Adobe has announced two new AI audio tools in its Firefly app: Generate Soundtrack and Generate Speech, reports The Verge.
With Generate Soundtrack (public beta), you upload a video and the tool analyzes it to produce synchronized background music. You can choose styles like lo-fi, hip-hop, EDM, or moods such as “sentimental” or “aggressive,” then receive four distinct variations, each up to five minutes long. Importantly, Adobe says the model was trained exclusively on licensed music, ensuring the output is cleared for commercial use, and sidestepping copyright issues that have tripped up other providers.
Generate Speech (also public beta) supports more than 50 voices in over 20 languages. It lets you convert text into voice-over narration, adjusting parameters such as speed, pitch, emotion, and pronunciation. This simplifies adding voice-overs to video content without needing voice actors or recording setups.
Adobe is also developing a web-based version of Firefly for video editing: a multi-track timeline interface integrating these audio tools, titles, and editing capabilities. This promises to streamline workflow for creators who want a unified suite rather than using separate tools for music, speech, footage, and editing.
For content creators, educators, social-media producers, and other professionals, this means faster turnaround and fewer technical barriers in video creation. For engineers and production teams, the commercial safety of the training set matters, reducing the risk of takedowns or licensing disputes. In the broader scheme, Adobe’s move signals that audio generation is joining image and video in the mainstream of generative AI tools.