Problem in reading BMP file .....

Question

I've facing some inconsistency in reading BMP headers. There are two of them, and I've outlined them below....
Here's a description of the BMP headers, just for refreshing ....
1.
     typedef struct {
     BITMAPFILEHEADER bmfHeader;
     BITMAPINFOHEADER bmiHeader;
     GLubyte *image_data;
     } BITMAP_IMAGE;
     typedef struct tagBITMAPFILEHEADER {
          WORD bfType;               // 2B
          DWORD bfSize;               // 4B
          WORD bfReserved1;     // 2B
          WORD bfReserved2;     // 2B
          DWORD bfOffBits;          // 4B
     } BITMAPFILEHEADER, *PBITMAPFILEHEADER;
     typedef struct tagBITMAPINFOHEADER{
     DWORD biSize;
     LONG biWidth;
     LONG biHeight;
     WORD biPlanes;
     WORD biBitCount;
     DWORD biCompression;
     DWORD biSizeImage;
     LONG biXPelsPerMeter;
     LONG biYPelsPerMeter;
     DWORD biClrUsed;
     DWORD biClrImportant;
     } BITMAPINFOHEADER, *PBITMAPINFOHEADER;
2. The file I'm reading test2.bmp has the following parameters
     File Type                     Windows 3.x Bitmap (BMP)
     Width                         4
     Height                         4
     Horizontal Resolution     96
     Vertical Resolution          96
     Bit Depth                    24
     Color Representation     True Color RGB
     Compression                    (uncompressed)
     Size                         102 bytes
     Size on disk               4.00 KB (4,096 bytes)
3. Output of the program is....
          File Headers are...
          bfType :424d
          bfSize :102
          bfReserved1 :0
          bfReserved2 :0
          bfOffBits :54
          File Info Headers are...
          biSize :40
          biWidth :4
          biHeight :4
          biPlanes :0
          biBitCount :1
          biCompression :1572864
          biSizeImage :48
          biXPelsPerMeter :196
          biYPelsPerMeter :234881220
          biClrUsed :234881024
          biClrImportant :0
4.      readInt()     4
          Reads four input bytes and returns an int value.
     readUnsignedShort()     2
          Reads two input bytes and returns an int value in the range 0 through 65535.
     readByte()
          Reads one Byte
5. The bfType fields value should be 'BM' whose Hex is 0x4D42 according to this link
http://edais.earlsoft.co.uk/Tutorials/Programming/Bitmaps/BMPch1.html
But the Hex I get is 424d.How come ?
6. When reading bfSize (see ## ), we should read 4 bytes as per the above structure. But I don't
get the answer of 102 when I do readInt() (which reads 4B). On the contrary, when I do
readByte(), thats when I the right file size of 102.
Why ?
Rest all the output look ok.
import java.io.*;
class BMPReader{
     private byte data[];
     private DataInputStream dis = null;
     public BMPReader(String fileName){
          readFile(fileName);
     private void readFile(String fileName){
          try{
               dis = new DataInputStream(new FileInputStream(new File(fileName)));
          catch(Exception e){
               System.out.println(e);
     public void printFileHeader(){
          System.out.println("File Headers are...");
          try{
               // converting the 2 bytes read to Hex
               System.out.println("bfType :" + Integer.toString(dis.readUnsignedShort(),16));
               System.out.println("bfSize :" + dis.readInt());
               //System.out.println("bfSize :" + dis.readByte());// ## identifier, should read 4B here instead of 1B
               System.out.println("bfReserved1 :" + dis.readUnsignedShort());     // bfReserved1
               System.out.println("bfReserved2 :" + dis.readUnsignedShort());     // bfReserved2
               System.out.println("bfOffBits      :" + dis.readInt());               // bfOffBits
          catch(Exception e){
               System.out.println(e.toString());
     public void printFileInfoHeader(){
          System.out.println();
          System.out.println("File Info Headers are...");
          try{
               System.out.println("biSize           :" + dis.readInt());                         //DWORD
               System.out.println("biWidth      :" + dis.readInt());                     //LONG
               System.out.println("biHeight      :" + dis.readInt());                     //LONG
               System.out.println("biPlanes      :" + dis.readUnsignedShort());          //WORD
               System.out.println("biBitCount      :" + dis.readUnsignedShort());     //WORD
               System.out.println("biCompression     :" + dis.readInt());               //DWORD
               System.out.println("biSizeImage      :" + dis.readInt());               //DWORD
               System.out.println("biXPelsPerMeter :" + dis.readInt());          //LONG
               System.out.println("biYPelsPerMeter :" + dis.readInt());          //LONG
               System.out.println("biClrUsed          :" + dis.readInt());                    //DWORD
               System.out.println("biClrImportant      :" + dis.readInt());               //DWORD
          catch(Exception e){
               System.out.println(e.toString());
public class BMPReadForum{
     public static void main(String args[]){
          String fName = "test2.bmp";
          BMPReader bmpReader = new BMPReader(fName);
          bmpReader.printFileHeader();
          bmpReader.printFileInfoHeader();
}

Answer 1

I stumbled across this thread while messing with numbers coming out of my palm (via pilot-xfer) that are unsigned 32 bit values read into a 4 element byte array. I would very much like to turn these things into longs. I looked over the above posted code (for which I am most grateful) and made some alterations that seemed prudent, would someone be so kind as to verify that this is doing what I would like it to do?
private long makeUInt(byte[] b) {
    // Convert a four byte array into an unsigned
    // int value. Created to handle converting
    // binary data in files to DWORDs, or
    // unsigned ints.
    long result = 0;
    int bit;
    int n = 0;    // I assumed this was the power to take the base to - 0 for lsb
    // I further assumed that we need to move backwards through this array,
    // as LSB is in the largest index, is this right?
    for (int a=3; a >= 0; a--) {
        // Similarly, I need to step through this backwards, for the same
        // reason
        for (int i=7; i >= 0; i--) {
            bit = b[a] & 1;
            result += bit * Math.pow(2, n++);
            b[a] >>>= 1;
    return result;
}So, as you see - I assumed the "n" term was a count for what power of 2 we were in in the bit string, and I further assumed that, since for me the MSB is in the byte array element with the largest index, I needed to work backwards through this thing.
Does this seem reasonable?
Thanks a lot
Lee