Use medium precision in blur shaders

To avoid artifacts on low-profile systems

client/shaders/blur_h/opengl_fragment.glsl

@@ -2,7 +2,7 @@
 
 uniform sampler2D rendered;
 uniform vec2 texelSize0;
-uniform lowp float bloomRadius = 3.0;
+uniform mediump float bloomRadius = 3.0;
 
 #ifdef GL_ES
 varying mediump vec2 varTexCoord;
@@ -13,13 +13,13 @@ centroid varying vec2 varTexCoord;
 void main(void)
 {
  // kernel distance and linear size
- lowp float n = 2. * bloomRadius + 1.;
+ mediump float n = 2. * bloomRadius + 1.;
 
  vec2 uv = varTexCoord.st - vec2(bloomRadius * texelSize0.x, 0.);
  vec4 color = vec4(0.);
- lowp float sum = 0.;
- for (lowp float i = 0.; i < n; i++) {
- lowp float weight = pow(1. - (abs(i / bloomRadius - 1.)), 1.3);
+ mediump float sum = 0.;
+ for (mediump float i = 0.; i < n; i++) {
+ mediump float weight = pow(1. - (abs(i / bloomRadius - 1.)), 1.3);
  color += texture2D(rendered, uv).rgba * weight;
  sum += weight;
  uv += vec2(texelSize0.x, 0.);

client/shaders/blur_v/opengl_fragment.glsl

@@ -2,7 +2,7 @@
 
 uniform sampler2D rendered;
 uniform vec2 texelSize0;
-uniform lowp float bloomRadius = 3.0;
+uniform mediump float bloomRadius = 3.0;
 
 #ifdef GL_ES
 varying mediump vec2 varTexCoord;
@@ -13,13 +13,13 @@ centroid varying vec2 varTexCoord;
 void main(void)
 {
  // kernel distance and linear size
- lowp float n = 2. * bloomRadius + 1.;
+ mediump float n = 2. * bloomRadius + 1.;
 
  vec2 uv = varTexCoord.st - vec2(0., bloomRadius * texelSize0.y);
  vec4 color = vec4(0.);
- lowp float sum = 0.;
- for (lowp float i = 0.; i < n; i++) {
- lowp float weight = pow(1. - (abs(i / bloomRadius - 1.)), 1.3);
+ mediump float sum = 0.;
+ for (mediump float i = 0.; i < n; i++) {
+ mediump float weight = pow(1. - (abs(i / bloomRadius - 1.)), 1.3);
  color += texture2D(rendered, uv).rgba * weight;
  sum += weight;
  uv += vec2(0., texelSize0.y);