Comment on page
AI Enhance [BETA]
AI Enhance is a deep learning super-resolution technique used to upscale Wave 2K clips to 4K. It is a Deep Neural Network that has been custom-trained for the Wave 2K frames, to take into account the subsampling used by the Wave sensor to achieve higher frame rates in 2K. So, the Ai Enhance can reduce subsampling and aliasing artifacts around edges during the super-resolution process.
With AI Enhance enabled, 2K clips (Width setting of 2048) will be processed and exported at the equivalent 4K resolution with the same aspect ratio.
A dedicated graphics card with at least 4GB of VRAM is required to use AI Enhance. The deep neural network is implemented in GPU hardware, so a higher-end desktop GPU, or M1 Pro or above will greatly improve the frame rate achievable with AI Enhance turned on.
The figure below shows how AI Enhance improves details on a 2K recording.
Comparison showing Ai Enhance: notice the detail in the bumblebee's wings, legs, and eyes.
AI Enhance is still a preliminary feature and improvements are still in-work. AI Enhance works best on scenes with clear focus and good lighting (avoid underexposure).
When using AI Enhance in Wave Player, a preview of AI Enhance will activate after the video is paused for approximately 1 second. During normal playback and “scrubbing” on the timeline, the AI Enhance is temporarily disabled to allow faster preview at the tradeoff of reduced preview screen quality, since AI Enhance consumes significant GPU resources. This reduced preview effect is only for the user interface in Wave Player: images and movies exported from the Wave Player will have full image quality of AI Enhance, if AI Enhance is turned on before the Export.
When using AI Enhance in Wave Player, AI Enhance will be active at 'Full' preview quality, using lower preview quality will disable the feature during scrubbing, but it will be enabled during export and in the export queue.