Ios 具有不透明度的GPUImageLightedBlendFilter

使用

GPUImage

我试图用不透明度复制Photoshop Lightning混合模式。不幸的是，使用

gpuimageBlendedFilter

时，alpha通道没有任何效果

Brad确认alpha可能存在以下问题：

我已经成功地使用符合alpha值的

CoreImage

复制了PS混合

CIImage *ciImage1 = [[CIImage alloc] initWithImage:input1];
CIImage *ciImage2 = [[CIImage alloc] initWithImage:input2];

// Alpha adjustment for input1
CIFilter *alphaFilter = [CIFilter filterWithName:@"CIColorMatrix"];
CGFloat rgba[4] = { 0.0, 0.0, 0.0, 0.5 };
CIVector *alphaVector = [CIVector vectorWithValues:rgba count:4];
[alphaFilter setValue:alphaVector forKey:@"inputAVector"];
[alphaFilter setValue:ciImage1 forKey:kCIInputImageKey];

// Lighten blend
CIFilter *blendFilter = [CIFilter filterWithName:@"CILightenBlendMode"];
[blendFilter setValue:alphaFilter.outputImage forKey:kCIInputImageKey];
[blendFilter setValue:ciImage2 forKey:kCIInputBackgroundImageKey];

我尝试过两个版本的

GPUImage

（它们使用不同的方法调整

input1

的alpha）

或：

两个

GPUImage

实现都返回输出，就好像

input1

的alpha是

1.0

我看过互联网上不同来源的混合模式文档，它们都使用以下公式：

max(blend, base)

查看

gpuimageLightedBlendFilter

实现中的着色器，它也使用相同的公式：

lowp vec4 textureColor = texture2D(inputImageTexture, textureCoordinate);
lowp vec4 textureColor2 = texture2D(inputImageTexture2, textureCoordinate2);
gl_FragColor = max(textureColor, textureColor2);

然而，Photoshop和CoreImage似乎对alpha有一些额外的操作（可能类似于Gimp:）

---

任何人都知道如何将alpha通道包括在

gpuimageLightedBlendFilter

公式中？

下面是

gpuimageLightedBlendFilter

的着色器代码，具有不透明度和数量

precision highp float;
uniform sampler2D inputImageTexture;
uniform sampler2D inputImageTexture2;
uniform float alpha;   // used for opacity....
uniform float amount;  // amount of blend....
varying vec2 textureCoordinate;

void main ()
{
    // Get samples from both layers
    vec4 dst = texture2D(inputImageTexture, textureCoordinate);
    vec4 src = texture2D(inputImageTexture2, textureCoordinate);

src.a *= alpha;
vec4 colour = vec4(0.0, 0.0, 0.0, 0.0);

colour = vec4(max(dst, src).rgb, src.a) * src.a + dst * (1.0 - src.a);
colour = clamp(colour, 0.0, 1.0);
gl_FragColor.xyz = mix(dst, colour, amount).rgb;
gl_FragColor.w = 1.0;
}

---

非常适合我……

我发现此着色器代码最适合我的情况：

lowp vec4 textureColor = texture2D(inputImageTexture, textureCoordinate);
lowp vec4 textureColor2 = texture2D(inputImageTexture2, textureCoordinate2);

textureColor.rgb *= textureColor.a;
textureColor2.rgb *= textureColor2.a;
lowp vec4 textureOut = max(textureColor, textureColor2);
textureOut.rgb /= textureOut.a;

gl_FragColor = textureOut;

可在此处找到：

precision highp float;
uniform sampler2D inputImageTexture;
uniform sampler2D inputImageTexture2;
uniform float alpha;   // used for opacity....
uniform float amount;  // amount of blend....
varying vec2 textureCoordinate;

void main ()
{
    // Get samples from both layers
    vec4 dst = texture2D(inputImageTexture, textureCoordinate);
    vec4 src = texture2D(inputImageTexture2, textureCoordinate);

src.a *= alpha;
vec4 colour = vec4(0.0, 0.0, 0.0, 0.0);

colour = vec4(max(dst, src).rgb, src.a) * src.a + dst * (1.0 - src.a);
colour = clamp(colour, 0.0, 1.0);
gl_FragColor.xyz = mix(dst, colour, amount).rgb;
gl_FragColor.w = 1.0;
}

lowp vec4 textureColor = texture2D(inputImageTexture, textureCoordinate);
lowp vec4 textureColor2 = texture2D(inputImageTexture2, textureCoordinate2);

textureColor.rgb *= textureColor.a;
textureColor2.rgb *= textureColor2.a;
lowp vec4 textureOut = max(textureColor, textureColor2);
textureOut.rgb /= textureOut.a;

gl_FragColor = textureOut;