New audio system (OpenAL replacement)

IMPORTANT

Status: WIP
Minimal version: 2.0

Overview

The system replaces the legacy OpenAL / OpenAL Soft audio stack with a modern pipeline: better spatialization, improved mixing, and richer processing without third-party sound libs. You can author custom filters and precise audio manipulations.

Reverb is powered by Resonance Audio.

Dev: SteamAudio (Phonon)

An initial SteamAudio (Phonon) implementation is present. To try it instead of Resonance Audio, remove #define DISABLE_STEAM_AUDIO.

Implemented:

• HRTF
• Algorithmic reverb (no geometry)

Key changes

---

1. Dynamic Range Compression

Before:

• No built-in dynamic range control
• Rendered in u8 (0–256) [1 byte]

Now:

• Optional dynamic compression on the final mix
• Processing and rendering in float (0.0–10.0) [4 bytes]
• User-adjustable compression strength

Benefits:

• Better audibility of quiet sounds
• Protects from clipping on loud peaks
• More even loudness balance

---

2. Audio quality and content requirements

Before:

• OpenAL Soft smoothed many artifacts automatically

Now:

• Manual preparation of audio is required
• IX-Ray SDK tools do not fix defects

.wav requirements:

• Sample rate: 44100 Hz
• Clean start/end (no noise)
• Fade-in / fade-out recommended
• Prefer pre-processing (Reaper, Audition, RX)

---

3. Sample rate

• Audio engine runs at a fixed 44100 Hz
• Other sample rates are unsupported

---

4. Concurrent sound limits

Before:

• 256 slot limit; overflow dropped sounds

Now:

• Up to 4096 sounds simultaneously
• Enables denser, richer soundscapes

---

5. Reverb system

EFX (legacy)

• Type: Preset-Based Auxiliary Reverb
• Model: Standard OpenAL presets (room, tunnel, hall)
• Limitations:
  • Limited physical accuracy
  • Minimal level adaptation
  • Reverb applied even outside zone bounds

---

Resonance Audio (new)

• Type:
  Parametric Algorithmic Reverb

• Approach:

  • Computes reverb params (RT60, decay, frequency response) from:
    • area size
    • materials
    • source/listener positions
  • Tails are generated algorithmically (no convolution, no FDN)
  • Early reflections are approximated with a simplified room model

• Benefits:

  • Context-aware sound
  • Auto-adapts to environment
  • Easy integration, minimal tuning
  • Sounds outside the reverb zone play dry (no sampling)

---

Reverb comparison

Parameter	EFX (OpenAL)	Resonance Audio
Type	Preset algorithmic	Parametric algorithmic
Early reflections	Presets	Geometry-based modeling
Late reverb	Algorithmic	Algorithmic, auto-tuned
Realism	Medium	High (environment-adaptive)
Adaptiveness	Low	High
Tuning	Manual	Nearly automatic

---

6. Sound Debug Info

Tools to visualize scene audio:

• Sound slots
• Active sound files
• Reverb zones on the level

Tips for modders

1. Audio prep

• Use clean .wav 44100 Hz
• Apply fade-in/fade-out to all files
• Remove clicks, chopped tails, noise
• Verify in-game or via Play In Editor in IX-Ray SDK

2. Dynamic compression awareness

• Check loudness at different DRC levels
• Avoid cranking everything to max
• Balance sources manually

3. Working with reverb (Resonance Audio)

• Assign correct surface materials (good practice in X-Ray SDK overall)
• Configure base presets in Shader Editor
• Pick presets matching expected zone intensity (test with Play In Editor)

---

Technical notes

• Vorbis is used for storage.
• Conversion is performed via IX-Ray SDK tools.