The Resolution Enhancement tab is used to control the Multi-Frame Super-Resolution processing. It provides an effective means for increasing the native video resolution by utilizing information redundancy between successive frames in a video.
Applications and benefits include:
- noise reduction
- suppression of atmospheric distortions (mirage-like wobbling)
- reduction of compression blocking artifacts
|Resolution Filter - Compact||Resolution Filter - Expanded|
|This slider controls the number of successive frames used by the algorithm. The maximal number of frames can be set by going to Tools > Options > Processing > Cache Size|
This combo box is used to set the upscale factor applied to the input frame by the Multi-Frame Super-Resolution algorithm. Available options depend on the input resolution and can be 1x, 2x, 3x, or 4x.
For example, applying a Super-Resolution factor of 2x to an input video with a resolution of 320x240 results in a video with a resolution of 640x480.
|This slider controls the suppression of outlier-related processing artifacts. The first phase of the Super-Resolution algorithm is motion estimation between successive frames. However, this can generate ghosting artifacts in some cases. The Fusion parameter can reduce ghosting|
|This slider can be used to control the tradeoff between sharpness and noise suppression. Settings range from 0 (maximum noise suppression) to 1 (maximum sharpness)|
|ROI Mode||This button allows the user to enhance, preview, and save the ROI with a wider range of Super-Resolution (upscale) factors.|
NOTE: The maximum output video size is 1080p. Hence higher resolution video has a lower available Super-Resolution factor. However, a larger Super-Resolution factor can be set by limiting the video with ROI mode or cropping the video.
- To enhance the precision of the motion estimation, select an ROI which doesn’t contain outliers.
- Choosing the appropriate Frames value can be challenging. The optimal value depends on the desired performance, Super-Resolution upscale factor, and scene dynamics. For static scenes, the best results are achieved by setting the Frames value to 9 or higher and the Super Resolution set to 2x.
- With dynamic scenes, it’s best to start at the lowest setting and work your way up. With static scenes, it’s best to start high and work your way down.
- Super Resolution yields the best results for static scenes when used in conjunction with Deblurring.