Mcd.db

From PhatHack Wiki
Revision as of 23:55, 11 March 2007 by Davec (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

PhatDB

Introduction

Phatdb is the name I have given to the mcd.db file format created by PhatNoise Media Manager software to be playable by Kenwood Car Stereo head units supporting ACDrive Technology.

The PhatDB is a database of song metadata. It contains information on all tracks, artists, albums, genres and playlists in a compact format. The concept is similar to the itunesdb database found on ipods, though phatdb is less complex.

This document describes the format of the PhatDB file.

A Free Software program and library to build mcd.db files is available here:

http://phatdb.sourceforge.net/

File Format

All multibyte types are stored in big endian (network) byte order. They must be byte-swapped to be read on little-endian PC platforms.


The overall structure looks like this:

Header
Tracks List
Artist List
Album List
Genre List
Playlists List

Header

The file begins with a header containing the following elements:

Offset Name Type Value Description
0x00000000 magic uint32 0x3af7d87c File Signature
0x00000004 unknown uint32 0x00000002 Version?
0x00000008 codepage uint32 0x000001b6 Codepage 437
0x0000000c unknown uint32 0x00000002 Constant?
0x00000010 n_tracks uint32   Total Number of Tracks
0x00000014 n_artists uint32   Total Number of Artists
0x00000018 n_albums uint32   Total Number of Albums
0x0000001c n_genres uint32   Total Number of Genres
0x00000020 n_playlists uint32   Total Number of Playlists
0x00000024 artists_len uint32   Len of Artists Array
0x00000028 albums_len uint32   Len of Albumns Array
0x0000002c genres_len uint32   Len of Genres Array
0x00000030 playlists_len uint32   Len of Playlist Array
0x00000034 unknown uint32 0x00000000 Constant?

Track List

The File Header is immediately followed by an array of n_tracks track structures (offsets are relative to start of track structure). Tracks are stored in alphabetical order.:

Offset Name Type Value Description
0x00000000 artists_id uint32   Artist ID
0x00000004 album_id uint32   Album ID
0x00000008 genre_id uint32   Genre ID
0x0000000c unknown uint32 0x00000000 Constant?
0x00000010 title phat_str   Track Title
0x0000xxxx path phat_str   Track Path

Where the phat_str structure looks like this:

Offset Name Type Value Description
0x00000000 string_length uint16   Length of String (bytes)
0x00000002 string_width uint16   Width of String?
0x00000004 data string   Character Array

When string_width is 2 (Track Title) data is UCS-2 encoded. When string_width is 1 (Track Path), data is CP437 encoded. data is null-terminated and padded to a multiple of 4 bytes. string_length is the length of data (including padding).

List Sections

Immediately following the track listing, there are sections for artists, albums and genres. All lists are in alphabetical order. Each section has the following format:

Offset Name Type Value Description
0x00000000 item_list phat_list[]   PhatDB Item List
0x00000xxx track_item_arr uint32[]   Track/Item Assignment Array
0x00000xxx item_index uint32[28]   Item Index


The structure of the phat_list is as follows:

Offset Name Type Value Description
0x00000000 array_offset uint32   Item Array Offset
0x00000004 name phat_str   Item Name String

array_offset is an index into track_item_arr. This array contains a mapping of track id's relating to that item (artist, album, genre). The length of the each array is given by the fields in the header artists_len, albums_len and genre_len. item_index seems to be an index for skipping through large arrays, perhaps indexed alphabetically?

Playlists

The playlists section immediately follows the genre section. It is identical to the sections above except that it does not have the item_index.