libjpeg turbo android jpeg compression

Am using libjpeg-turbo from here to compress images using JPEG compression, since

Bitmap.compress(Bitmap.CompressFormat.JPEG, quality, ByteArrayOutputStream);

causes OOM for big Images in my Android Project.

Previously i was using libjpeg-turbo from here which caused me a lot of trouble so i shifted to the above Github by sorccu (I believe this was the latest or working fine library).

Android Studio - 1.4 RC1 Android Ndk - r10e Android SDK min = 14, compiled = 22

Also with Benjamin code i was trying with old Gradle 1.3.0 which caused lot of errors, after few weeks of trying i changed to gradle-experimental:0.2.0 and everything compiled Good.

Since i feel sorccu Github easy to integrate and use, shifted to it. then used the turbojpeg-jni.c in the library along with the java code for integration in android (both available in the library).

I was able to integrate and compile using ndk-build to build the .so file and copied to lib folder and was able to use it in the java code.

module name in gradle = jpegturbo module name in android.mk = jpegturbo

Built it as a shared library

library loading code in java = System.loadLibrary("jpegturbo"); (org.libjpegturbo.turbojpeg.TJLoader.java)

There are few problems i face now, please help me to resolve it,

1.The main issue i can't compress the image using it, the compressed image is corrupted may be am doing it wrong, below is the java code

static {
    try {
        tj = new TJCompressor();
    } catch (Exception e) {
        e.printStackTrace();
    }
 }

File file = new File(pathName);
int actualBytes = (int) file.length();

byte[] srcBuf = new byte[actualBytes];
byte[] dstBuf;

FileInputStream fIn;
try {
    fIn = new FileInputStream(file);
    fIn.read(srcBuf);
    fIn.close();
    } catch (FileNotFoundException e) {
        e.printStackTrace();
    } catch (IOException e) {
        e.printStackTrace();
    }

tj.setSourceImage(srcBuf, 0, 0, options.outWidth, 0, options.outHeight, TJ.PF_RGB);
tj.setJPEGQuality(quality);
tj.setSubsamp(TJ.SAMP_444);
dstBuf =  tj.compress(TJ.FLAG_ACCURATEDCT, options.outWidth, options.outHeight);

int size = tj.getCompressedSize();

tj.close();

FileOutputStream fos;
   try {
       fos = new FileOutputStream(file1);
       fos.write(dstBuf);
       fos.close();
   } catch (FileNotFoundException e) {
       e.printStackTrace();
   } catch (IOException e) {
       e.printStackTrace();
   }catch (OutOfMemoryError e) {
       e.printStackTrace();
   }

Log.e("SIZE ", String.valueOf(size));

May be the approach or logic in the above block can be wrong, help me !!!

2. i didnt change the application.mk from the github,

APP_ABI := armeabi armeabi-v7a armeabi-v7a-hard arm64-v8a x86 x86_64

i use below command to build .so file

ndk-build NDK_PROJECT_PATH=. APP_BUILD_SCRIPT=./Android.mk APP_ABI=armeabi-v7a LOCAL_ARM_NEON=true ARCH_ARM_HAVE_NEON=true

it generates .so file for ABI armeabi-v7a alone, i manually copy the.so file for other ABI's. i even tried below commands but its not helping

ndk-build NDK_PROJECT_PATH=. APP_BUILD_SCRIPT=./Android.mk APP_ABI=all

ndk-build NDK_PROJECT_PATH=. APP_BUILD_SCRIPT=./Android.mk

Also the final apk works in my mobile Sony Xperia U ST25i, but not in the emulator it says the .so file cannot be found, i didn't check in other devices

Note: - i may be wrong about the library author, but both the mentioned library creators have done a great job

i also referred the following links,

1. link 1
2. link 2
3. link 3

Update solved the issue 2, by changing Application.mk as APP_ABI := armeabi-v7a armeabi-v7a-hard x86 x86_64 arm64-v8a APP_PLATFORM := android-14

here's my ndk-build command ndk-build NDK_PROJECT_PATH=. APP_BUILD_SCRIPT=./Android.mk NDK_APPLICATION_MK=./Application.mk

After this app works ok in emulator too

Finally i made it work, with the help of this question

Thanks to the user who posted it.

The code exactly meet my requirement and am able to compress images.

Note: but problem for compressing big images (> 5-30 MB)

UPDATE:-

Problem for compressing big images (> 5-30 MB) can be solved by specifying the resizing dimensions as 1/8 of its original width and height (minimum).