With the popularization of smartphones also came the popularization of tools like Shazam, which allow you to recognize songs by listening to a small fragment. This, however, has never completely solved the problem of having a song in mind but not a real audio fragment to be able to recognize it. Google claims to be able to fix it, its AI now recognizes songs by humming, whistling, or trying to sing them in some way or another.
In an event related to Google searches, the company announced different news related to its search engine and how it has improved and will continue to improve shortly. One of the most curious functions is probably the one that Google has announced regarding the detection of songs. They say that now the search engine will be able to recognize songs by humming them.
Singing, whistling, and humming
The new functionality is already available today in the Google application for smartphones and in Google Assistant. Just ask Google “What is this song?” (or derivations of this question) and later hum, whistle, or sing it. Google will show different results ordered from highest to lowest probability that they are the song you are looking for. From there it is possible to listen to the song to see if it is indeed that.
How does Google do this? As they comment, they use a machine learning model that converts what is hummed or whistled into a “numerical sequence”. After that, it’s a matter of comparing said number sequence with the existing and already scanned song sequences. The more the two sequences match, the more likely it is that it is the song you are looking for.
They say they have trained the learning model based on people singing, whistling, and humming. In addition, it does not take into account the instruments or the vocal quality of each user. It is the difficulty and the challenge of these functionalities of the Google search engine: understanding the human being when he is most human. That is, to get closer to natural language and not depend so much on the specific data that it receives. Like when he understands what we mean by “Aguanchu bi fri”.
The function is available today on Android and iOS through the Google app, as well as Google Assistant. At the moment in iOS, it is limited to English, but for Android, it is available in more than 20 languages, according to the company.
We have verified that it works in Spanish (you have to say the phrase and hum often, if we wait to hum it will tell us to download Shazam), and according to Google it will reach more languages, let’s see if it doesn’t take too long to expand to iOS as well.
Sharlene Meriel is an avid gamer with a knack for technology. He has been writing about the latest technologies for the past 5 years. His contribution in technology journalism has been noteworthy. He is also a day trader with interest in the Forex market.
