关于android缩放bitmap，BitmapFactory.Options.inSampleSize的优化算法

安利一下这个 GitHub 上别人汉化好的 Google Android Training 

地址是 http://hukai.me/android-training-course-in-chinese/index.html

这个 GitHub 项目地址是 https://github.com/kesenhoo/android-training-course-in-chinese

----------------------------------------------上面是广告---------------------------------------------------------------------------------------------------------------------------------------------------------------------

在android官网网站上的 Develop > Training >  Displaying Bitmaps Efficiently > Loading Large Bitmaps Efficiently 章节看到的计算 

BitmapFactory.Options.inSampleSize的算法，没看懂这官方示例代码的计算逻辑（好吧，这其实怪我数学不好~~~ - -b）

然后根据这个值的要求，我写了一段更符合逻辑更容易看懂的代码~~ duang~duang~duang~ 如下所示（至于这个值怎么计算的，有什么用，如果你不造的话，可以去了解下，据说这部分内容在android开发过程中还挺重要的、挺基础的~~据说。。。）：

        // 计算缩放比
        // Calculate the largest inSampleSize value that is a power of 2 and keeps both
        // height and width larger than the requested height and width.
        int n = 0;
        int inSampleSize = 1;
        long startTime = System.nanoTime();
        while (imageHeight / Math.pow(2, n) > screenHeight
                && imageWidth / Math.pow(2, n) > screenWidth) {
            n++;
        }
        inSampleSize = (int) Math.pow(2, n);
        long endTime = System.nanoTime();
        Log.i("----scs", "计算 inSampleSize 耗时 " + (endTime - startTime) + " ns.");
        Log.i("----scs", "inSampleSize: " + inSampleSize);
相对应的，官方的计算方法是：

        long startTime1 = System.nanoTime();
        if (imageHeight > screenHeight || imageWidth > screenWidth) {

            final int halfHeight = imageHeight / 2;
            final int halfWidth = imageWidth / 2;

            // Calculate the largest inSampleSize value that is a power of 2 and keeps both
            // height and width larger than the requested height and width.
            while ((halfHeight / inSampleSize) > screenHeight
                    && (halfWidth / inSampleSize) > screenWidth) {
                inSampleSize *= 2;
            }
        }
        long endTime1 = System.nanoTime();
        Log.i("----scs", "计算 inSampleSize 耗时 " + (endTime1 - startTime1) + " ns.");
        Log.i("----scs", "inSampleSize: " + inSampleSize);
然后，对比了一下，耗时如下：

06-24 09:45:57.745    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ imageWidth:1920, imageHeight: 1200
06-24 09:45:57.746    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ screenWidth:800, screenHeight: 480
06-24 09:45:57.747    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ 计算 inSampleSize 耗时 211846 ns.
06-24 09:45:57.747    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ inSampleSize: 4
06-24 09:45:57.747    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ 计算 inSampleSize 耗时 3077 ns.
06-24 09:45:57.748    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ inSampleSize: 4

翻转了一下屏幕：
06-24 09:46:31.276    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ imageWidth:1920, imageHeight: 1200
06-24 09:46:31.278    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ screenWidth:480, screenHeight: 800
06-24 09:46:31.278    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ 计算 inSampleSize 耗时 12846 ns.
06-24 09:46:31.278    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ inSampleSize: 2
06-24 09:46:31.278    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ 计算 inSampleSize 耗时 3385 ns.
06-24 09:46:31.278    8801-8801/com.jumple.scs.myapp1.app I/----scs﹕ inSampleSize: 2

果然第一种写法的耗时比较长（估计是因为第一种计算过程中使用的是double 而官方的方法只使用了int值来计算吧），哔哔了这么多，其实只是想说......我需要去补习数学了......