At its annual MAX conference, Adobe is introducing new AI-based video features alongside the latest Firefly AI image model 5. Audio capabilities, in particular, are being expanded. In July, AI-generated, royalty-free sound effects were already introduced; now, Firefly can also artificially generate instrumental background music and synthesize voices for voice-over speakers.

AI Soundtracks in Firefly
AI Soundtracks in Firefly
The new AI function "Generate Soundtrack" in Adobe Firefly analyzes uploaded videos and then creates suitable music of any length. According to Adobe, these royalty-free soundtracks can be used without hesitation even in commercial projects.
After analyzing the video to be scored, Firefly suggests various prompts for generating the musical accompaniment. Alternatively, users can enter their own description for a custom soundtrack. A tag-based prompt builder helps define the vibe, style, genre, and purpose, as well as set the length and tempo.
Various versions are then generated, from which the desired track can be downloaded and used as music (WAV) or video.
AI Voice Generator
Speaker voices for commercially usable voiceovers can now also be generated directly in Adobe Firefly. In addition to Adobe's own Firefly Voice Generator, users can also access the newly integrated Elevenlabs Multilingual v2 model.
The AI-based "Generate Speech" feature supports over 20 languages, with up to 70 different voices available for selection. The text template is simply entered via copy/paste or uploaded in DOCX or TXT format. Emotion, tempo, emphasis, and pronunciation can be specifically adjusted.
These new, generative audio tools are now available as premium features for Firefly users, requiring generative credits.