- In addition to the score, they also create sound effects and dialogue that “fit the character and tone of the video.”
-
Google has announced an update to its video-to-audio technology. The software now enables “synchronized audio-visual generation,” allowing it to generate soundscapes, effects, and dialogue for videos, archival material, silent films, and other types of footage. The generated audio “matches the character and tone of the video,” Google said. Said A new blog post explains that an unlimited number of soundtracks are possible for any video input. Unlike existing video-to-audio solutions, Google's technology can understand raw pixels and doesn't necessarily require text prompts, the post states. “Video generation models are advancing at an incredible pace, but many current systems can only produce silent output,” the post states. “Early results show that this technology could be a promising approach to bring generated movies to life.” The tool is still undergoing safety checks before being released to the public, and is part of Google's AI lab, DeepMind. Check out some test clips below.
For more information, read Google's blog post.
