Losing width and height information of Image???

Question

Hi,
I am losing the width and height information between converting an image to byte[] and then back to Image
public void testImageToBufferAndBack(Image takenImage){
BufferedImage bufimage = this.ImageToBufferedImage(takenImage);
System.out.println("buffimage"+bufimage.getWidth()+" "+bufimage.getHeight(null)); //works fine
ByteArrayOutputStream stream = new ByteArrayOutputStream();
try {
ImageIO.write(bufimage, "jpeg", stream);
catch (IOException ex) {
byte[] bytes = stream.toByteArray();
byte[] pictureBytes = bytes;
Image testImage = Toolkit.getDefaultToolkit().createImage(pictureBytes);;
System.out.println("tttttttteeeeeeeessssssttt "+testImage.getWidth(null)+" "+testImage.getHeight(null));
Output:
buffimage320 240
tttttttteeeeeeeessssssttt -1 -1
Could someone point me out my mistake please??

Answer 1

Your mistake is not realizing that java.awt.Image loads images asynchronously. When Toolkit's
createImage method returns an Image, that image contains no pixel data -- as you've seen above, it doesn't
even know the size of the image! I knew this was the behaviour of the createImage method that took a URL,
but was surprised here that this is also the case with the version that takes a byte[], but undoubtably, that is
what is happening here. The way to fix this is always the same: use a MediaTracker, as my demo below
illustrates.
A couple points:
1. If you render your image in a custom component like this it will eventually appear:
g.drawImage(image, x, y, this); //for exampleThat's because Component's default behaviour as an ImageObserver (the reason for the final this)
is to continually repaint as more pixel data becomes available. This is not always the best solution -- what
if you need to know the image's dimension sooner rather than later?
2. ImageIcon can load an image for you, by using a MediaTracker behind the scenes:
Image image = ...
new ImageIcon(image);
//now image is loaded...Just be aware that if you use its constructors that don't take Image, but for example URLs
URL url = ...
Image image = new ImageIcon(url).getImage();Behind the scenes it's first getting the Image with Toolkit's getImage (not createImage). Toolkit's
getImage caches the image, which may not be what you want if you are processing many images --
why have the image loiter in memory if there is no advantage to that?
import java.awt.*;
import java.awt.image.*;
import java.io.*;
import java.net.*;
import javax.imageio.*;
public class Example {
    public static void main(String[] args) throws IOException {
        String url = "http://today.java.net/jag/bio/JagHeadshot-small.jpg";
        BufferedImage original = ImageIO.read(new URL(url));
        displayDimensions("original buffered image", original);
        ByteArrayOutputStream out = new ByteArrayOutputStream();
        ImageIO.write(original, "jpeg", out);
        byte[] bytes = out.toByteArray();
        Image image = Toolkit.getDefaultToolkit().createImage(bytes);
        displayDimensions("created image", image);
        loadImage(image, new Label()); //any component will do
        displayDimensions("loaded image", image);
        System.exit(0);
    //handy dandy utility method
    public static void loadImage(Image image, Component comp) throws IOException {
        MediaTracker mt = new MediaTracker(comp);
        mt.addImage(image, 0);
        try {
            mt.waitForID(0);
        } catch (InterruptedException e) {
            throw new RuntimeException("Unexpected", e);
        if (mt.isErrorID(0))
            throw new IOException("Error during image loading");
    static void displayDimensions(String msg, Image image) {
        System.out.println(msg + ", w=" + image.getWidth(null) + ", h=" + image.getHeight(null));
}Output:
original buffered image, w=235, h=240
created image, w=-1, h=-1
loaded image, w=235, h=240