7.9.4. AUDIO-04 — 3D Spatial Audio

This tutorial covers 3D positional audio: placing the listener and sound sources in space, moving sources in real time, and choosing attenuation models. Best experienced with headphones.

7.9.4.1. Listener Setup

set_head_position(pos, dir, vel) defines the listener in 3D space:

pos — world position (float3)
dir — facing direction (float3, should be normalized)
vel — velocity (float3, used for Doppler calculations)

// Listener at origin, facing forward (+Y), stationary
set_head_position(float3(0, 0, 0), float3(0, 1, 0), float3(0, 0, 0))

The coordinate system is left-handed: +X is right, +Y is forward, +Z is up.

7.9.4.2. Playing 3D Sound

play_3d_sound_loop_from_pcm places a looping sound at a 3D position with a specified attenuation model:

var samples <- generate_click_burst()
let sid = play_3d_sound_loop_from_pcm(
    float3(5, 0, 0),              // position: 5 meters to the right
    linear_attenuation(20.0),     // audible up to 20 meters
    MA_SAMPLE_RATE, 1,
    samples
)

The attenuation model determines how volume falls off with distance (see below). Short impulsive sounds (clicks, noise bursts) are easier to localize than pure tones.

7.9.4.3. Moving a Sound Source

set_position(sid, pos, vel) updates the position and velocity of a 3D sound. Call it in a loop to animate:

let radius = 5.0
let num_steps = 100
for (i in range(num_steps)) {
    let angle = 2.0 * PI * float(i) / float(num_steps)
    let x = radius * cos(angle)
    let y = radius * sin(angle)
    let pos = float3(x, y, 0)

    // Tangential velocity enables Doppler effect
    let vx = -radius * sin(angle) * 2.0 * PI / 5.0
    let vy =  radius * cos(angle) * 2.0 * PI / 5.0
    let vel = float3(vx, vy, 0)

    set_position(sid, pos, vel)
    sleep(50u)
}

The velocity vector does not move the source — it only affects the Doppler pitch shift. You must update pos yourself each frame.

7.9.4.4. Attenuation Models

Four built-in attenuation models control how volume decreases with distance. Each constructor takes a max_distance parameter:

Constructor	Falloff	Character
`inverse_distance_attenuation`	1 / d	Natural, gradual rolloff
`linear_attenuation`	1 - d / max	Straight line to silence at max distance
`quadratic_attenuation`	1 - (d / max)²	Faster than linear near max distance
`inverse_square_attenuation`	1 / d²	Physically accurate, rapid falloff

Example — comparing all four at different distances:

let max_dist = 30.0
let sid_inv  = play_3d_sound_loop_from_pcm(pos,
    inverse_distance_attenuation(max_dist), MA_SAMPLE_RATE, 1, samples_a)
let sid_lin  = play_3d_sound_loop_from_pcm(pos,
    linear_attenuation(max_dist), MA_SAMPLE_RATE, 1, samples_b)
let sid_quad = play_3d_sound_loop_from_pcm(pos,
    quadratic_attenuation(max_dist), MA_SAMPLE_RATE, 1, samples_c)
let sid_isq  = play_3d_sound_loop_from_pcm(pos,
    inverse_square_attenuation(max_dist), MA_SAMPLE_RATE, 1, samples_d)

7.9.4.5. HRTF

HRTF (Head-Related Transfer Function) is enabled by default in the audio engine. It applies binaural filtering so that sounds placed in 3D space are perceived at their correct spatial position when listening through headphones. No additional setup is required — all play_3d_sound_* functions automatically benefit from HRTF processing.

7.9.4.6. Running the Tutorial

daslang.exe tutorials/dasAudio/04_spatial_audio.das

The tutorial places a clicking sound to the right, orbits it around the listener over 5 seconds, then demonstrates the four attenuation models at varying distances. Use headphones for the best spatial experience.