Adobe announced a new upgrade to its Firefly video model on Thursday, introducing a new third-party artificial intelligence (AI) model that is available on the platform. The California-based software giant said it will now improve motion generation on firefly video models to make it even more natural and smooth. Additionally, the company adds advanced video controls to allow users to produce more consistent video output. Additionally, Adobe has also introduced four new third-party AI models that have been added to the Firefly board.
In a blog post, the software giant detailed the new features and tools Adobe Firefly users will receive immediately. These features are only accessible to paid subscribers, some of which are currently limited to web apps.
Adobe's Firefly video models are already creating videos with realistic, physics-based movement. Now, the company is making enhanced motion generation capabilities to provide a smoother, more natural transition. These improvements apply to both 2D and 3D content, increasing the motion fidelity of not only characters but also elements such as floating bubbles, rustling jewels, and drifting clouds.
The recently released Firefly app also has support for new third-party AI models. Adobe presents Upscalers and Moonvalley's Marey, images and videos from Topaz Labs. These will be added to your firefly board soon. Meanwhile, Luma AI's Ray 2 and Pika 2.2 AI models are already available on the board and will support video generation capabilities immediately (currently only available for image generation).
Coming to the new video controls, Adobe added additional tools to encourage prompts and reduce the need to do inline editing. With the first tool, users can upload videos as references, and Firefly follows the original configuration with the generated output.
Another new inclusion is the Style Presets tool. Users who generate AI videos will now be able to select styles such as Claymation, Anime, Line Art, 2D along with prompts, and Firefly adheres to the instruction of the style of the final output. Keyframe trimming is now possible even in the prompt stage. Users can upload the first and last frames of the video, and Firefly generates videos that match the format and aspect ratio.
Apart from this, Adobe is also introducing a new tool called Beta-generated Sound Effects. This tool allows users to create custom audio using audio or text prompts and overlay it on AI-generated videos. When using voice, users can also determine the timing and intensity of the sound, as they generate custom audio that matches the energy and rhythm of the sound.
Finally, the company is also introducing text into the avatar feature that converts scripts into avatar-driven videos. Users can also select their preferred avatar from Adobe's pre-listed library, customize the background, and select the accents of the generated speech.
