AMD has integrated a new engine into FFmpeg, the popular free video program, that allows using artificial intelligence to improve image quality, detect objects, or separate backgrounds. This feature is compatible with processors, graphics cards, and AI units from AMD and other brands. For video editors, this means faster and more accessible processes on common computers, without the need for specialized hardware.
How the new AI engine in FFmpeg works 🚀
The engine is based on Vulkan and DirectML, allowing AI models to run on varied hardware, from AMD Radeon GPUs to NVIDIA GeForce. FFmpeg can now apply filters such as super-resolution, noise removal, and object segmentation in real-time. The advantage is that open source makes it easy for developers and users to integrate these capabilities into existing workflows. Everything is documented so anyone can try it without being a machine learning expert.
Now even your grandma can separate backgrounds without crying 😅
Yes, because before, separating the background of a video with AI required sacrificing an entire weekend and an offering to the GPU gods. With this FFmpeg update, any mortal with a normal PC can do it while making coffee. That said, don't expect the AI to fix that blurry video from Christmas dinner: for miracles, better talk to another saint.