Adobe has unveiled a groundbreaking new upscaling technology that utilizes AI to enhance video quality significantly. This new feature, known as VideoGigaGAN, promises to upscale videos by up to eight times without the typical artifacts such as flickering or distortion commonly seen with other methods.
Unlike traditional Video Super Resolution (VSR) techniques that often introduce unwanted artifacts, VideoGigaGAN focuses on maintaining image sharpness and detail. By implementing advanced technologies such as temporal attention, feature propagation, anti-aliasing, and HF shuttle, Adobe’s researchers have managed to create a system that produces impressive results.
While VideoGigaGAN excels at enhancing videos with richer details like skin textures, fine hairs, and intricate patterns, it is essential to note that the system generates some details that may not exist in the original footage.
This AI-powered upscaling technology has the potential to revolutionize video production by enabling creators to enhance the quality of low-resolution footage significantly. However, it is worth mentioning that the use of AI to enhance people in videos remains a controversial topic.
As Adobe continues to explore the possibilities of integrating VideoGigaGAN into its products like Premiere Pro and After Effects, other tech giants such as NVIDIA, Microsoft, and Blackmagic Design are also actively developing similar upscaling solutions. The competition in this space indicates a growing demand for advanced video enhancement tools in the industry.
