Adobe is taking generative creativity to the next level with its latest AI-driven filmmaking tools — including a unique feature that turns silly voice noises into believable sound effects.
As part of its Firefly app, Adobe has launched a beta version of the “Generate Sound Effects” tool, which allows users to use onomatopoeic voice recordings (like “clip clop” or “snap”) to create realistic audio synced with video. Paired with a short text description — for example, “hooves on concrete” — the tool generates four sound effect options that align with the video’s motion. This offers more flexibility than Google’s Veo 3, especially when syncing custom audio with visuals.
This new tool builds upon Adobe’s earlier “Project Super Sonic”, first previewed at the Max event in October. While it doesn’t support speech synthesis, it does enable users to generate a wide range of impact and ambient sounds, such as footsteps, zippers, twigs snapping, and city or nature ambience.
In addition to sound generation, Adobe has upgraded its Firefly text-to-video generator with more refined controls:
- Composition Reference: Users can upload a video clip to guide the layout of the AI-generated footage.
- Keyframe Cropping: Creators can upload the first and last frames, and the tool will generate the in-between footage accordingly.
- Style Presets: Users can apply quick visual themes like anime, vector art, and claymation to their video output.
Although some results, like the “claymation” style, didn’t impress during live demos, Adobe is continuing to evolve Firefly’s capabilities, aiming to remain a leader in creative software. According to Adobe’s Generative AI head Alexandru Costin, the company plans to extend these features to third-party AI models, opening more options for users.
In a competitive space where OpenAI and Google are leading in raw model development, Adobe is focusing on enhancing creative control and integration to maintain its dominance among creative professionals.
