Google’s artificial intelligence unit, DeepMind, has announced an AI tool that will voice AI-generated videos. However, the results are not very good for now.
Google’s Artificial Intelligence research lab, DeepMind, has introduced a new tool called V2A (short for “audio from video”) that voices AI-generated videos. According to DeepMind’s statement, V2A uses both the written description and the pixel-by-pixel analysis of the video while creating the sounds. This means it can generate sound even without a written description.
It should be noted that AI tools for voicing videos are important. Although many AI tools can create videos from text or images, these tools mostly produce silent videos. However, we all know that video is a medium that is as auditory as it is visual. When we see a video of a speeding train but cannot hear the sound of the tracks, wagons, and locomotive, the video loses its impact.
Example of V2A voicing an AI-generated guitar playing video
DeepMind claims that V2A is unrivaled in producing image-compatible sound and automatically synchronizing sound with the image. However, when we look at the examples shared, it is difficult to say that we are impressed. V2A produces sound that matches the image, but it seems to be dubbing a stock sound that is only somewhat in line with the image.
The examples below can help you understand how rudimentary V2A still is when it comes to dubbing videos. Still, it’s important to remember that DALL-E once drew a feathered yellow ball when asked to draw a canary. Artificial intelligence tools evolve as they are used.
You may also like this content
- Meta Building World’s Fastest AI Supercomputer for Metaverse
- Artificial Intelligence Will Make Decisions Instead Of People
- Bill Gates: Artificial Intelligence Over Web3 and Metaverse
-
-
-
-
-
-
-
-
-
-
