0
0
mirror of https://github.com/obsproject/obs-studio.git synced 2024-09-20 13:08:50 +02:00
obs-studio/libobs/data/bicubic_scale.effect
James Park ba21fb947e libobs: Fix various alpha issues
There are cases where alpha is multiplied unnecessarily. This change
attempts to use premultiplied alpha blending for composition.

To keep this change simple, The filter chain will continue to use
straight alpha. Otherwise, every source would need to modified to output
premultiplied, and every filter modified for premultiplied input.

"DrawAlphaDivide" shader techniques have been added to convert from
premultiplied alpha to straight alpha for final output. "DrawMatrix"
techniques ignore alpha, so they do not appear to need changing.

One remaining issue is that scale effects are set up here to use the
same shader logic for both scale filters (straight alpha - incorrectly),
and output composition (premultiplied alpha - correctly). A fix could be
made to add additional shaders for straight alpha, but the "real" fix
may be to eliminate the straight alpha path at some point.

For graphics, SrcBlendAlpha and DestBlendAlpha were both ONE, and could
combine together to form alpha values greater than one. This is not as
noticeable of a problem for UNORM targets because the channels are
clamped, but it will likely become a problem in more situations if FLOAT
targets are used.

This change switches DestBlendAlpha to INVSRCALPHA. The blending
behavior of stacked transparents is preserved without overflowing the
alpha channel.

obs-transitions: Use premultiplied alpha blend, and simplify shaders
because both inputs and outputs use premultiplied alpha now.

Fixes https://obsproject.com/mantis/view.php?id=1108
2019-05-08 20:26:52 -07:00

183 lines
4.0 KiB
Plaintext

/*
* bicubic sharper (better for downscaling)
* note - this shader is adapted from the GPL bsnes shader, very good stuff
* there.
*/
uniform float4x4 ViewProj;
uniform texture2d image;
uniform float4x4 color_matrix;
uniform float2 base_dimension_i;
uniform float undistort_factor = 1.0;
sampler_state textureSampler {
Filter = Linear;
AddressU = Clamp;
AddressV = Clamp;
};
struct VertData {
float4 pos : POSITION;
float2 uv : TEXCOORD0;
};
VertData VSDefault(VertData v_in)
{
VertData vert_out;
vert_out.pos = mul(float4(v_in.pos.xyz, 1.0), ViewProj);
vert_out.uv = v_in.uv;
return vert_out;
}
float weight(float x)
{
float ax = abs(x);
/* Sharper version. May look better in some cases. */
const float B = 0.0;
const float C = 0.75;
if (ax < 1.0)
return (pow(x, 2.0) *
((12.0 - 9.0 * B - 6.0 * C) * ax +
(-18.0 + 12.0 * B + 6.0 * C)) +
(6.0 - 2.0 * B))
/ 6.0;
else if ((ax >= 1.0) && (ax < 2.0))
return (pow(x, 2.0) *
((-B - 6.0 * C) * ax + (6.0 * B + 30.0 * C)) +
(-12.0 * B - 48.0 * C) * ax +
(8.0 * B + 24.0 * C))
/ 6.0;
else
return 0.0;
}
float4 weight4(float x)
{
return float4(
weight(x - 2.0),
weight(x - 1.0),
weight(x),
weight(x + 1.0));
}
float AspectUndistortX(float x, float a)
{
// The higher the power, the longer the linear part will be.
return (1.0 - a) * (x * x * x * x * x) + a * x;
}
float AspectUndistortU(float u)
{
// Normalize texture coord to -1.0 to 1.0 range, and back.
return AspectUndistortX((u - 0.5) * 2.0, undistort_factor) * 0.5 + 0.5;
}
float2 pixel_coord(float xpos, float ypos)
{
return float2(AspectUndistortU(xpos), ypos);
}
float4 pixel(float xpos, float ypos, bool undistort)
{
if (undistort)
return image.Sample(textureSampler, pixel_coord(xpos, ypos));
else
return image.Sample(textureSampler, float2(xpos, ypos));
}
float4 get_line(float ypos, float4 xpos, float4 linetaps, bool undistort)
{
return
pixel(xpos.r, ypos, undistort) * linetaps.r +
pixel(xpos.g, ypos, undistort) * linetaps.g +
pixel(xpos.b, ypos, undistort) * linetaps.b +
pixel(xpos.a, ypos, undistort) * linetaps.a;
}
float4 DrawBicubic(VertData v_in, bool undistort)
{
float2 stepxy = base_dimension_i;
float2 pos = v_in.uv + stepxy * 0.5;
float2 f = frac(pos / stepxy);
float4 rowtaps = weight4(1.0 - f.x);
float4 coltaps = weight4(1.0 - f.y);
/* make sure all taps added together is exactly 1.0, otherwise some
* (very small) distortion can occur */
rowtaps /= rowtaps.r + rowtaps.g + rowtaps.b + rowtaps.a;
coltaps /= coltaps.r + coltaps.g + coltaps.b + coltaps.a;
float2 xystart = (-1.5 - f) * stepxy + pos;
float4 xpos = float4(
xystart.x,
xystart.x + stepxy.x,
xystart.x + stepxy.x * 2.0,
xystart.x + stepxy.x * 3.0
);
return
get_line(xystart.y , xpos, rowtaps, undistort) * coltaps.r +
get_line(xystart.y + stepxy.y , xpos, rowtaps, undistort) * coltaps.g +
get_line(xystart.y + stepxy.y * 2.0, xpos, rowtaps, undistort) * coltaps.b +
get_line(xystart.y + stepxy.y * 3.0, xpos, rowtaps, undistort) * coltaps.a;
}
float4 PSDrawBicubicRGBA(VertData v_in, bool undistort) : TARGET
{
return DrawBicubic(v_in, undistort);
}
float4 PSDrawBicubicRGBADivide(VertData v_in) : TARGET
{
float4 rgba = DrawBicubic(v_in, false);
float alpha = rgba.a;
float multiplier = (alpha > 0.0) ? (1.0 / alpha) : 0.0;
return float4(rgba.rgb * multiplier, alpha);
}
float4 PSDrawBicubicMatrix(VertData v_in) : TARGET
{
float3 rgb = DrawBicubic(v_in, false).rgb;
float3 yuv = mul(float4(saturate(rgb), 1.0), color_matrix).xyz;
return float4(yuv, 1.0);
}
technique Draw
{
pass
{
vertex_shader = VSDefault(v_in);
pixel_shader = PSDrawBicubicRGBA(v_in, false);
}
}
technique DrawAlphaDivide
{
pass
{
vertex_shader = VSDefault(v_in);
pixel_shader = PSDrawBicubicRGBADivide(v_in);
}
}
technique DrawUndistort
{
pass
{
vertex_shader = VSDefault(v_in);
pixel_shader = PSDrawBicubicRGBA(v_in, true);
}
}
technique DrawMatrix
{
pass
{
vertex_shader = VSDefault(v_in);
pixel_shader = PSDrawBicubicMatrix(v_in);
}
}