Anyone that has worked with me will know that I love low level and/or old technology. So when you mix the two together, I am in my element! QuickDraw is an old, low level library for handling media. It first appeared on the earliest versions of the Macintosh System Software. It should have been right up my street.
It turns out old technology is not always documented well. In some cases it’s not really documented at all. Once upon time QuickDraw was very well documented. Well it was well documented from an API standpoint. The internals were… well there was some information about how it worked, but precious little. A document does still exist online for it. It’s in the archived section of Apple’s old websites. It’s basically a PDF of a scanned document. Not fantastic, but better than nothing.
QuickDraw it seems went through several iterations. Thankfully, by the time the Escape Velocity series came on to the scene. Version 2 of QuickDraw the default, and thus all picture resources appear to be encoded as such. This is both a blessing and curse.
Version 2 is the only one that I’ve had any luck in finding documents for. Even then there isn’t much. That document mentions a tiny amount about Version 1. The only thing I could learn was, that Version 1 was basically easier to parse. Damn. That means more work in trying to get the picture decoder up and running.
There are 2 major difficulties with the Picture Resources in QuickDraw. These are of course only the challenges that I’ve encountered. It turns out that EV Nova data files only seem to use a thin slice of the Picture Resource functionality.
Yes. The Picture Resource uses a series of op-codes to encode data. This is done primarily so that pictures can be encoded as a type of vector graphic. This is pretty cool considering when the format was designed. However there is only really one op-code that is needed for EV Nova. That is the Draw Direct Packed Bits operation. This opcode basically reads a massive chunk of data from the resource. This data is an encoded bitmap.
This is an algorithm that is used to compress/pack bitmap information into a smaller space. The exact algorithm is notoriously difficult to find online and there appears to be multiple variants of the algorithm.
On top of those issues, the colour space for the original pictures was RGB 555, which is no longer a ready defined colour space on the Mac. So the bitmap information once unpacked needs to be translated into RGBA 8888.
When it goes wrong…
There is so much minutia in the format that it is extremely common to end up with results that appear something like you can see above, or worse. A lot of images will still fail to parse correctly due to some unexpected opcodes that I’ve yet to handle or decipher.
And this is just the picture resource. There is still the cicn, rlë, ppat, and other QuickDraw formats to handle.