Adobe has unveiled Project Sound Lift, an AI-powered audio tool designed to enhance video post-production by allowing users to separate audio recordings into distinct tracks, including voices, non-speech sounds, and background noise.
Project Sound Lift integrates with Adobe’s existing Enhance Speech technology, currently available in Premiere Pro. Although I have not tried it yet, if it works as promised, this tool will be particularly notable for its ability to handle real-world recordings, which often contain a mix of noise, reverb, and overlapping speakers – conditions that have traditionally challenged audio AI models.
This kind of AI-based signal processing is not new; there are several plugins and standalone apps that offer similar features. In fact, this is the type of tech used to separate John Lennon’s voice track from a mixed audiotape so they could release “Now and Then” as a finished track.
If you’re wondering… this type of audio “magic” is not possible without AI. Imagine combining a cup of red liquid and a cup of blue liquid. The result would be two cups of purple liquid. How would you separate the two cups of purple liquid back into a cup of red and a cup of blue? Sound Lift is the audio equivalent. It takes any audio track and lets you separate the component parts into separate audio tracks. It’s super fun and super useful. I can’t wait to try it.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.