Java SE (Java Platform, Standard Edition)

Hi,

Im trying to solve something similar to the FFT problem found by this user (7429284).

The current method of conversion :

 short[] unsignedByteArray;
        int[] unsignedShortArray;
        shortArray = array;
        isBigEndian = bigEndian;
        int N = shortArray.length;

        unsignedShortArray = new int[N];
        for (int i = 0; i < N; i++) {
            unsignedShortArray[i] = shortArray[i] + 128 * 257;
        }
        byteArray = new byte[N * 2];
        unsignedByteArray = new short[N * 2];
        if (isBigEndian) {
            for (int i = 0; i < N; i++) {
                unsignedByteArray[i * 2] = (short) (unsignedShortArray[i] / 256);
                unsignedByteArray[i * 2 + 1] = (short) (unsignedShortArray[i] % 256);
            }
        } else {
            for (int i = 0; i < N; i++) {
                unsignedByteArray[i * 2 + 1] = (short) (unsignedShortArray[i] / 256);
                unsignedByteArray[i * 2] = (short) (unsignedShortArray[i] % 256);
            }
        }
        for (int i = 0; i < N * 2; i++) {
            byteArray[i] = (byte) (unsignedByteArray[i] - 128);
        }

ive tried tackling the problem 2 different ways, neither of which remove the noise from the output

attempt 1: More noise than current code

        shortArray = array;
        int N = shortArray.length;

        byte out[] = new byte[N * 2];
        for (int i = 0; i < N; i++) {
            out[i * 2] = (byte) (array[i] >>> 8);//most significant
            out[(i * 2) + 1] = (byte) array;//least significant

}

return out;





attempt 2: less noise than attempt 1, still more than current
        shortArray = array;
        int N = shortArray.length;
        int hexBase = 0xff;

        byte out[] = new byte[N * 2];
        for (int i = 0; i < N; i++) {
            out[i*2] = (byte) (hexBase & array);

out[(i*2)+1] = (byte) ((byte) ((hexBase <<8) & array[i]) >> 8);



}

return out;







I think solving this problem is beyond my capability, but if someone could help in producing a less noisy output im sure many more would benefit from a fully operational FFT implementation