Tuesday, April 21, 2009

JPEG chopped and screwed 2

Here's the latest (disclaimer: some photoshop used as well).

I'm starting to learn how to make educated guesses about what part of the image will be effected.

The following bullets refer to the text file which results when opening up a JPEG in TextEdit or another basic text editor.

  • The top section should be left untouched, this seems to contain the code that identifies the file as a JPEG (even slight changes to this code may render your image file unreadable)

  • The next few lines after the above section often relate to color. These are fun to play with, and only require very small changes. Try removing single characters, or small groups of characters for big results.

  • The remainder of the text document describes the structure of the image. It also contains color information

    • Changes are less drastic than in other parts of the document (sometimes very large sections of text can be deleted without much change to the image)

    • Deleting too much information will result in grey bars, or the removal of whole lines of the image

    • Copying, pasting, and writing new text seems to have the best effects in this area (though I have been using the delete key almost exclusively)

    • Generally, the image is described from the top down. Ex: changes made to the bottom of the document are likely to affect the bottom of the image.

  • JPEGs created in Photoshop are structured a bit differently than those found on the Web, and are harder (in my experience) to work with in this way

Oh, also... I have found that saving the files as "Western (Mac OS Roman)" works best for me. I originally tried "Unicode (UTF-8)" but it seems that this doesn't have a large enough alphabet to accommodate all the strange characters created by the JPEG.

No comments: