The Infinity Engine Reimplementation Project written in Python by Sebastian Hagen
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is even with uyjulian:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
fonts
src
.gitignore
COPYING
README.md
be17p.py
e17p.py
setup.py

README.md

e17p

Purpose

The 'infinity series' is a set of proprietary visual novels produced by the semi-defunct company KID ('Kindle Imagine Develop'). Various proprietary engines for these games are available, including for Win32 and PSP, but at the time of writing there is no Free engine capable of playing these games natively on Unix-like OSes.

The purpose of e17p is to create such an engine by reverse-engineering the storage formats used by these games, and writing a free parser based on the knowledge gained this way.

Work is currently focused on Ever17, though some preliminary work has been done on the Remember11 formats as well.

License

GNU GPL v2, or (at your option) any later version. See the included COPYING file for the license text for version 2.

Dependencies

For everything

  • Python 3.x.

For the GUI VN engine (when that gets written)

  • Pygame

For CPS->PNG conversion

  • convert(1) from ImageMagick

Formats and parsers (Ever17)

LNK files

The top-level media data (*.dat) files of Ever17 are in a format that has been dubbed 'LNK'. Almost all of the VN data used by the Ever17 engine is stored in these files; the known exceptions are savegame data and movie files. LNK files are simple tarball-like archives.

Extraction syntax:
./e17p.py ever17.ff.lnk foo.dat outdir

CPS files (from bg.dat, chara.dat)

CPS files contain image data. This data is both obfuscated and compressed. The deobfuscation and decompression algorithms have been reimplemented. The actual plaintext image format contained inside would be more appropriately referred to as 'PRT'; however, plain PRT files have so far not been observed in the Ever17 archives, and as such the parsers for both are currently integrated.

PRT files use a bitmap format; the main image data is typically stored in RGB with a color depth of 24 bits. Character art also includes alpha channel information, which is stored seperately after the color bitmap.

CPS information is extracted using ./e17p.py ever17.ff.cps. Usage modes:
Just unobfuscation: ./e17p.py ever17.ff.cps -u foo.cps
Unobfuscation and decompression: ./e17p.py ever17.ff.cps -d foo.cps
Unobfuscate, decompress, and convert to BMP: ./e17p.py ever17.ff.cps -b foo.cps
(Note that standards compliance of the produced BMPs is somewhat iffy; in any case, some programs won't be able to read them. The primary use of this output mode is to feed it to convert(1) for recompression, which works fine.)

Unobfuscate, decompress, and recompress to PNG:
./e17p.py ever17.ff.cps -p foo.cps
(Being options, these can be mixed freely; see --help for details.)

Alternatively, this program can also be used directly on CPS-containing LNK files, e.g.: ./e17p.py ever17.ff.cps -p -o outdir --lnk chara.dat

scr (script.dat)

Are script files, containing both text (conversation/narration) data as well as the main game logic. We have a functioning tokenizer for the first, as well as largely functioning tokenizer and partial parser (everything except system.scr and startup.scr can be tokenized; semantics of many opcodes are still unknown or questionable, however) for the second.
See --help for the full list of options.
Simple usage:
Dump parsed colorized conversation data: ./e17p.py ever17.ff.scr -tcu foo.scr bar.scr ...
Dump parsed colorized event script data: ./e17p.py ever17.ff.scr -tsu foo.scr bar.scr ...

WAF (bgm.dat, se.dat)

Audio files for background music and sound effects; contain uncompressed data in ADPCM format. Conversion to wav is possible using:
./e17p.py ever17.ff.waf foo.waf -w

WAF (voice.dat)

Audio files containing voice clips. While these have the same extension as the files described in the previous section, the format is different in that the entire file contents have been RLE encoded (using the same scheme as used in CPS files) and prepended with an additional ('lnd') header. ff_dat_waf is capable of recognizing these files by header; as such, the syntax for wav conversion is the same as for BGM and SE WAF files:
./e17p.py ever17.ff.waf foo.waf -w

SCR (saver.dat)

Windows screen savers. These are stored obfuscated; the unobfuscation algorithm has not been reimplemented yet.

JPG (wallpaper.dat)

Jpeg image files, stored for wallpaper export. These are stored obfuscated, using essentially the same obfuscation algorithm as for SCR files.

WAV (sysvoice.dat)

Wave audio files, stored for system voice set export. These are stored obfuscated, using essentially the same obfuscation algorithm as for SCR files.

Formats and parsers (Remember11)

AFS files

Are top-level aggregates of smaller files. Aside from file contents, they also store names and various metadata of unknown meaning. This file format includes an RLE compression algorithm that works on a per-file basis, and a single AFS archive can contain both compressed and uncompressed files in arbitrary combination.

BIP files (mac.afs)

Event scripts, possibly mixed with other data. There's an experimental and incomplete tokenizer for the script bytecode; recommended usage:
./e17p.py remember11.ff.event_script -qrtu --afs --cont mac.afs

T2P files (chr.afs, ev.afs)

Image files that come in two flavors: BGRA bitmap images, and wrapped PNG files. The T2P header is 64 bytes; wrapped PNG get an additional inner header that always starts with the bytes 'PNGFILE', and is another 124 bytes in length. A converter script from both types to PNG is available; usage example:
./e17p.py remember11.ff.t2p -o outdir --png --afs chr.afs