Skip to content

@igv igv released this Nov 9, 2017 · 28 commits to master since this release

Updated 22.4.2019: Improved 8-0-4-1 and 16-0-4-1.

FSRCNN extended with local residual learning and recursive blocks with pre-activation structure.
All networks were trained with --distort=True (means they can reduce compression artifacts)
Older mpv versions, prior to Jul 8 2018, require --fbo-format=rgba16hf (rgba16f for opengl)
For example if you placed this shader in the same folder as your mpv.conf, then your config should look like this (regardless of the operating system):

profile=gpu-hq
glsl-shader="~~/FSRCNNX_x2_8-0-4-1.glsl"

For upscaling images use 56-16-4-1, it's too slow for real-time without MRT support in mpv.

Assets 6

@igv igv released this Oct 4, 2017 · 28 commits to master since this release

Shader for mpv.
Trained with --distort=True (means it can remove / reduce compression artifacts)
Requires --fbo-format=rgba16f (rgba16hf for vulkan and d3d11)
For example if you placed this shader in the same folder as your mpv.conf, then your config should look like this:

profile=gpu-hq
fbo-format=rgba16f
glsl-shaders="~~/FSRCNN_x2_r2_32-0-2.glsl"
Assets 4

@igv igv released this Sep 25, 2017 · 28 commits to master since this release

Shader for mpv.
Trained with --distort=True (means it can remove / reduce compression artifacts)
Requires --fbo-format=rgba16f (rgba16hf for vulkan and d3d11)
For example if you placed this shader in the same folder as your mpv.conf, then your config should look like this:

profile=gpu-hq
fbo-format=rgba16f
glsl-shaders="~~/FSRCNN_x2_r2_16-0-2.glsl"
Assets 5

@igv igv released this Sep 2, 2017 · 28 commits to master since this release

Shader for mpv.
Trained with --distort=True (means it can remove / reduce compression artifacts)
Requires --fbo-format=rgba16f (rgba16hf for vulkan and d3d11)
For example if you placed this shader in the same folder as your mpv.conf, then your config should look like this:

profile=gpu-hq
fbo-format=rgba16f
glsl-shaders="~~/FSRCNN_8-0-2_x2.glsl"
Assets 5
Sep 1, 2017

0.1

Adjust deconvolution filter size for different scaling factors
You can’t perform that action at this time.