Adobe Firefly is now capable of creating music, speech, and sound effects.

      Adobe Firefly’s audio tools became widely available on Thursday. The tools for generating music, speech, and sound effects have now joined the image, video, and design features in its creative AI studio. Each tool operates on a distinct model: Generate Music utilizes the Firefly Music Model to create original tracks that align with a video's length and mood, Generate Speech employs the Firefly Speech Model with an option for ElevenLabs, allowing customization of voice, pacing, and emotion, while Generate Sound Effects uses the Firefly Audio Model to synchronize sounds with the timing of a clip.

      Adobe highlights the primary applications for these tools, including social media content, live vlogs, short films, product tutorials, and podcast segments, and there’s no separate subscription fee for accessing them.

      According to Adobe, they aim to address the challenge of switching between tabs. Creators often shift across different workspaces, applications, and services, disrupting their workflow, especially with sound, which is typically the last component to be incorporated and the first noticed by the audience.

      The feature set isn’t particularly groundbreaking; numerous tools already generate music from prompts. However, Adobe emphasizes the specific selling point throughout its announcement.

      "The product is the license." The phrase “commercially safe” recurs frequently. Adobe asserts that Generate Music produces “universally licensed original tracks.” This means that the music can “travel with your content without concerns about takedowns.” The audio is “designed for actual production.”

      This focus on legal safety rather than sound quality is the central message. Adobe trains Firefly models using licensed and public domain material, enabling it to make such assurances. Since the launch of Firefly, this differentiation has underpinned the company’s generative strategy, and audio is the latest format to be introduced.

      The timing of this release appears strategic. In July, a German court ruled that Suno violated copyright, marking a significant legal decision in Europe, brought forth by the collecting society GEMA. Since then, Suno has been facing increasing legal challenges, leading to watermarking and fingerprinting its output and limiting downloads. Nonetheless, AI-generated music continues to emerge and has entered the charts this year, with artists openly using it. Adobe is entering a market ready for this technology, where demand is evident, leaving the lingering question of whether a track will withstand scrutiny from rights holders.

      Two creators mentioned in the announcement highlight this commercial issue, noting that licensing background music poses a recurring challenge when collaborating with brands, with both testimonials provided by Adobe.

      Adobe is also reselling tools from its competitors. Firefly doesn’t rely solely on Adobe’s models; it incorporates technologies from Google, ElevenLabs, Kling AI, Luma AI, OpenAI, and Runway, with Google’s Gemini Omni Flash added recently. This model accommodates video, audio, and image inputs alongside text, allowing a rough concept to evolve into a storyboarded draft in a single interface.

      The ElevenLabs partnership is particularly noteworthy. Adobe developed its own speech model but chose to include a competitor’s as well. ElevenLabs was rumored to be in discussions for a tender offer at a valuation of $22 billion in July, signifying that it is not a minor player being quietly assimilated.

      Adobe seems content to serve as the platform rather than focus on specific models, as long as work occurs within its environment. This represents a different bet compared to many AI companies, and it only pays off if users initiate their projects within Firefly.

      However, this raises an interesting contradiction in the licensing narrative. Adobe can guarantee tracks from its own music model, but extending that assurance across a studio utilizing six other companies' systems is more complex, making Adobe's claims of commercial safety particularly tied to its own Firefly models.

      Additionally, Adobe has now made the Adobe Firefly AI Assistant freely accessible, with a no-cost tier and daily generation limits. The assistant was previously introduced in beta earlier this year, gaining creative capabilities in June when Adobe unveiled its largest AI initiative yet. Among its most utilized features are Create Storyboard and Create Brand Kit, with the assistant also managing batch edits and branded mockups. A related feature called Elements retains characters, locations, and objects across projects.

      The availability of the free assistant and the lack of a dedicated audio subscription indicates that Adobe is cultivating user habits, banking on the idea that creators who assemble projects within Firefly will eventually pay for the finishing touches.

      Adobe supports its claims with research from the Berklee College of Music, which found that nearly four in five video creators, musicians, and marketers post video content daily or several times a week, with every participant confirming the use of music in their videos. Adobe financed the study, and its own footnote notes that the company “did not control respondent recruitment, survey responses, analysis, or conclusions,” with the Berklee lab maintaining editorial control of the report. This disclosure is more extensive than many statistics provided by vendors, yet it is still a figure that Adobe paid for, used