Google has released a new app called blob Opera—a machine learning application that allows songwriters of all skill levels to compose songs—operas or otherwise—performed onscreen by four singing “blobs.” In announcing the new app on its website, Google suggests such compositions might just be the perfect holiday present. The app was created by artist David Li.
The idea, Freya Murray, program manager for Google Arts & Culture suggests, is to explore what she describes as humanity’s first instrument—the voice. The app is available on Google’s Arts & Culture, EXPERIMENTS page. Users can program four singing blobs with assistance from a behind-the-scenes machine learning algorithm. Programming is done by dragging sliders for each of the blobs for different parts of the score. blob Opera then puts them all together (in melodies and harmonies) into a composition and the blobs perform it. The blobs are colorful armless and legless creatures with mouths, eyes and Santa hats. They sing upright next to one another, as if a small choir.
The blobs have been programmed to replicate the style of opera singers by training them with operas sung by real singers: Christian Joel, Frederick Tong, Joanna Gamble and Olivia Doutney. And to that end, each of the four blobs performs one of four singing voice ranges: tenor, bass, mezzo-soprano and soprano. The performances are not the voices of the human singers—instead, they are generalized versions of opera singers based on training.
Users who wish to experiment more deeply can edit their songs by changing pitch, vowel sounds and the notes they originally chose when composing a song. They can also take over during a performance, allowing one of the blobs to sing a solo on the fly. They can also share their performances with others. And for those who are not pleased with their compositions, the blobs also perform standard songs such as Christmas carols.
blog.google/outreach-initiativ … ive-song-blob-opera/
© 2020 Science X Network
Google releases Blob Opera just in time for the holidays (2020, December 17)
retrieved 22 December 2020
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.