Working on documenting, hacking HE SCUMM data.

[GeorgeQGreg]

Hello, I am working on documenting and hacking the HE SCUMM format, well specifically that of the classics such as Putt-Putt Joins the Parade. Right now I'm just working on the DOS demo of Parade, which come to think of it is probably pretty stupid on my part, as some formats are different in the demo. The goal of all this is to be able to translate the classics games into other languages, as besides Russian, Hebrew and Arabic, they were never translated. I'd like to be able to translate at least the subtitle data, and translating the speech and graphics would be nice too.

Anyway, here's what think I know so far:

File structure is pretty similar to DOTT
Background and object data seems to be the same format as DOTT
Actors however are in a different format. Possibly similar to Full Throttle?
Game data itself is encrypted with an XOR key of 69. Boy, those devs sure were mature. Bizarrely, the speech is also XOR encrypted. Beyond the XOR, text is pretty plain, just surrounded with opcodes, some of which the wiki's documentation may be able to help with.
I have a vague idea of what some opcodes do.
I have no idea of how the game searches it's data. Ergo, if I make the main data file bigger to put more data in it, there's a good chance that some of it will be in a different place, and that would screw things up. I think it's in the HE0, in some way.
The speech samples seem to be done this way: ASCII sequence TALK, followed by two null bytes, then two hex bytes that differ, the ASCII HSHD, three null bytes, hex code 18, two more null bytes and then 11khz 8 bit unsigned PCM data.

Corrections are welcome.