View source for Tomba! 2: The Evil Swine Return/Technical information

This page contains technical information on the ROM data.

__TOC__

== Video analysis ==
{| class="wikitable"
|+
!Video description
!Link
|-
|Compression in Tomba 2
|https://youtu.be/f6Vh9b1Kiw8
|-
|CVRAM files
|https://youtu.be/iMHLHivs4mM
|}

==TOMBA2.IDX==
The IDX is the most important file in the entire game. It consists of several chunks of data, all with a size of <code>0x800</code>. One Chunk represents one area or "scene" so to speak and contains explicit information on what data to load to RAM for the assets and data and which VRAM to use, etc. We aren't loading anything to RAM, but extracting this data to real binary files to look closer. 

For example, the demo IDX is <code>0x18000</code> bytes big, so it contains <code>0x30</code> chunks, as <code>0x18000/0x800=0x30</code>. To get to the address of a given chunk_index, just seek to chunk_index * <code>0x800</code>. A chunk is made from two parts; the starting <code>0x100</code> part and the trailing <code>0x700</code> part. The first <code>0x100</code> bytes of a given chunk are:
 <code>'''"<5I" = img_start, img_end, dat_start, dat_end, pointer_amount'''</code>

* <code>img_start</code> and <code>img_end</code> determine the range of bytes to get from the IMG to create CVRAM for chunk_index.
*<code>dat_start</code> and <code>dat_end</code> determine the range of bytes to get from the DAT to create SDAT for chunk_index. Notice: neither of these variables determines a "size" or "amount." Only the range.
*pointer_amount determines the amount of pointers to read immediately after, so read <code>"<~I"</code> (where <code>~</code> is <code>pointer_amount</code>).

The list you get from doing so will contain items that each hold two values:
 '''<code>item >> 24 = dat_id</code>
 <code>'''item & 0x00FFFFFF = dat_ptr'''</code>
These two values should be put into a tuple, and all the

tuples should be put to a list called <code>sdat_pointers</code>.
 <code>'''sdat_pointers = [ (dat_id,dat_ptr), (dat_id,dat_ptr), ... ]'''</code>
Any data left up until the trailing <code>0x700</code> part should be ignored.

The trailing <code>0x700</code> bytes of a given chunk are:
 <code>'''"<224I" = tomba_list'''</code>

*Iterate through this list with a step of 2 with iterator t, where:

  <code>'''tomba_list[t], tomba_list[t+1] = dat_tomba_start, dat_tomba_end'''</code>
Much like its <code>img_</code> and <code>dat_</code> counterparts, these values represent addresses in

the DAT that contain relevant Tomba model data to use for that particular <code>chunk_index</code>.

There are multiple addresses in <code>tomba_list</code>, but any value of <code>0</code> should be ignored.

What the information in one chunk_index in the IDX gives us:
{| class="wikitable"
|+
|'''CVRAM file'''
|a small piece cut from the IMG file (Use unIMG.py to decompress)
|-
|'''SDAT file'''
|a small piece cut from the DAT file
|-
|'''A list called <code>sdat_pointers</code>'''
|a list with pointers to all data in the SDAT. Every pointer is accompanied by an index value (why?)
|-
|'''A list called <code>tomba_list</code>'''
|a list with data ranges to Tomba in his various suits
|}
== TOMBA2.DAT ==
DAT file contains:

- spirtes

- 3D models (characters and level geometry)

- text https://www.youtube.com/watch?v=Bbk2eXQPL68

- asset animation data

- level collision data

- "drawmaps"

- background tiles 

==BIN files==
BIN files contain asset placement data in every level. The file seems to be compressed.