Blur

自从iOS系统引入了Blur效果，也就是所谓的毛玻璃、模糊化效果，磨砂效果，各大系统就开始竞相模仿，这是一个怎样的效果呢，我们现来看一些图：

这些就是典型的Blur效果，在iOS和MIUI中还有很多，这里就不再贴图了。

实现

实现Blur效果主要有两种方式，一个是通过RenderScript来做，另一种是通过直接对像素点来进行算法处理。

RenderScript是API11之后才引入的，所以对版本有限制，而且RenderScript确实挺复杂的，虽然使用他的Blur功能很简单，但是要真正搞懂，不是一天两天的事。

本文主要介绍另一种算法来实现Blur，这个算法是目前市面上对Blur效果处理比较好的一种算法了，研究的前沿网址请戳 我是Blur 。

经典图：

有兴趣的朋友可以去看看。

使用

下面我们来看看如何在Android中来使用Blur，当然，我们需要使用上面提到的FastBlur

package com.xys.blur;import android.graphics.Bitmap;/** * Created by paveld on 3/6/14. */public class FastBlur {    public static Bitmap doBlur(Bitmap sentBitmap, int radius, boolean canReuseInBitmap) {        // Stack Blur v1.0 from        // http://www.quasimondo.com/StackBlurForCanvas/StackBlurDemo.html        //        // Java Author: Mario Klingemann <mario at quasimondo.com>        // http://incubator.quasimondo.com        // created Feburary 29, 2004        // Android port : Yahel Bouaziz <yahel at kayenko.com>        // http://www.kayenko.com        // ported april 5th, 2012        // This is a compromise between Gaussian Blur and Box blur        // It creates much better looking blurs than Box Blur, but is        // 7x faster than my Gaussian Blur implementation.        //        // I called it Stack Blur because this describes best how this        // filter works internally: it creates a kind of moving stack        // of colors whilst scanning through the image. Thereby it        // just has to add one new block of color to the right side        // of the stack and remove the leftmost color. The remaining        // colors on the topmost layer of the stack are either added on        // or reduced by one, depending on if they are on the right or        // on the left side of the stack.        //        // If you are using this algorithm in your code please add        // the following line:        //        // Stack Blur Algorithm by Mario Klingemann <[email protected]>        Bitmap bitmap;        if (canReuseInBitmap) {            bitmap = sentBitmap;        } else {            bitmap = sentBitmap.copy(sentBitmap.getConfig(), true);        }        if (radius < 1) {            return (null);        }        int w = bitmap.getWidth();        int h = bitmap.getHeight();        int[] pix = new int[w * h];        bitmap.getPixels(pix, 0, w, 0, 0, w, h);        int wm = w - 1;        int hm = h - 1;        int wh = w * h;        int div = radius + radius + 1;        int r[] = new int[wh];        int g[] = new int[wh];        int b[] = new int[wh];        int rsum, gsum, bsum, x, y, i, p, yp, yi, yw;        int vmin[] = new int[Math.max(w, h)];        int divsum = (div + 1) >> 1;        divsum *= divsum;        int dv[] = new int[256 * divsum];        for (i = 0; i < 256 * divsum; i++) {            dv[i] = (i / divsum);        }        yw = yi = 0;        int[][] stack = new int[div][3];        int stackpointer;        int stackstart;        int[] sir;        int rbs;        int r1 = radius + 1;        int routsum, goutsum, boutsum;        int rinsum, ginsum, binsum;        for (y = 0; y < h; y++) {            rinsum = ginsum = binsum = routsum = goutsum = boutsum = rsum = gsum = bsum = 0;            for (i = -radius; i <= radius; i++) {                p = pix[yi + Math.min(wm, Math.max(i, 0))];                sir = stack[i + radius];                sir[0] = (p & 0xff0000) >> 16;                sir[1] = (p & 0x00ff00) >> 8;                sir[2] = (p & 0x0000ff);                rbs = r1 - Math.abs(i);                rsum += sir[0] * rbs;                gsum += sir[1] * rbs;                bsum += sir[2] * rbs;                if (i > 0) {                    rinsum += sir[0];                    ginsum += sir[1];                    binsum += sir[2];                } else {                    routsum += sir[0];                    goutsum += sir[1];                    boutsum += sir[2];                }            }            stackpointer = radius;            for (x = 0; x < w; x++) {                r[yi] = dv[rsum];                g[yi] = dv[gsum];                b[yi] = dv[bsum];                rsum -= routsum;                gsum -= goutsum;                bsum -= boutsum;                stackstart = stackpointer - radius + div;                sir = stack[stackstart % div];                routsum -= sir[0];                goutsum -= sir[1];                boutsum -= sir[2];                if (y == 0) {                    vmin[x] = Math.min(x + radius + 1, wm);                }                p = pix[yw + vmin[x]];                sir[0] = (p & 0xff0000) >> 16;                sir[1] = (p & 0x00ff00) >> 8;                sir[2] = (p & 0x0000ff);                rinsum += sir[0];                ginsum += sir[1];                binsum += sir[2];                rsum += rinsum;                gsum += ginsum;                bsum += binsum;                stackpointer = (stackpointer + 1) % div;                sir = stack[(stackpointer) % div];                routsum += sir[0];                goutsum += sir[1];                boutsum += sir[2];                rinsum -= sir[0];                ginsum -= sir[1];                binsum -= sir[2];                yi++;            }            yw += w;        }        for (x = 0; x < w; x++) {            rinsum = ginsum = binsum = routsum = goutsum = boutsum = rsum = gsum = bsum = 0;            yp = -radius * w;            for (i = -radius; i <= radius; i++) {                yi = Math.max(0, yp) + x;                sir = stack[i + radius];                sir[0] = r[yi];                sir[1] = g[yi];                sir[2] = b[yi];                rbs = r1 - Math.abs(i);                rsum += r[yi] * rbs;                gsum += g[yi] * rbs;                bsum += b[yi] * rbs;                if (i > 0) {                    rinsum += sir[0];                    ginsum += sir[1];                    binsum += sir[2];                } else {                    routsum += sir[0];                    goutsum += sir[1];                    boutsum += sir[2];                }                if (i < hm) {                    yp += w;                }            }            yi = x;            stackpointer = radius;            for (y = 0; y < h; y++) {                // Preserve alpha channel: ( 0xff000000 & pix[yi] )                pix[yi] = (0xff000000 & pix[yi]) | (dv[rsum] << 16) | (dv[gsum] << 8) | dv[bsum];                rsum -= routsum;                gsum -= goutsum;                bsum -= boutsum;                stackstart = stackpointer - radius + div;                sir = stack[stackstart % div];                routsum -= sir[0];                goutsum -= sir[1];                boutsum -= sir[2];                if (x == 0) {                    vmin[y] = Math.min(y + r1, hm) * w;                }                p = x + vmin[y];                sir[0] = r[p];                sir[1] = g[p];                sir[2] = b[p];                rinsum += sir[0];                ginsum += sir[1];                binsum += sir[2];                rsum += rinsum;                gsum += ginsum;                bsum += binsum;                stackpointer = (stackpointer + 1) % div;                sir = stack[stackpointer];                routsum += sir[0];                goutsum += sir[1];                boutsum += sir[2];                rinsum -= sir[0];                ginsum -= sir[1];                binsum -= sir[2];                yi += w;            }        }        bitmap.setPixels(pix, 0, w, 0, 0, w, h);        return (bitmap);    }}

算法就是这样了，不要问我懂不懂，你懂的。

使用

如何在程序中使用呢，也很简单：

package com.xys.blur;import android.app.Activity;import android.graphics.Bitmap;import android.graphics.BitmapFactory;import android.graphics.Canvas;import android.graphics.Paint;import android.graphics.drawable.BitmapDrawable;import android.os.Bundle;import android.view.View;import android.view.ViewTreeObserver.OnPreDrawListener;import android.widget.ImageView;public class Test extends Activity { @Override protected void onCreate(Bundle savedInstanceState) {  super.onCreate(savedInstanceState);  setContentView(R.layout.main);  final ImageView imageView = (ImageView) findViewById(R.id.image1);  final Bitmap bitmap = BitmapFactory.decodeResource(getResources(),    R.drawable.blur);  imageView.getViewTreeObserver().addOnPreDrawListener(    new OnPreDrawListener() {     @Override     public boolean onPreDraw() {      blur(bitmap, imageView);      return true;     }    }); } private void blur(Bitmap bkg, View view) {  long startMs = System.currentTimeMillis();  float scaleFactor = 8;  float radius = 2;  Bitmap overlay = Bitmap.createBitmap(    (int) (view.getMeasuredWidth() / scaleFactor),    (int) (view.getMeasuredHeight() / scaleFactor),    Bitmap.Config.ARGB_8888);  Canvas canvas = new Canvas(overlay);  canvas.translate(-view.getLeft() / scaleFactor, -view.getTop()    / scaleFactor);  canvas.scale(1 / scaleFactor, 1 / scaleFactor);  Paint paint = new Paint();  paint.setFlags(Paint.FILTER_BITMAP_FLAG);  canvas.drawBitmap(bkg, 0, 0, paint);  overlay = FastBlur.doBlur(overlay, (int) radius, true);  view.setBackground(new BitmapDrawable(getResources(), overlay));  System.out.println(System.currentTimeMillis() - startMs + "ms"); }}

为什么我们要在addOnPreDrawListener中来调用blur方法呢，看我前面的文章的朋友应该会知道，这样做是为了能够在onCreate中获取控件尺寸，通过scaleFactor和radius两个参数，我们来控制Blur的程度。

代码中还有几点需要解释下：

1、我们为什么要通过scaleFactor来缩小图片：由于在做Blur的时候，图片精度本来就要降低，那么我们为什么不先降低精度再去处理呢，这样的效果就是巨大的缩小了生成时间

2、我们给Paint提供了FILTER_BITMAP_FLAG标示，这样的话在处理bitmap缩放的时候，就可以达到双缓冲的效果，模糊处理的过程就更加顺畅了

3、如果我们要做某一部分的Blur效果，一般是将这部分图片裁减下来，然后Blur后设给某个控件的背景

最终效果如下：

效果已经出来了，Demo很简单，只是为了演示使用方法。

以上。