Adobe, VideoGigaGAN introduced an experimental artificial intelligence feature called. The company states that the feature can upscale the video by 8 times without the usual artifacts such as flickering or distortion.
According to Adobe, VideoGigaGAN outperforms other Video Super Resolution ( Leave your VSR) methods behind tiring. The feature also manages to add sharpness and detail.
The system produces details that do not exist from the ground up, and the details it adds, such as skin textures, fine feathers, bird feather details and more, are impressively realistic. It looks like.
According to Adobe researchers, the model is based on a large-scale image upsampler called GigaGAN. While previous VSR models had difficulty rendering rich details in the results, Adobe used “temporal attention” (reducing artifacts that accumulate over time), feature arc to create the final result. It combined moderation (adding detail where there was none), anti-aliasing, and something called “HF shuffling” (shutting out high-frequency features).
If this feature is added to products like Premiere Pro or After Effects, it will allow video producers to make low resolution footage look much better. It will be possible for it to work.
