Adobe, known for its creative software suite, has showcased a new tool that utilizes AI to modify the emotions conveyed in voice-overs. The technology, demonstrated by Oriol Nieto at Adobe’s MAX Sneaks event, enables users to alter the tone and style of a voice-over to better match the desired emotional context of a video.
The AI model analyzes video scenes, identifies emotional cues, and automatically inserts appropriate sound effects. For example, in a scene featuring an alarm clock, the AI added a relevant sound effect, enhancing the viewer’s experience. However, there were instances of imperfection, such as unrealistic alarm sounds or misplaced effects during character interactions.
Instead of manual editing, Adobe employs a conversational interface akin to ChatGPT to interact with the AI model and request specific changes. This intuitive approach streamlines the editing process and ensures precise modifications like adding ambient sounds to scenes seamlessly.
While these experimental features are not yet available to users, Adobe has a track record of integrating Sneaks event innovations into its software suite. This development comes amidst a backdrop of AI-related concerns in the creative industry, particularly regarding the use of AI to replicate voice actors’ performances.
Source: WIRED