FP TrendingOct 16, 2020 19:17:20 IST
Google has introduced a new search tool called ‘hum to search’. The feature allows users to hum, or sing or even whistle a song and let machine learning find out the track. Taking to a blog, Krishna Kumar, Senior Product Manager, Google Search revealed that Google can now help users figure out songs sans lyrics, artist name or even a perfect pitch. According to Google, users can now “hum, whistle or sing a melody to Google to solve your earworm.”
The feature is currently available in English on iOS and in more than 20 languages on Android.
As per the post, users need to open the latest version of the Google app or find your Google Search widget on their mobile device and tap the mic icon and say “what’s this song?” or click the “Search a song” button.
Once done, they need to start humming for 10-15 seconds. On Google Assistant, one needs to simply say, “Hey Google, what’s this song?” and then hum the tune.
The feature is currently available in English on iOS and in more than 20 languages on Android. Google hopes to expand the feature to more languages in the future.
Once a user has finished humming, the machine learning algorithm helps identify the possible song matches. One does not need to be pitch perfect for the feature to work and Google will show the most likely options based on the tune.
Users can select the best match and explore information on the song and artist, view any music video accompanying the song or listen to the song on one’s favourite music app. They can also find the lyrics, read the analyses and even check out other recordings of the same song.
As per Google, a song’s melody is like a fingerprint with a unique identity. Google has built machine learning models that can match the music to that unique “fingerprint.”
When a user hums, the machine learning models transform the audio into a number-based sequence representing the song’s melody. It compares the sequences to thousands of songs from around the world and identify potential matches in real time.
Credit: www.firstpost.com
