Overview

Short for libMessageStudio, LMS is a proprietary tool utilized by Nintendo developers to store localized text. The following page describes the general structure that all LMS file formats follow.

Format List

• MSBT
• MSBP
• MSBF

File Header

Offset	Type	Description
0x0	MAGIC	Magic number
0x8	u16	BOM
0xA	u16	Unused
0xC	u8	Message encoding (0=UTF-8, 1=UTF-16, 2=UTF-32)
0xD	u8	Version number
0xE	u16	Number of blocks
0x10	u16	Unused
0x12	u32	Filesize
0x16	padding [10]	Padding

Section Header

Offset	Type	Description
0x0	MAGIC	Block Magic
0x4	u32	Block size (excludes header)
0x8	padding [8]	Padding
0x10	-	Block data

Labels

Labels in LMS file formats are looked up with a hash table. The amount of slots are stay consistent with the format itself no matter the amount of items there may be.

Format	Slot Count
MSBT	101
MSBP	29
MSBF	59

The following hash algorithm is used:

def calculate_hash(label: str, slot_count: int):
    hash = 0
    for character in label:
        hash = hash * 0x492 + ord(character)
    return (hash & 0xFFFFFFFF) % slot_count

Section Format

Offset	Type	Description
0x0	u32	Slot count
0x4	Hash slots [slotCount]	Array of hash slots
	Labels	Label strings

Hash Slot

Offset	Type	Description
0x0	u32	Number of labels
0x4	u32	Offset to labels

Label

Offset	Type	Description
0x0	u8	Label length
0x1	labelLength	Label string
	u32	Item index