Adobe’s New AI Works Like a Photoshop for Audio
Adobe's experimental Project Sound Lift can take audio recordings and isolate voices or remove distracting background sounds
Adobe has already demonstrated experimental AI-powered tools that can dissect an image, making it easy to remove, replace, or adjust the appearance of objects and people. Project Sound Lift promises to do the same for audio, disassembling recorded audio into pieces so sounds can be remixed or removed entirely. For example, you could take a finished song, pull out the bass track, remix it, and place it back in, essentially using the tool to make different audio "layers," like a photoshop for music.
Adobe shared a demonstration of Sound Lift to YouTube today, although it appears to simply be a clip from the many sneak peeks of experimental features the company presented at its annual Adobe Max conference in October.
Professional recordings are typically made with many microphones, each capturing a very specific sound separate from the others. This allows each sound, such as the instruments and performers, to be individually adjusted and mixed together at varying levels for a final recording.
When capturing audio using a device like a smartphone or even professional-grade cameras with just a single microphone, all the sounds are recorded as a single track, permanently merged together so specific sounds can't be removed or adjusted. It's like trying to remove a single ingredient from a baked cake.
Project Sound Lift leverages AI to extract and isolate specific sounds from a single recording. In one demonstration, the tool is able isolate the voice of an announcer in a noisy auditorium, the talent speaking on stage and even the applause of the audience, which otherwise overpowers everything. In another example, the sounds of people talking in the background can be removed so that only the person speaking directly to a camera is heard.
There are already audio processing tools that can eliminate certain frequencies from a recording, allowing, for example, the hum of a fan to be removed. With Project Sound Lift, an audio recording can be disassembled and brought into an editing tool where individual sounds can be adjusted and remixed at different levels. With a recording of a live concert, you may not want to completely eliminate the sound of an excited crowd, but with Project Sound Lift, you can ensure that over-zealous fans aren't drowning out the talent on stage.
- Photoshop Officially Comes to Browsers With AI Tools and a New Paywall
- Adobe’s New Label Sorts Real Images and AI Content
- Adobe’s Generative AI is Catching Up to Dall-E
- Adobe’s $20 Billion Bid for Figma Hits a Roadblock in Europe
- Record A Deepfake In Drake’s Voice? YouTube Reportedly Working On Official Audio Cloning
- Book Authors Demand Big Tech Pay for Using Copyrighted Works To Train AI
- Tetris Creator Stuns 13-Year-Old Who Beat the Game With Surprise Zoom Call: ‘This Is So Cool’Tech
- Apple May Face Sweeping Antitrust Lawsuit From Justice Department Over iPhone: ReportBusiness
- Elon Musk’s Take on DEI Slammed by Azealia Banks: ‘No, Stupid’Entertainment
- A Real-Life Spider-Man? This Engineer Made Amazing, Spectacular Web ShootersTech
- This $2,149 Smart Toilet Seat Brings Alexa Into Your BathroomTech
- Why the Next Moon Landing May Be the Most Important Since Apollo 11Tech
- How My AI Coach Helps Me Stay on TrackBusiness
- The BlackBerry Dream Lives On in This iPhone Keyboard CaseTech
- You Can Buy Your Own Custom Version of ChatGPT Next WeekTech
- Mesmerizing Videos Reveal Stormy Weather on a Hellish Planet Where It Rains IronTech
- Largest Known Male of World’s Deadliest Spider Captured in AustraliaTech
- There’s a New ‘Jaws’ Pinball Machine, and We’re Going to Need a Bigger BallTech
