Skip to content

Latest commit

 

History

History
68 lines (58 loc) · 4.04 KB

Header.md

File metadata and controls

68 lines (58 loc) · 4.04 KB
Raw

The first few KBytes

Rough notes from reading the hexdump of the stuff before the ogg files.

Header:

The first 32-bit-word is an offset into the file.

At that position, there is a sequence of of more offsets (32-bits). I call this the main table.

Every object in the book has a 16 bit OID code. When the user points at any object the tiptoi identifies its 16 bit OID code, subtracts the first used OID code and takes the result as the offset to find the corresponding jump table. The base for this offset is the beginning of the maintable + 8 (first two 32bit words are skipped)

For most files(?), the main table consists of

  • 32bit: last used OID code
  • 32bit: first used OID code
  • Then, 32-bit offsets that point to (what I call) jump tables (see below).
  • At some of these offsets are 0xFFFFFFFF. These indicate that the corresponding OID code is not used within the book.
  • These correspond linearly to the OID codes. E.g. WWW_Bauernhof: The first piglet has OID-code 1499, the corresponding jump table is at 0x766A. This offset is the 100th entry of the main table (4 bytes for each entry). So possibly 4*(OID - 1401) = main table index + 8. The Value 1401 in this example is taken from the second 32bit word in the main table, it represents the first used OIS code within the current book.
  • The end of the offsets can be found at (maintable + 8 + 4*(last used OID code - first used OID code).

There is more data contained at the beginning of the file:

  • 0x0000 the first 32-bit word pointer to the main table
  • 0x0004 The second 32-bit word is a pointer to the media table.
  • 0x0008 unknown 32-bit word, value always 0x000238b
  • 0x000c points to a 16bit word that is alwaya 0x0000 (no exceptions found so far)
  • 0x0010 32-bit-word (0x0001703b) is an offset into the file. This offset is always direcIts value is always 2 bytes after the There is a number (32-bit, 0x000 000c = 12) followed by that many offsets right after the list. These offsets point to right after the list and spread out towards the next known block (the media file table), so its objects are relatively large (~1k). The objects themselves contains numbers and offsets.
  • 0x0014 The sixth 32-bit-word is a small number containing the product id code (== OID code of the power on symbol on page 1)
  • 0x0018 The seventh 32-bit-word points to the data directly following the list of the welcome samples. Maybe this means something like "end of table data"?
  • 0x001c The eighth 32-bit-word is a small number, differing between books.
  • 0x0020 is a byte that contains the lenght of the Chomptech data format version string.
  • 0x0021 Then follows the string CHOMPTECH DATA FORMAT CopyRight 2009 Ver2.5.0908 or similar, resembling the length given at `0x0020
  • directly after that there is a date string like 20111024 followed by zero bytes up to 0x005f. For some books the date contains a language string, e.g 20111024GERMAN or 20111002DUTCH
  • 0x0060 For some books a 32bit address pointing to a place where another pointer can be found. Otherwise 0
  • 0x0064 to 0x0070 are 0
  • 0x0071 This points to a list that contains the address of the welcome samples played when the power on symbol is pointed at.
  • 0x0075 to 0x008b are 0
  • Beginning at 0x008c some books have more data (otherwise 0): 0x008c oooooooo 32-bit offset 0x0090 oooo 16-bit offset? 0x0092 nnnn 16-bit number 0x0094 oooo 0x0096 nnnn 0x0098 oooo 0x009a nnnn 0x009c 0000 0x009e 0000 0x00a0 oooo 0x00a2 nnnn 0x00a4 0000 or 0001 0x00a6 0000 0x00a8 oooo 0x00aa nnnn
    • from 0x00ac to 0x01ff all bytes are 0

Jump table locations

Where are the others referenced from?

For example there is a (very small, one line with one F command) jump-table at 0x35C0 in WWW_Bauernhof.gme, but that offset does not occur in the file. Maybe just some dead code.