GTA V Radio Dumps

This repository contains raw data dumps, scripts and processed JSON files related to GTA V's radio stations.

📦 Repository Structure

• Root Directory: Contains all Python scripts for parsing, hash reversing, and XML processing. info_merged.json includes all data from /processed merged into a single file.
• /raw: Holds raw XML data parsed from audio metadata files (dat151.rel, dat54.rel, dat4.rel) and AWC files (in /raw/tracks) using CodeWalker. Nametables and global text tables are also extracted with CodeWalker, but some missing nametables are sourced from Monkys-Audio-Research (game.dat151.nametable, sounds.dat54.nametable)
• /processed: Final JSON output for each DLC, ready for use

📁 JSON Output Format

Each JSON file in /processed follows this structure:

Radio Station

• FlagsValue - Hex string containing unprocessed flags
• Flags - List of enabled flag names
• Genre
• AmbientRadioVol
• RadioName – Alternative to index
• TrackLists – Collection of tracklists tied to this station
• Speech - (optional) Dictionary of DJ Speech Context, indexed first by category and then context

TrackList

• DlcPath - (merged JSON only) Absolute path to the DLC's audio folder
• FlagsValue – Hex string that usually doesn't contain any meaningful flags
• Category – Usage context (e.g., music, id, mono_solo, ad)
• Tracks – Array of tracks included

Speech Context

• Variations - Number of available variations
• ContainerPath - Absolute path to the speech container

Track

• Id
• DlcPath - (optional) Absolute DLC audio path, if the track is from a different folder
• TrackList - (optional) Indicates which tracklist this audio is duplicated from
• Path – Path relative to the DLC folder
• Duration – Length of the track (in ms)
• Intro - (optional) Describes exclusive intro speeches
• Markers - Dictionary of:
  • Track Markers
  • Beat Markers
  • DJ Markers
  • Rockout Markers

Intro

• Variations - Number of exclusive intro variations
• ContainerPath - Relative path to the intro container within the DLC

Track Marker

• Offset – Timestamp of the marker (in ms)
• Id - Internal ID for track title/artist lookup from text table
• Title – Track title (for talkshows might be Artist instead)
• Artist – Track artist

Beat Marker

• Offset – Timestamp of the beat (in ms)
• Value – Beat type

DJ Marker

• Offset - Timestamp of the marker (in ms)
• Value - String that describes if an intro or outro is starting or ending ("{intro/outro}_{start/end}")

Rockout Marker

• Offset - Timestamp of the marker (in ms)
• Value - String "start" or "end"

📜 Scripts

Requirements:

• Python >= 3.10
• lxml:
  Install with pip:

  pip install lxml

Usage:

• main.py - The main entry point. This script processes radio data from the /raw directory.
  
  At the bottom of main.py, you'll find this line:

  export_dlc_radio_info(all_stations, dlc)

  • First argument: the set of stations to process (defaults to all)
  • Second argument: the DLC name to process:

  DLC Tag	Files Processed
  "" (empty)	game.dat151.rel.xml; sounds.dat54.rel.xml; speech.dat4.rel.xml
  "dlchei4"	dlchei4_game.dat151.rel.xml; dlchei4_sounds.dat54.rel.xml; dlchei4_speech.dat4.rel.xml

• xml_utils.py - Utilities for handling XML

  • TypeIndex - Caches lookups for XML elements to speed up parsing
  • to_dict - Recursively converts an XML element to Python dictionary
  • marker_dict_awc() / marker_dict_xml() - Converts marker containers into readable dictionaries

• hash_utils.py - Utilities for hash operations

  • HashMap - Loads .nametable and .gxt2 files into a hash lookup table
  • gxt2_binary - Parses .gxt2 binary files (global text table) and turns them into hash maps
  • joaat() - Hashes strings using JOAAT (case-insensitive)
  • format_hash() / parse_hash_string() - Converts hashes to/from string representations (1048674328 <=> "hash_3E818018")

📚 Related Projects

• GTA V Radio implementation repository - HintSystem/GTA-V-Radio