Closed id-ekaagr closed 6 months ago
I get this when I run clang-format
:
// from https://github.com/hyprwm/Hyprland/issues/1140#issuecomment-1335128437
precision mediump float;
varying vec2 v_texcoord;
uniform sampler2D tex;
const float temperature = 2600.0;
const float temperatureStrength = 1.0;
#define WithQuickAndDirtyLuminancePreservation
const float LuminancePreservationFactor = 1.0;
// function from https://www.shadertoy.com/view/4sc3D7
// valid from 1000 to 40000 K (and additionally 0 for pure full white)
vec3 colorTemperatureToRGB(const in float temperature) {
// values from:
// http://blenderartists.org/forum/showthread.php?270332-OSL-Goodness&p=2268693&viewfull=1#post2268693
mat3 m = (temperature <= 6500.0)
? mat3(vec3(0.0, -2902.1955373783176, -8257.7997278925690),
vec3(0.0, 1669.5803561666639, 2575.2827530017594),
vec3(1.0, 1.3302673723350029, 1.8993753891711275))
: mat3(vec3(1745.0425298314172, 1216.6168361476490,
-8257.7997278925690),
vec3(-2666.3474220535695, -2173.1012343082230,
2575.2827530017594),
vec3(0.55995389139931482, 0.70381203140554553,
1.8993753891711275));
return mix(
clamp(vec3(m[0] / (vec3(clamp(temperature, 1000.0, 40000.0)) + m[1]) +
m[2]),
vec3(0.0), vec3(1.0)),
vec3(1.0), smoothstep(1000.0, 0.0, temperature));
}
void main() {
vec4 pixColor = texture2D(tex, v_texcoord);
// RGB
vec3 color = vec3(pixColor[0], pixColor[1], pixColor[2]);
#ifdef WithQuickAndDirtyLuminancePreservation
color *= mix(1.0,
dot(color, vec3(0.2126, 0.7152, 0.0722)) /
max(dot(color, vec3(0.2126, 0.7152, 0.0722)), 1e-5),
LuminancePreservationFactor);
#endif
color = mix(color, color * colorTemperatureToRGB(temperature),
temperatureStrength);
vec4 outCol = vec4(color, pixColor[3]);
gl_FragColor = outCol;
}
What is the source/rationale behind your formatting?
Manually formatted the code for making it more readable.
This issue can safely be closed.
PS: Let me know what you think of this shader config. :)
// link: https://github.com/hyprwm/Hyprland/issues/1140#issuecomment-1614863627
precision mediump float;
varying vec2 v_texcoord;
uniform sampler2D tex;
const float X = 2.0;
const float R = X * X;
const float G = R * R;
const float B = G * G;
const float VIB_VIBRANCE = 1.0 / (R + B + G);
const vec3 VIB_RGB_BALANCE = vec3(
R,
G,
B
);
const vec3 VIB_coeffVibrance = VIB_RGB_BALANCE * -VIB_VIBRANCE;
void main() {
vec4 pixColor = texture2D(tex, v_texcoord);
vec3 color = vec3 (
pixColor[0],
pixColor[1],
pixColor[2]
);
vec3 VIB_coefLuma = vec3(
R / (R + B + G),
G / (R + B + G),
B / (R + B + G)
) * X;
float luma = dot(VIB_coefLuma, color);
float max_color = max(color[0], max(color[1], color[2]));
float min_color = min(color[0], min(color[1], color[2]));
float color_saturation = max_color - min_color;
vec3 p_col = vec3(vec3(vec3(vec3(sign(VIB_coeffVibrance) * color_saturation) - 1.0) * VIB_coeffVibrance) + 1.0);
pixColor[0] = mix(luma, color[0], p_col[0]);
pixColor[1] = mix(luma, color[1], p_col[1]);
pixColor[2] = mix(luma, color[2], p_col[2]);
gl_FragColor = pixColor;
}
It looked good for saving people's eyes from strain (keeping vibrancy), so I felt the need to share. Thanks for this amazing utility.
Is that supposed to be vibrance
or blue-light-filter
?
This is what it looks like on my system:
vibrance
:
blue-light-filter
:
We can better call it a highlight-filter
.
It's desirable by people who use light-colour-themes
in their IDE's like VSCodium
.
What are the drawbacks with blue-light-filter
:
yellow
colour for highlighting instance of clicked variable
or selected text
, stuff in yellow colour looks pretty dull.What this filter does:
Thanks loqusion.
Just to clarify, does highlight-filter
serve a similar purpose to blue-light-filter
? Or are there situations where a user might want to switch between highlight-filter
and blue-light-filter
(e.g. based on what's displayed on the screen)? I'm wondering if this should be implemented as a blue-light-filter
variant that is customizable in hyprshade.toml
, or if it should be added as a separate highlight-filter.glsl
.
blue-light-filter
is very scientific, highlight-filter
is a filter for particular use case, it's also very aggressive on blue
colour. So highlight-filter
should be kept in a separate category initially. Once highlight-filter
is tested enough it should become a variant of blue-light-filter
, someday.
Here's the latest code for the filter:
// link: https://github.com/hyprwm/Hyprland/issues/1140#issuecomment-1614863627
precision mediump float;
varying vec2 v_texcoord;
uniform sampler2D tex;
const float X = 1.0;
const float N = 1.0 + X;
const float R = N * N;
const float G = R * R;
const float B = G * G;
const float S = R + G + B;
const float VIB_VIBRANCE = 1.0 / S;
const vec3 VIB_RGB_BALANCE = vec3(
R,
G,
B
);
const vec3 VIB_coeffVibrance = VIB_RGB_BALANCE * -VIB_VIBRANCE;
void main() {
vec4 pixColor = texture2D(tex, v_texcoord);
vec3 color = vec3(pixColor[0], pixColor[1], pixColor[2]);
vec3 VIB_coefLuma = vec3(R / S, G / S, B / S) * R;
float luma = dot(VIB_coefLuma, color);
float max_color = max(color[0], max(color[1], color[2]));
float min_color = min(color[0], min(color[1], color[2]));
float color_saturation = max_color - min_color;
vec3 p_col = vec3(vec3(vec3(vec3(sign(VIB_coeffVibrance) * color_saturation) - 1.0) * VIB_coeffVibrance) + 1.0);
pixColor[0] = mix(luma, color[0], p_col[0]);
pixColor[1] = mix(luma, color[1], p_col[1]);
pixColor[2] = mix(luma, color[2], p_col[2]);
gl_FragColor = pixColor;
}
variable X
stands for Intensity. Ideally X
should be anything between 0.0
to 1.0
.
Please replace the code below with blue-light-shader.glsl. This one is properly formatted.