android - PNG 混合模式

Question

在 Photoshop 中有几种混合模式，例如:

• 颜色加深
• 相乘
• 变暗
• 减轻
• 叠加

(更多信息在这里:http://help.adobe.com/en_US/photoshop/cs/using/WSfd1234e1c4b69f30ea53e41001031ab64-77eba.html)

例如，对于模式Color Burn“查看每个 channel 中的颜色信息并通过增加两者之间的对比度使基色变暗以反射(reflect)混合色。与白色混合不会产生任何变化".

这是我的代码:

Bitmap blendBitmap = BitmapFactory.decodeStream(ctx.getAssets().open(filename));
Canvas canvas = new Canvas(srcBitmap);
canvas.drawBitmap(blendBitmap, 0, 0, null); // ?
p.recycle();
p = null;

是否可以应用 Color Burn 混合模式，而不是简单地将图像绘制在其他图像之上(就像在这个小代码中一样)？

Answer 1

这并不难。

我使用 NDK(因为性能)来处理像素。混合模式的信息非常有用:How does photoshop blend two images together?

我最终的解决方案是:

#define ChannelBlend_ColorBurn(A, B)  ((uint8_t) ((B == 0) ? B : max(0, (255 - ((255 -  
#define ChannelBlend_Alpha(A, B, O)    ((uint8_t) (O * A + (1 - O) * B))
#define ChannelBlend_AlphaF(A, B, F, O) (ChannelBlend_Alpha(F(A, B), A, O))

typedef struct {
    uint8_t red;
    uint8_t green;
    uint8_t blue;
    uint8_t alpha;
} rgba;

// Blend
JNIEXPORT void 
JNICALL Java_com_package_Filter_jniBlend(JNIEnv* env, jobject obj, jobject bitmapA,
    jobject bitmapB, jobject bitmapOut, jint mode) {

    // Properties
    AndroidBitmapInfo   infoA;
    void*               pixelsA;
    AndroidBitmapInfo   infoB;
    void*               pixelsB;
    AndroidBitmapInfo   infoOut;
    void*               pixelsOut;

    int ret;

    // Get image info
    if ((ret = AndroidBitmap_getInfo(env, bitmapA, &infoA)) < 0 ||
        (ret = AndroidBitmap_getInfo(env, bitmapB, &infoB)) < 0 ||
        (ret = AndroidBitmap_getInfo(env, bitmapOut, &infoOut)) < 0) {
        return;
    }
    // Check image
    if (infoA.format != ANDROID_BITMAP_FORMAT_RGBA_8888 ||
        infoB.format != ANDROID_BITMAP_FORMAT_RGBA_8888 ||
        infoOut.format != ANDROID_BITMAP_FORMAT_RGBA_8888) {
        return;
    }
    // Lock all images
    if ((ret = AndroidBitmap_lockPixels(env, bitmapA, &pixelsA)) < 0 ||
        (ret = AndroidBitmap_lockPixels(env, bitmapB, &pixelsB)) < 0 ||
        (ret = AndroidBitmap_lockPixels(env, bitmapOut, &pixelsOut)) < 0) {
        LOGE("Error! %d", ret);
    }

    int h = infoA.height;
    int w = infoA.width;
    int wh = w * h;

    int n;
    rgba* inputA = (rgba*) pixelsA;
    rgba* inputB = (rgba*) pixelsB;
    rgba* output = (rgba*) pixelsOut;

    rgba pA, pB;

    int x, y;
    for (y = 0; y < h; y++) {
        for (x = 0; x < w; x++) {
            n = y * w + x;

            pA = inputA[n];
            pB = inputB[n];

            float alpha = (float) pB.alpha / 255.0;

            output[n].red = ChannelBlend_AlphaF(pA.red, pB.red, ChannelBlend_ColorBurn, alpha);
            output[n].green = ChannelBlend_AlphaF(pA.green, pB.green, ChannelBlend_ColorBurn, alpha);
            output[n].blue = ChannelBlend_AlphaF(pA.blue, pB.blue, ChannelBlend_ColorBurn, alpha);
        }
    }

    // Unlocks everything
    AndroidBitmap_unlockPixels(env, bitmapA);
    AndroidBitmap_unlockPixels(env, bitmapB);
    AndroidBitmap_unlockPixels(env, bitmapOut);
}

提高性能的小评论:在我对单个像素进行方程式之后，最好将值存储在某种缓存中，以便下次无需方程式即可更快地访问它。