Tech startup ElevenLabs has launched an artificial intelligence (AI) program to automatically generate ambient sounds based on text commands. This situation was officially reported page Companies in social network X (formerly Twitter).
The authors of the new neural network demonstrated its capabilities using the example of silent video created using OpenAI’s Sora generative model.
“We used text cues like ‘waves crashing’, ‘metal clanking’, ‘birds chirping’ and ‘race car engine’ to create the sound and overlaid them with some of our favorite clips from the OpenAI Sora announcement.” at ElevenLabs.
The demo video shows other examples of the algorithm in action, including the noise of a metropolitan street, the mechanical hum of a robot, and the barking of puppies.
ElevenLabs is known as the developer of an artificial intelligence system that converts text into synthesized speech and automatic video dubbing that supports more than 20 languages.
Previously OpenAI presented Sora neural network that produces photorealistic videos based on text description.