要解决不希望的PorterDuff效果
，首先使用最简单的方法，就像OP的问题一样，aPath.arcTo(*, *, *, *, false)就足够了——注意它是arcTo，不是addArc，以及添加arc之前的false手段no forceMoveTo——不需要PorterDuff。
Path arcPath = new Path();
@Override
protected void onDraw(Canvas canvas) {
arcPath.rewind();
arcPath.moveTo(arcRectF.centerX, arcRectF.centerY);
arcPath.arcTo(arcRectF, -90, currentAngleSweep, false);
arcPath.close();
canvas.clipPath(arcPath, Region.Op.DIFFERENCE);
canvas.drawBitmap(bitmap, circleSourceRect, circleDestRect, arcPaint);
}
如果您确实需要 PorterDuff，主要用于复杂的颜色变形，例如混合渐变，请不要将具有 PorterDuff 过滤效果的颜色或形状或位图直接绘制到提供的默认画布上onDraw(Canvas)，请使用一些带有 alpha 通道的缓冲/目标位图 [s]- - 并且 - 存储PorterDuffsetHasAlpha(true)过滤的结果，最后将位图绘制到默认画布上，除了矩阵更改之外不应用任何过滤。
这是一个创建边界模糊圆形图像的工作示例：
import android.annotation.SuppressLint;
import android.content.Context;
import android.graphics.Bitmap;
import android.graphics.Canvas;
import android.graphics.Color;
import android.graphics.Matrix;
import android.graphics.Paint;
import android.graphics.Path;
import android.graphics.PorterDuff;
import android.graphics.PorterDuffXfermode;
import android.graphics.RadialGradient;
import android.graphics.Rect;
import android.graphics.RectF;
import android.graphics.Shader;
import android.support.annotation.Nullable;
import android.util.AttributeSet;
import android.widget.ImageView;
/**
* Created by zdave on 6/22/17.
*/
public class BlurredCircleImageViewShader extends ImageView {
private Canvas mCanvas;
private Paint mPaint;
private Matrix matrix;
private static final float GRADIENT_RADIUS = 600f; //any value you like, but should be big enough for better resolution.
private Shader gradientShader;
private Bitmap bitmapGradient;
private Bitmap bitmapDest;
private Canvas canvasDest;
public BlurredCircleImageViewShader(Context context) {
this(context, null);
}
public BlurredCircleImageViewShader(Context context, @Nullable AttributeSet attrs) {
this(context, attrs, 0);
}
public BlurredCircleImageViewShader(Context context, @Nullable AttributeSet attrs, int defStyleAttr) {
super(context, attrs, defStyleAttr);
mPaint = new Paint(Paint.ANTI_ALIAS_FLAG);
matrix = new Matrix();
int[] colors = new int[]{Color.BLACK, Color.BLACK, Color.TRANSPARENT};
float[] colorStops = new float[]{0f, 0.5f, 1f};
gradientShader = new RadialGradient(GRADIENT_RADIUS, GRADIENT_RADIUS, GRADIENT_RADIUS, colors, colorStops, Shader.TileMode.CLAMP);
mPaint.setShader(gradientShader);
bitmapGradient = Bitmap.createBitmap((int)(GRADIENT_RADIUS * 2), (int)(GRADIENT_RADIUS * 2), Bitmap.Config.ARGB_8888);
bitmapDest = bitmapGradient.copy(Bitmap.Config.ARGB_8888, true);
Canvas canvas = new Canvas(bitmapGradient);
canvas.drawRect(0, 0, GRADIENT_RADIUS * 2, GRADIENT_RADIUS * 2, mPaint);
canvasDest = new Canvas(bitmapDest);
}
@Override
protected void onMeasure(int widthMeasureSpec, int heightMeasureSpec) {
super.onMeasure(widthMeasureSpec, heightMeasureSpec);
int width = getMeasuredWidth();
setMeasuredDimension(width, width);
}
@Override
protected void onDraw(Canvas canvas){
/*uncomment each of them to show the effect, the first and the third one worked, the second show the same problem as OP's*/
//drawWithLayers(canvas); //unrecommended.
//drawWithBitmap(canvas); //this shows transparent as black
drawWithBitmapS(canvas); //recommended.
}
@SuppressLint("WrongCall")
private void drawWithLayers(Canvas canvas){
mPaint.setXfermode(new PorterDuffXfermode(PorterDuff.Mode.DST_IN));
float width = canvas.getWidth();
float hWidth = width / 2;
//both saveLayerAlpha saveLayer worked here, and if without either of them,
//the transparent area will be black.
//int count = canvas.saveLayerAlpha(0, 0, getWidth(), getHeight(), 255, Canvas.ALL_SAVE_FLAG);
int count = canvas.saveLayer(0, 0, getWidth(), getHeight(), null, Canvas.ALL_SAVE_FLAG);
super.onDraw(canvas);
float scale = hWidth/GRADIENT_RADIUS;
matrix.setTranslate(hWidth - GRADIENT_RADIUS, hWidth - GRADIENT_RADIUS);
matrix.postScale(scale, scale, hWidth, hWidth);
gradientShader.setLocalMatrix(matrix);
canvas.drawRect(0, 0, width, width, mPaint);
canvas.restoreToCount(count);
}
@SuppressLint("WrongCall")
private void drawWithBitmap(Canvas canvas){
super.onDraw(canvas);
float scale = canvas.getWidth() / (GRADIENT_RADIUS * 2);
matrix.setScale(scale, scale);
mPaint.setXfermode(new PorterDuffXfermode(PorterDuff.Mode.DST_IN));
canvas.drawBitmap(bitmapGradient, matrix, mPaint); //transparent area is still black.
}
@SuppressLint("WrongCall")
private void drawWithBitmapS(Canvas canvas){
float scale = canvas.getWidth() / (GRADIENT_RADIUS * 2);
int count = canvasDest.save();
canvasDest.scale(1/scale, 1/scale); //tell super to draw in 1/scale.
super.onDraw(canvasDest);
canvasDest.restoreToCount(count);
mPaint.setXfermode(new PorterDuffXfermode(PorterDuff.Mode.DST_IN));
canvasDest.drawBitmap(bitmapGradient, 0, 0, mPaint);
matrix.setScale(scale, scale); //to scale bitmapDest to canvas.
canvas.drawBitmap(bitmapDest, matrix, null);
}
}
一些注意事项： 1、此视图扩展ImageView不View，存在一些差异。
2、为什么drawWithLayers--saveLayer或saveLayerAlpha--不推荐： a、它们不确定，有时不能正常工作（显示透明为黑色），特别是对于ViewwhoesonDraw(Canvas)是空的，而ImageView.onDraw(Canvas)用aDrawable来画一些；b、价格昂贵，分配off-screen bitmap存储临时绘制结果，没有任何资源回收机制的明确线索。
3、使用自己的bitmap[s]，更适合自定义资源回收。
有人说，每次绘制都不能使用PorterDuff，不分配bitmap[s]，因为在绘制/布局/测量之前无法确定位图的宽高。
您可以使用缓冲位图 [s] 进行 PorterDuff 绘图：
首先，分配一些足够大的位图 [s]。
然后，使用一些矩阵在位图[s] 上绘制。
然后，将位图 [s] 绘制到带有一些矩阵的 dest 位图中。
最后，用一些矩阵将 dest 位图绘制到画布中。
有些人推荐 setLayerType(View.LAYER_TYPE_SOFTWARE, null)，这对我来说不是一个选项，因为它会导致 onDraw(Canvas) 在循环中被调用。
结果图像
源图像
