Writing 8-bit tiff images?

Processing Project Guidance
1

Hi, we are using Processing to acquire data and write a video stream to disk. However, we are unable to figure out how to write one-channel 8-bit grayscale tiff images instead of the standard RGB images. Does anyone have some pointers on how to do this? Thank you.

1 Like
2

@tmhoog

Look at :

3

Thanks for the pointer. We want to write only a single channel 8-bit grayscale image. My understanding reading the forum so far is that we need to write a custom function for this.

4

Yes, I think there’s not a built-in function in Processing for converting an image in 8-bit colors.

5

This is a hack function, but it works:

void convertAndSave8bitGray(PImage img) {
  Image iimg = img.getImage();
  BufferedImage bimage = new BufferedImage(iimg.getWidth(null), iimg.getHeight(null), BufferedImage.TYPE_INT_ARGB);
  Graphics2D bGr = bimage.createGraphics();
  bGr.drawImage(iimg, 0, 0, null);
  bGr.dispose();
  BufferedImage src = bimage;
  BufferedImage dest = new BufferedImage(src.getWidth(), src.getHeight(), BufferedImage.TYPE_BYTE_GRAY);
  ColorConvertOp cco = new ColorConvertOp(src.getColorModel()
    .getColorSpace(), dest.getColorModel().getColorSpace(), null);
  cco.filter(src, dest);
  File outputfile = new File(sketchPath("test2.jpg"));
  try {
    ImageIO.write(dest, "jpg", outputfile);
  } 
  catch(IOException e) {
  }
}
1 Like
6

That’s great. Thanks so much.

7

For beginners looking to retrieve grayscale data or create grayscale images more generally (not 8-bit tiff specifically), there are also the HSB color functions – saturation() and brightness(), which return single values. Brightness in particular will return a single greyscale value for each pixel on the canvas or on a PImage / PGraphics, and you can use this as a simple built-in way to loop over pixels[] and either convert to grayscale or create a grayscale copy that can then be saved to disk.

https://processing.org/reference/brightness_.html

You can also set colorMode(HSB) to hue-saturation-brightness and then use tint() to remove the saturation on display, creating a grayscale image – although I believe it will still save as grayscale data in RGB if using the default saveFrame / saveImage.

https://processing.org/reference/tint_.html

1 Like