Why wont this work?

Out of context: Reply #15

  • Started
  • Last post
  • 17 Responses
  • jawbreaker0

    you might be referring to this bug...

    bitmap distortion problems
    From: Tinic Uro

    Date: Mon, 30 Dec 2002 16:26:40 -0800

    Yes, we know this is a bug and a quite pressing issue for a lot of
    people. We refer to it as the 'bitmap wrapping bug'. The last minor
    revision of the Flash player fixed the 'bitmap shifting bug' which
    occured on 32bit displays and varied alpha values or masked bitmaps. You
    usually saw this shifting when fading in bitmaps. The 98% alpha
    workaround usually fixed this issue while having a impact on overall
    performance (anything with alpha is drawn with a performance hit).

    The 'bitmap wrapping bug' will be fixed in one of the next major Flash
    player releases. Why major version you will ask?

    Unfortunatly this is a bug which will require a change in your .fla
    files in many cases. Most authors apply a workaround on the stage, e.g.
    cut off one pixel row from their bitmap or use the transparent border
    workaround. We do not think it is acceptable to expect people to change
    their .fla files for a minor revision and remove workarounds.

    For the sake of stability and backwards compability the fix will only be
    applied for new content. You will have to deal with issue for <= 6.x
    Flash content.

    Personally I really feel your pain, I had to deal with this myself. It
    makes features like dynamically loading JPEG files quite limited. We are
    commited to address the issue though and I welcome any feedback on this
    issue or any other issues with bitmaps you might have. The earlier we
    can get our hand on test files to verify our fixes, the better.

    A more technical explaination of what happens: Due to the fact that we
    use 16.16 fixed point integer values for our tranformation matrixes,
    some bits get lost during some operations, namely matrix
    multiplications. Usually the TX and TY field of a transformation matrix
    should be both 0 for the most common case. Because of the inaccuracy it
    becomes 0xFFFFxxxx internally which at the end gets rounded off to -1
    instead of 0. This is why the bitmaps are shifted by one pixel. This
    also only happens if you use antialiasing.

    MORE:
    http://chattyfig.figleaf.com/cgi…

View thread