/SimsReiaParser

Libraries to parse and (maybe) encode .reia file for sims2

Primary LanguagePythonMIT LicenseMIT

SimsReiaParser

Libraries to parse and (maybe) encode .reia file for sims2

Belladonna.Cove.mp4

Format Details

.reia files are how Sims2 stores the videos for its neighborhood previews in its main menu. They seem to be a relatively custom format, this is an overview on how it encodes video data:

Header

.reia files first have a header with metadata, the fields are:

Offsets Type Description (values in hex)
0x00 - 0x03 Magic 52 49 46 46 "RIFF"
0x04 - 0x07 uint32 Size of the "chunk" which is the rest of the file
0x08 - 0x0f Magic 52 65 69 61 68 65 61 64 "Reiahead"
0x10 - 0x13 uint32 18 Size of the metadata to follow (always 24 in practice)
0x14 - 0x17 uint32 01 Unknown, always 1
0x18 - 0x1b uint32 Horizontal resolution in pixels
0x1c - 0x1f uint32 Vertical resolution in pixels
0x20 - 0x23 uint32 Frames per second numerator (see below)
0x24 - 0x27 uint32 FPS denominator. FPS given by fps_numerator/fps_denominator
0x28 - 0x2b uint32 Number of frames in the file
... frames See below

Frames

After the header, there are a sequence of frames. Each frame is

Offsets Type Description (values in hex)
0x00 - 0x03 Magic 66 72 6d 65 "frme"
0x04 - 0x07 uint32 Size of the frame data to follow
... Frame Data See below

Note that the frame data is padded to be 2-byte aligned. So if the frame size was 3, there will be 4 bytes of data instead of 3 before the next frme magic bytes.

Frame Data - Run Length Encoding

The format uses some interesting tricks for compression: run-length encoding, encoding differences in pixel values between frames and reusing blocks of pixels that stay the same.

Each pixel has a standard 24-bit color depth storing 8-bits for the RGB channels in little endian.

Images are stored in blocks of 32x32 pixels. So a reia file with a resolution of 128x128 is constructed of 4x4 = 16 blocks. Each block is parsed as follows:

  • A single byte is read. 0 indicates this 32x32 block should be reused from the previous frame. Any other value means the block data is to follow.

  • A byte is read and treated as a signed 8-bit integer n:

    • If the value is negative, then a single pixel value should be read and repeated (-n) + 1 times.
    • If the value is positive, then n + 1 unique pixel values should be read.

  • The pixel values retrieved are relative to the previous frame unless it is the first frame of the file. So a change from a red value of 200 to 205 is encoded as 5. A change from 255 to 0 is encoded as 1.

    In order to get the actual pixel value you must use (previous + current) & 0xFF.