MDSSharp 1.2.1

MDSSharp

MDSSharp is a managed C# library for reading Alcohol 120% disc images (.mds descriptor + .mdf data) and Daemon Tools MDS v2 / MDX descriptors. It parses the MDS session and track tables, locates and reassembles the data file (including .i00/.i01 and .001/.002 split sets), strips CD subchannel tails from oversized sectors, and writes CUE sheets so the image can be converted by chdman or other tools.

The library is the Alcohol 120% conversion engine used by CHD Studio, where it prepares .mds images for conversion to CHD with chdman or CHDSharp.

• Repository: https://github.com/purelogiccode/CHDStudio
• Package: https://www.nuget.org/packages/MDSSharp

Table of contents

• Features
• Requirements
• Installation
• Quick start
• Usage examples
  • Parse an MDS descriptor
  • Prepare an image for conversion
  • Join a split volume set
  • Strip subchannel data
  • Generate a CUE sheet
  • Inspect tracks
• Error handling
• API reference
  • MdsParser
  • MdsInputPreparer
  • SplitImageJoiner
  • Models
• Supported images
• How it works
• Building from source
• Version history
• License

Features

• MDS descriptor parsing — reads the medium type, session and track tables of Alcohol 120% descriptors, including track number, mode, sector size, start LBA, subchannel mode and ADR/CTL.
• MDS v2 / Daemon Tools descriptors — decrypts the AES-256 + zlib-obfuscated descriptor (RIPEMD-160 PBKDF2 key derivation, no external crypto dependency), parses the wider v2 session/track/footer blocks and resolves their UTF-16 data file names. Compressed track data (per-track compression tables with stored/RLE/deflate groups) and encrypted track data (AES-256 LRW with GF(2^128) tweaks, including TAGES-style password-less images) are decoded automatically, and single-file .mdx containers are read end to end.
• Data file resolution — uses the file names embedded in the track footers (single-byte or UTF-16), falling back to the matching .mdf, a split .i00 first volume, or an unambiguously named file one folder down.
• Multi-file and split sets — joins descriptors that name several data files, plus .i00/.i01 (Alcohol) and .001/.002 (byte-splitter) sets, tolerating case differences and archive-looking extensions.
• Subchannel stripping — rewrites 2448-byte (2352 data + 96 subchannel) and 2368-byte (2352 + 16) images down to the 2352-byte sectors chdman reads.
• Pregap handling — when the descriptor records pregaps that the data file does not contain, rebuilds them as zeros so the CUE can place INDEX 00 without shifting later tracks.
• CUE sheet generation — writes a single-file CUE with TRACK/INDEX 00/INDEX 01 entries, formatted as MM:SS:FF and encoded UTF-8 without a BOM for chdman.
• Medium-aware output — DVD media goes to the encoder as a cooked image; CD media with 2048-byte sectors gets a MODE1/2048 cue; Mode 2 sectors map by the low nibble of the mode byte, as libmirage's reverse engineering established.
• No native dependencies — pure managed code, cross-platform.

Requirements

The package contains a separate assembly for each target framework, so the correct build is selected automatically by NuGet.

Read-only: MDSSharp reads Alcohol 120% images and produces CUE/BIN output. It cannot create or modify .mds/.mdf files, and it does not produce ISO images.

Installation

dotnet add package MDSSharp

or with the Package Manager console:

Install-Package MDSSharp

Quick start

An Alcohol image is a .mds descriptor next to its .mdf data file (or split .i00/.001 volumes). Point the library at the .mds; the data file is resolved automatically.

using MDSSharp;

var disc = MdsParser.Parse(@"C:\Games\Game.mds");
Console.WriteLine(disc.Summary);

var prepared = await MdsInputPreparer.PrepareAsync(
    disc,
    workDir: @"C:\Temp\mds-work",
    onLog: Console.WriteLine,
    token: CancellationToken.None
);

if (!prepared.Success)
{
    Console.Error.WriteLine(prepared.FailureReason);
    return;
}

// Hand one of these to chdman (createcd for the cue, createdvd for the image).
Console.WriteLine(prepared.CuePath ?? prepared.DvdImagePath);

Usage examples

Parse an MDS descriptor

MdsParser.Parse reads the descriptor without touching the data file, so it is cheap enough for inspection or pre-flight checks.

using MDSSharp;

if (!MdsParser.IsMdsFile(mdsPath))
{
    Console.Error.WriteLine("Not an Alcohol MDS descriptor (missing \"MEDIA DESCRIPTOR\" signature).");
    return;
}

var disc = MdsParser.Parse(mdsPath);

Console.WriteLine($"Sessions: {disc.SessionCount}");
Console.WriteLine($"Data:     {disc.MdfPath ?? "missing"}");
Console.WriteLine(disc.Summary);

IsMdsFile only checks the MEDIA DESCRIPTOR signature and never throws, while Parse throws when the file is missing or not a usable descriptor.

Prepare an image for conversion

MdsInputPreparer.PrepareAsync performs the whole pipeline: resolves the data file, joins split volumes into the work directory, strips subchannel data when present, and writes a CUE sheet. The returned Result tells you exactly what to convert.

using MDSSharp;

var disc = MdsParser.Parse(mdsPath);

var result = await MdsInputPreparer.PrepareAsync(
    disc,
    workDir,
    message => Console.WriteLine(message),
    cancellationToken
);

if (result.Success && result.CuePath is not null)
{
    // 2352-byte raw CD (with or without stripped subchannel): convert as a CD.
    RunChdman("createcd", result.CuePath);
}
else if (result.Success && result.DvdImagePath is not null)
{
    // 2048-byte cooked sectors: the .mdf is a DVD image.
    RunChdman("createdvd", result.DvdImagePath);
}
else
{
    Console.Error.WriteLine(result.FailureReason);
}

Join a split volume set

SplitImageJoiner detects and concatenates numbered pieces, whether the set uses Alcohol's .i00 style, the .001 style, or plain names. A set is only accepted when the next piece actually exists, so a lone file ending in .001 is left alone.

using MDSSharp;

var parts = SplitImageJoiner.TryGetVolumeSet(@"C:\Games\Game.i00");
if (parts is not null)
{
    Console.WriteLine($"{parts.Count} parts, {SplitImageJoiner.GetTotalBytes(parts):N0} bytes");
    var written = await SplitImageJoiner.JoinAsync(parts, @"C:\Temp\Game.bin", cancellationToken);
    Console.WriteLine($"Joined {written:N0} bytes");
}

Strip subchannel data

Alcohol often rips 2448-byte sectors (2352 data + 96 bytes of subchannel). chdman does not read them, so StripSubchannelAsync keeps only the first 2352 bytes of every sector. It returns null on success or a user-facing reason on failure.

using MDSSharp;

var failure = await MdsInputPreparer.StripSubchannelAsync(
    @"C:\Games\Game.mdf",
    @"C:\Temp\Game.bin",
    sectorSize: MdsDisc.RawPlusSubchannelSize,
    cancellationToken
);

Console.WriteLine(failure is null ? "Stripped." : failure);

Generate a CUE sheet

WriteCueAsync writes a single-file CUE with one TRACK/INDEX 01 entry per track. FormatMsf converts an absolute LBA to the MM:SS:FF a CUE expects.

using MDSSharp;

var disc = MdsParser.Parse(mdsPath);
var cuePath = await MdsInputPreparer.WriteCueAsync(disc, workDir, "Game.bin", cancellationToken);

Console.WriteLine($"Wrote {cuePath}");
foreach (var track in disc.Tracks)
{
    Console.WriteLine($"  TRACK {track.Number:00} {track.CueTrackType} @ {MdsInputPreparer.FormatMsf(track.StartLba)}");
}

Inspect tracks

Each MdsTrack exposes its mode byte, sector size and start LBA, plus a CUE track type when the mode can be expressed in a CUE.

using MDSSharp;

var disc = MdsParser.Parse(mdsPath);

foreach (var track in disc.Tracks)
{
    Console.WriteLine($"Track {track.Number:00}: {track.CueTrackType ?? "unrepresentable"}");
    Console.WriteLine($"  Audio:  {track.IsAudio}");
    Console.WriteLine($"  Sector: {track.SectorSize} bytes");
    Console.WriteLine($"  Start:  LBA {track.StartLba} ({MdsInputPreparer.FormatMsf(track.StartLba)})");
}

Error handling

PrepareAsync and StripSubchannelAsync otherwise report failures through their result (Result.FailureReason, null/message) rather than exceptions, so batch processing can continue past a bad image. MDS v2/MDX images with compressed or password-less encrypted track data are decoded automatically; images whose track data needs a user password fail preparation unless the password is supplied as the optional PrepareAsync(..., password:) argument.

API reference

All types live in the MDSSharp namespace.

MdsParser

The descriptor parser.

MdsInputPreparer

Turns a parsed image into something chdman can read.

SplitImageJoiner

Models

Supported images

• Descriptors — MEDIA DESCRIPTOR signature, medium type, session table and track table, plus the per-track extra blocks (pregap and length) and footer blocks (data file names). Up to 99 sessions; lead-in and lead-out entries are skipped.
• MDS v2 / MDX descriptors — the Daemon Tools variant (MEDIA DESCRIPTOR + version 2). The descriptor is decrypted and decompressed transparently, and its 32-byte session, 80-byte track and 32-byte footer blocks are parsed, including UTF-16 file names. Compressed tracks are decoded from their compression tables (stored, run-length and deflate groups); encrypted tracks are decrypted with AES-256 LRW (password-less/TAGES or a supplied password); single-file .mdx containers keep their descriptor and data in one file, which is read directly.
• Per-track pregaps — each footer's track_data_length is the sectors actually stored, so StoredDataSectors tells whether a track's pregap is in the data file (LengthSectors = missing, LengthSectors + PregapSectors = stored, the layout libMirage validates). Preparation materializes only the missing, representable pregaps as zeros, so a v2 image that stores one track's pregap and omits another's still gets every INDEX 00/INDEX 01 at the right LBA. The first track's pregap before LBA 0 is never materialized.
• Medium types — CD, CD-R, CD-RW (0x00–0x02) and DVD, DVD-R (0x10, 0x12); other values fall back to inspecting the sector sizes.
• Track modes — audio, Mode 1 and the Mode 2 forms, selected by the low nibble of the mode byte and mapped to AUDIO, MODE1/2352/MODE1/2048 and MODE2/2352/MODE2/2336.
• Data files — one or more files named by the descriptor (single-byte or UTF-16 names, *.mdf wildcards), split .i00/.i01 sets, split .001/.002 sets.
• Sector sizes — 2352 (raw CD), 2448 (raw + 96 subchannel), 2368 (raw + 16 subchannel), 2336 (cooked Mode 2) and 2048 (cooked CD/DVD data).
• Pregaps — INDEX 00 is written when the data file carries the pregap sectors; descriptors whose pregaps are absent have them rebuilt as zeros first.

How it works

An Alcohol image is a binary descriptor plus at least one data file. MDSSharp parses the descriptor's medium type, session blocks (24 bytes each) and track blocks (80 bytes each), taking the mode, POINT (track number), sector size and start LBA per track, and reads the per-track extra block (pregap and length) and footer blocks (data file names). The data files are resolved from those names, or by base name in the descriptor's folder or exactly one subfolder when the names cannot be found.

The medium type and the stored sector layout determine what has to happen before conversion:

When the descriptor records pregaps and the track lengths add up to the file size, the data file does not contain the pregap sectors. Those descriptors are rebuilt into a .pregap.bin with zero-filled pregaps (stripping subchannel data in the same pass when present) so the cue's INDEX 00 points at real sectors. If the pregaps are already in the data file, INDEX 00 points at them directly and nothing is copied. For MDS v2 the decision is made per track from the footer's StoredDataSectors, so an image that keeps one track's pregap and omits another's is rebuilt selectively; a pregap before LBA 0 (the first track's) is never materialized.

CUE index values are derived from the descriptor's start LBA with FormatMsf (75 frames per second, 60 seconds per minute) and written as UTF-8 without a BOM.

Building from source

The library lives in the CHD Studio repository under MDSSharp/.

git clone https://github.com/purelogiccode/CHDStudio.git
cd CHDStudio
dotnet build MDSSharp/MDSSharp.csproj -c Release
dotnet pack MDSSharp/MDSSharp.csproj -c Release -o artifacts

The repository's xUnit suite lives in CHDStudio.Tests/:

dotnet test CHDStudio.Tests/CHDStudio.Tests.csproj -c Release

Version history

1.2.1

• Package metadata and README now reference the renamed CHD Studio repository (https://github.com/purelogiccode/CHDStudio, previously BatchConvertToCHD); no functional changes.

1.2.0

• MDS v2 and single-file .mdx support: the descriptor is decrypted and decompressed, compressed tracks are decoded from their compression tables and encrypted tracks use AES-256 LRW (password-less/TAGES or a supplied password).
• Per-track pregaps: StoredDataSectors exposes each footer's stored length, and preparation materializes only the missing, representable pregaps so mixed v2 layouts keep every INDEX 00/INDEX 01 at the right LBA.
• A footer count of zero still reads footer 0 like mdsx, and a later fragment of a split track that names no data file is reported instead of re-reading the first file.
• MdsDisc gained HasEncryptedTrackData, HasCompressedTrackData and IsMdxContainer.

1.1.0

• Renamed from Alcohol120Sharp; package metadata, README and icon added.
• Pregaps are described (INDEX 00, rebuilt when absent) and multi-file descriptors are joined in track order.
• Medium type, per-track extra blocks and footer file names are parsed, with the mode taken from the low nibble of the mode byte.

License

MDSSharp is released under the MIT license.