I have found the pointer table location for the good endings. This is really good news. It means my greatest fear, of the pointers being embedded in the code, aren't reality. I can work with this. But, I haven't been able to decode the pointer table yet. It's storing pointers for more than just text, and I'm not sure how the values are calculated. Some of the values are even repeated for some reason. I'm really not sure how to work with it yet, but at least I have a place to start now. I've messed around with it a little bit and have been successful just repeating different values in different places to kind of change the order the ending visuals and audio play out, but I'm still not sure how to calculate a new pointer value for newly inserted text.
There are actually two tables. There is a very small one right before the text, and a larger (although still pretty small compared to the ones I've been working with) at the very end of the file. I'm not sure what the difference is between the two either. Definitely need to do some more monkeying around and reverse engineering to figure it out. The table near the end of the file actually contains the references to the ending text and such though. I have figured that much out.
So, from what I've discovered with my latest testing, the pointer table only points to the start of the dialog when the picture changes. The dialog itself has the encodings for the audio files to play and line breaks and clearing the screen. The Kyosuke ending only has 4 pictures, so 4 pointers.
The pointer table near the end of the file appears at 0x1E0C. There are definitely more than 4 entries in this table, and I've found the pointers that appear to point to the text for each picture segment (except the first one, but it really isn't important as it can always start in the same place). Each line of dialog begins with a 2 byte code telling it which audio files to play. Each audio file is only 1 byte, so you can play 2 audio clips per dialog line. This is exactly how the bad endings work too. After the 2 byte audio command, there is always a 0x1E byte. I don't know what this is for. Changing it does not appear to affect anything. At the end of the dialog line, there is always a 0x0002 to end the dialog. I assume 0x00 ends the string and 0x02 clears the screen? I'm not sure. But not having the 0x02 breaks stuff. Some lines end with 0x04 or 0x07 instead, but there doesn't seem to be any difference between any of them.
I need to figure out how these pointers are calculated and what the meaning of the other pointers are.
For my own documentation purposes here are the details I know about the pointers.
the pointer table appears to start at 0x1E0C. The pointers I've found to be important are
0xECC7 = points to the first scene / picture
0xE8C8 = points to the second scene picture / text
0x14CA = points to the third scene picture / text
0x90CB = points to the fourth / final scene picture / text.
All the opcode stuff from before is still the same, as far as playing the audio clips and such.
Until I can figure out the pointer math (and maybe it's not even possible,) I've proceeded with doing it the old fashioned, pointer-less way. It's not as flexible, but I've checked with 13 of the19 endings so far and haven't run into any issues. It is limiting in the translation, since my text length has to match the original size (or close to it. There are some unused empty bytes in there sometimes that can add just enough flexibility). But I've already written an inserter for the good things using this method and it's working quite well. I need to finish creating the input files for the inserter, but it's progress at least.
| Process | Progress |
|---|---|
| Translation | 20% |
| Proof reading | 0% |
| Insertion | 10% |
| Testing | 0% |
Comments
Post a Comment