This collection contain my RG0 normal map upscaling models.

All models here are for upscaling *tangent-space* normal maps in RG0 format. RG0 means that the B channel is set to 0. These models will work not correctly if you give them images with non-zero B channel, so you either have to zero the B channel manually or use tool like chaiNNer to do it.

## DDS Compression

I made 3 versions: 
- Normal RG0 is for uncompressed normal map textures. Since it hasn't been trained on compression artifacts, it's highly sensitive to quantization artifacts and noise.
- Normal RG0 BC1 is for BC1-compressed DDS normal map textures.
- Normal RG0 BC7 is for BC7-compressed DDS normal map textures. This model sometimes produces images that aren't as sharp. In those cases, you can try the BC1 version to see whether it gives better results.

Normal Map Upscaling

A 4x upscaler for uncompressed normal maps with a zeroed-out B channel. The input is required to have no alpha and constant-zero blue channel. The model can handle some light compression artifacts but has trouble with quantization artifacts. The output normals will also have a constant-zero B channel. Use external software or image editing plugins to properly normalize the generated normals and generate the Z component (if necessary). E.g. chaiNNer can do this with Normalize Normal Map node. Do not rely on this network producing unit vectors.


More information: https://github.com/RunDevelopment/ESRGAN-models/blob/main/normals/README.md

Normal RG0

A 4x upscaler for BC1-compressed normal maps with a zeroed-out B channel. The output normals will also have a constant-zero B channel. Use external software or image editing plugins to properly normalize the generated normals and generate the Z component (if necessary). E.g. chaiNNer can do this with Normalize Normal Map node. Do not rely on this network producing unit vectors. 

**Dataset transformation:** \
The LR have been compressed with various BC1 compression settings (dithering, weighting) using Texconv version 2021.11.8.1. Since the contents of the B channel influence the R and G channels during compression, the LR were given a B channel that was either the Z component of the normal, a random constant color, or a random texture (usually the G channel of the associated albedo) before compression. The LR + B channel was then compressed and the resulting BC1-compressed DDS was converted back into a PNG and had its B channel zeroed out.

More information: https://github.com/RunDevelopment/ESRGAN-models/blob/main/normals/README.md

Normal RG0 BC1

A 4x upscaler for BC7-compressed normal maps with a zeroed-out B channel and no alpha. The input is required to have no alpha and constant-zero blue channel. BC7-compressed images with an alpha channel should be handled by the BC1 model instead. This is because BC7 achieves very different quality for images with and without an alpha channel. The output normals will also have a constant-zero B channel. The output normals will also have a constant-zero B channel. Use external software or image editing plugins to properly normalize the generated normals and generate the Z component (if necessary). E.g. chaiNNer can do this with Normalize Normal Map node. Do not rely on this network producing unit vectors. 

**Dataset transformation:** \
The LR have been compressed using Texconv version 2021.11.8.1. Since the contents of the B channel influence the R and G channels during compression, the LR were given a B channel that was either the Z component of the normal, a random constant color, or a random texture (usually the G channel of the associated albedo) before compression. The LR + B channel was then compressed and the resulting BC7-compressed DDS was converted back into a PNG and had its B channel zeroed out.

More information: https://github.com/RunDevelopment/ESRGAN-models/blob/main/normals/README.md

Normal RG0 BC7

A 4x upscaler for BC1-compressed wood textures. The model was trained on wood planks, wood stems, and a bit of tree bark. It performs reasonably well on woody textures and produces sharp outputs. However, it also tends to slightly shift the hue of the image, so results might appear slightly more red.

More information: https://github.com/RunDevelopment/ESRGAN-models/blob/main/wood/README.md

RunDevelopment's Models

RunDevelopment's Collections