videos associated with the song.

The new feature distinguishes itself by its swiftness in identification, requiring only three seconds of input. The technology behind this feature is based on machine learning algorithms that match hums or sung notes to a song's unique "fingerprint" or signature melody.

YouTube Is Testing AI-Generated Video Summaries
This test feature, currently limited to specific English-language videos and a selected user base, aims to provide brief overviews of video content without replacing the original human-written descriptions.

Although this feature is currently limited to a small number of Android users, its potential widespread implementation could revolutionize how people search for music and interact with the platform.