Super Resolution#
The Super Resolution filter, while preserving the content, improves the resolution of low-resolution videos, enhances the details, and sharpens the output.
This filter enhances the resolution of low-resolution videos, enhances the details, sharpens the output while preserving the content. Mode 0 is suitable for upscaling lossy content that has encoding artifacts, and Mode 1 is suitable for lossless videos:
Mode 0 enhances less and removes more encoding artifacts and is suited for lower-quality videos.
Mode 1 enhances more and is suited for higher quality lossless videos.
The filter’s input/output is as follows:
The input should be provided in a GPU buffer in BGR planar format, where each pixel component is a 32-bit float value.
The output of the filter is a GPU buffer in BGR planar format, where each pixel component is a 32-bit float value.
Table 1‑1 illustrates the scale and resolution support for input videos to be used with the ArtifactReduction and SuperRes effects.
Table 1‑1. Scale and Resolution Support for Input Videos
Effect |
Scale factor |
Input resolution range |
Output resolution range |
|---|---|---|---|
Artifact reduction |
Input and output range is the same as AR does not change the resolution |
[90p, 1080p] |
[90p, 1080p] |
Super resolution |
4/3x |
[90x90, 2160p] |
[120x120, 2880p] |
1.5x |
[90x90, 2160p] |
[135x135, 3240p] |
|
2x |
[90x90, 2160p] |
[180x180, 4320p] |
|
3x |
[90x90, 720p] |
[270x270, 2160p] |
|
4x |
[90x90, 540p] |
[360x360, 2160p] |
Note
A lower end GPU or some MIG configurations might not have enough memory to support the maximum permitted input resolutions of the Super Resolution feature. Refer to Using Multi-Instance GPUs (Linux only) for more information.