Google's Blob Opera AI Sings Carols and Helps Make Music
Opera lovers can finally take an active part in their passion by creating it with the help of artificial intelligence. The Google Arts & Culture initiative called Blob Opera enables us to create opera music in a very simple way.
If you have an artistic soul, and your dream is to compose music, then Google's latest experiment belonging to their project Arts & Culture will surely appeal to you - especially if you wanted to create opera compositions. The non-profit initiative called Blob Opera enables you to compose music with funny "blobs". With this tool, you can play Christmas carols, only with your mouse cursor. If you want to take part in the experiment, you can do it here.
Using the musical drops
After starting the experiment, you will receive a tutorial in which you will learn the basics of using musical blobs. These will drop one by one as soon as you learn to use the previous one. The blobs in question represent different voices. Starting with bass, through tenor, mezzo-soprano and ending with an ordinary soprano. Each blob has different singing heights and volumes, which can be modulated in different ways by moving the cursor up, down or sideways. You can also mute them with one mouse click. An important element here is the button responsible for the recording, because only after starting the recording will the sequence be saved. After the composition is finished, we can play back our song without the need to interfere with their behavior.
All the blobs are interactive - they are flexible depending on the sung voice, and their eyes follow the cursor. The tool also includes a Christmas mode, in which the blobs don Santa's caps and it starts snowing in the background. When enabled, we also have access to carols, which, like our own compositions, can be modulated.
How does it all work?
The experiment was created by David Li in cooperation with Google Arts & Culture. David used machine learning to create artificial intelligence that will sing on our command. The whole thing was to allow for the creation of beautiful songs without the need for musical knowledge or experience. However, the sound is not created by the computer. The author invited real opera musicians to the studio, who sang the appropriate sequences with their voices. The recording, which lasted 16 hours, was imported into a convolutional neural network, which learned how to reproduce every kind of voice from what it heard. This was followed by interactive blobs and the ability to modulate sound based on pitch and volume. Finally, we got the Blob Opera, a tool that enables us to play opera voices freely.
