Raw Audio (*.DGT)

.DGT files are the raw digital audio assets used in Darklands, most visibly in the intro sequence where they accompany .PAN visual assets. The format is about as minimal as digital audio gets: no header, no metadata, just a stream of unsigned 8-bit PCM samples.

Extension

*.DGT

Location

Game root

Byte order

— (byte stream; no multi-byte values)

Size

Variable — entire file is sample data

Compression

None

Magic

None — no header of any kind

Status

Partial

Source

Experimental — waveform and spectrogram analysis of OPENDARK.DGT

Source: experimental validation via waveform and spectrogram analysis of OPENDARK.DGT.

Format Specification

Binary Layout

Offset    Size    Description
0x0000    N       Raw PCM sample data (one byte per sample)

The entire file is sample data. There are no leading or trailing structures.

Playback Parameters

Because the format carries no metadata, all playback parameters are supplied externally by the engine.

Notes:

• 11025 Hz is plausible but produces slightly accelerated, higher-pitched audio.
• The exact sample rate is likely hardcoded or stored in the engine’s sound system configuration — determining this requires tracing the sound driver in DARKLAND.EXE.

Conversion

ffmpeg

ffmpeg -f u8 -ar 8000 -ac 1 -i OPENDARK.DGT OPENDARK.wav

Python

import numpy as np
import wave

data = np.fromfile("OPENDARK.DGT", dtype=np.uint8)

with wave.open("OPENDARK.wav", "wb") as f:
    f.setnchannels(1)
    f.setsampwidth(1)   # 8-bit = 1 byte per sample
    f.setframerate(8000)
    f.writeframes(data.tobytes())

Engine Integration

.DGT files are part of the presentation system, paired with .PAN image sequences in the intro and likely other cutscene-style sequences:

Load .PAN  → visual frame sequence
Load .DGT  → audio stream
Play both in sync via presentation loop

The engine almost certainly:

1. Loads the .DGT file into a memory buffer
2. Configures the audio device (sample rate, output channel)
3. Streams the raw buffer to the sound hardware

Given the era, the likely hardware path is Sound Blaster DAC output via DMA-based streaming or timer-driven buffer feeding (port 0x22C). This is consistent with the 8000 Hz sample rate, which is a typical Sound Blaster DSP rate.

Reimplementation

Decoding is trivial — no parsing required:

byte[] raw = File.ReadAllBytes("OPENDARK.DGT");

// Convert to normalised float samples for a modern audio API
float[] samples = new float[raw.Length];
for (int i = 0; i < raw.Length; i++)
    samples[i] = (raw[i] - 128) / 128f;

All complexity is in the engine’s playback layer (sample rate selection, hardware abstraction), not in the format itself.

Known Files

Open Questions

• What is the exact sample rate used by the engine — 8000 Hz, 11025 Hz, or something else?
• Is the sample rate fixed globally or configurable per-file?
• Is playback blocking, interrupt-driven, or DMA-streamed?

Answering these requires tracing the sound subsystem in DARKLAND.EXE. The audio driver layer is a key upcoming RE target.

Status