Understanding Android Image Compression: Fundamentals, Algorithms, and Luban Implementation

The article begins with a brief overview of image compression concepts such as ARGB colour model, bitmap representation, and the difference between bit depth and colour depth, illustrating how these affect memory usage and file size on Android devices.

It then describes Android's two main compression techniques: quality compression (adjusting JPEG/WEBP quality) and size (sampling) compression, providing Java code snippets for both approaches:

ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
bitmap.compress(Bitmap.CompressFormat.JPEG, quality, outputStream);

BitmapFactory.Options options = new BitmapFactory.Options();
options.inSampleSize = 2;
Bitmap bitmap = BitmapFactory.decodeFile("xxx.png", options);

The article introduces the Luban library, explains its background, and walks through its algorithmic steps—ratio classification, boundary checks, size calculations, and iterative compression—highlighting the empirical values derived from reverse‑engineering WeChat's compression strategy.

It also analyses the original Luban implementation, pointing out issues such as lack of memory pre‑checks, fixed quality value, missing format options, and inefficient stream handling, then proposes a modernised solution using coroutines, LiveData, and configurable parameters.

Further, the piece delves into the Android graphics stack: how Bitmap operations flow from Java to native Skia, the role of Huffman coding in libjpeg, and the impact of the optimize_coding flag on compression efficiency. It notes that Android 7.0+ enables optimized Huffman tables, reducing file size by roughly half.

Finally, a native C++ example demonstrates manual JPEG encoding with libjpeg‑turbo, showing how to extract RGB data from an Android Bitmap, configure the compressor (including optimize_coding = TRUE ), and write the compressed image to disk.