Audio

Sound in Tribes 2 is two datablocks and a handful of console functions. AudioProfile is the most-declared datablock in the game — 298 of them [script] — so this is a system you will touch constantly.

The two datablocks

flowchart LR
    AD["AudioDescription<br/>the category:<br/>3D? looping? which channel?<br/>falloff distances"] --> AP["AudioProfile<br/>a specific sound:<br/>filename + description + effect"]
    EP["EffectProfile<br/>force feedback"] --> AP

AudioProfile — a sound

datablock AudioProfile(DiscFireSound)
{
   filename    = "fx/weapons/spinfusor_fire.wav";
   description = AudioDefault3d;        // ← the AudioDescription
   preload     = true;
   effect      = DiscFireEffect;        // ← optional force-feedback EffectProfile
};

preload = true on anything that must not stutter the first time it plays — weapon fire, impacts, UI clicks. The shipped profiles use it selectively; looping ambient sounds mostly do not.

AudioDescription — the category

Thirteen ship [script]. Three examples:

datablock AudioDescription(AudioClosest3d)
{
   volume     = 1.0;
   isLooping  = false;
   is3D       = true;
   minDistance = 5.0;
   MaxDistance = 30.0;
   type       = $EffectAudioType;
   environmentLevel = 1.0;
};

datablock AudioDescription(AudioDefault3d)
{
   volume     = 1.0;
   isLooping  = false;
   is3D       = true;
   minDistance = 20.0;
   MaxDistance = 100.0;
   type       = $EffectAudioType;
   environmentLevel = 1.0;
};

datablock AudioDescription(ProjectileLooping3d)
{
   volume     = 1.0;
   isLooping  = true;
   is3D       = true;
   minDistance = 5.0;
   MaxDistance = 20.0;
   type       = $EffectAudioType;
   environmentLevel = 1.0;
};

Note the capital M in MaxDistance against lowercase minDistance. That is how the shipped datablocks spell it, consistently. Match it.

The shipped descriptions

Pick the existing description that matches your intent rather than declaring a new one. The distances were tuned together and mixing in an oddly-scaled sound is immediately noticeable.

Channels

$…AudioType globals name the mixer channels, each bound to a preference [script]:

alxSetChannelVolume( $EffectAudioType, $pref::Audio::effectsVolume );
alxSetChannelVolume( $VoiceAudioType,  $pref::Audio::voiceVolume );
alxSetChannelVolume( $ChatAudioType,   $pref::Audio::radioVolume );
alxSetChannelVolume( $MusicAudioType,  $pref::Audio::musicVolume );
alxSetChannelVolume( $GuiAudioType,    $pref::Audio::guiVolume );
alxSetChannelVolume( $RadioAudioType,  $pref::Audio::radioVolume );

Putting a sound on the right channel means the player’s volume sliders work as expected. A gameplay sound on $GuiAudioType cannot be turned down independently.

Playing sounds

alx is the OpenAL-style naming — Tribes 2 wraps Miles Sound System behind an OpenAL-shaped API.

Object audio slots

Slots are numbered and long-lived, which is how looping sounds are managed. The ELF gun reserves two [script]:

$ELFZapSound  = 2;
$ELFFireSound = 3;
…
%targeter.playAudio($ELFFireSound, ELFGunFireSound);
if(!%target.zapSound)
{
   %target.playAudio($ELFZapSound, ELFHitTargetSound);
   %target.zapSound = true;
}
…
%target.stopAudio($ELFZapSound);
%targeter.stopAudio($ELFFireSound);

Note the boolean guard — playAudio on a slot that is already playing restarts it, so a per-frame call without a guard produces a stutter rather than a loop. Define your slot numbers as globals, as Sierra did, so two systems do not collide on a slot.

Positional one-shots

serverPlay3D(GrenadeThrowSound, %pos);

Server-side, no object required, replicated. This is the right call for transient sounds at a location.

Sound in chat messages

defaultMessageCallback plays a WAV embedded in a message via the ~w marker [script]:

$MaxMessageWavLength = 5200;

%wavStart = strstr( %message, "~w" );
if ( %wavStart != -1 )
{
   %wav = getSubStr( %message, %wavStart + 2, 1000 );
   %wavLengthMS = alxGetWaveLen( %wav );
   if ( %wavLengthMS <= $MaxMessageWavLength )
   {
      %handle = alxCreateSource( AudioChat, %wav );
      alxPlay( %handle );
   }
   else
      error( "WAV file \"" @ %wav @ "\" is too long! **" );

   %message = getSubStr( %message, 0, %wavStart );
   if ( %message !$= "" )
      addMessageHudLine( %message );
}

So a server message of the form '\c2Enemy spotted!~wfx/voice/spotted.wav' prints text and plays a sound. This is how voice binds work, and the 5.2-second cap is the anti-spam guard.

Deployables use it for the error beep [script]:

%errorSnd = '~wfx/misc/misc.error.wav';

Force feedback — EffectProfile

140 declared [script]. They drive Immersion TouchSense hardware, which almost nobody has today.

datablock EffectProfile(DiscFireEffect)
{
   effectname = "weapons/spinfusor_fire";
   minDistance = 2.5;
   maxDistance = 2.5;
};

effectname refers to an .ifr (Immersion Force-feedback Resource) in base.vl2’s effects/ directory. You do not need to declare these — a new AudioProfile with no effect field works fine. The shipped weapon files declare a full set purely for completeness.

Recipe: sounds for a new weapon

Follow the shipped order — effects, then sounds, then everything that references them:

//------------------------------------------------------------------------------
// MyMod — Burst Spinfusor sounds
//------------------------------------------------------------------------------

datablock AudioProfile(BurstDiscFireSound)
{
   filename    = "fx/weapons/spinfusor_fire.wav";     // reuse a stock file
   description = AudioDefault3d;
   preload     = true;
};

datablock AudioProfile(BurstDiscReloadSound)
{
   filename    = "fx/weapons/spinfusor_reload.wav";
   description = AudioClosest3d;
   preload     = true;
};

datablock AudioProfile(BurstDiscLoopSound)
{
   filename    = "fx/weapons/spinfusor_idle.wav";
   description = ClosestLooping3d;
};

datablock AudioProfile(BurstDiscDryFireSound)
{
   filename    = "fx/weapons/spinfusor_dryfire.wav";
   description = AudioClose3d;
   preload     = true;
};

then reference them from the image’s state machine:

   stateSound[2] = BurstDiscLoopSound;      // Ready
   stateSound[3] = BurstDiscFireSound;      // Fire
   stateSound[4] = BurstDiscReloadSound;    // Reload
   stateSound[6] = BurstDiscDryFireSound;   // DryFire

Shipping your own .wav files

Place them at MyMod/audio/<yourpath>/<name>.wav and reference them as "<yourpath>/<name>.wav" — the audio/ category prefix is implied. See Mod paths and overrides.

Because audio files are referenced by a datablock rather than carried in it, a client without your .wav gets silence, not an error. New audio requires a client-side download; reusing stock sounds does not. See Datablocks.

Reverb environments

AudioEnvironment datablocks (3 shipped [script]) define reverb spaces, applied by the mission and scaled per-sound by environmentLevel on the AudioDescription. Rarely worth authoring for a gameplay mod; relevant if you are building interiors.

Under the community patches

This is the most-changed subsystem in section 03 — though almost none of it affects how you author audio.

OpenAL Soft joins Miles

The QoL patch replaces Mss32.dll with a proxy that routes to OpenAL Soft (soft_oal.dll), and offers both backends [patch-script]:

$pref::Audio::drivers = "Miles\tOpenAL";

The reason is practical: Miles’ DirectSound path degraded on Windows Vista and broke harder on Windows 10. OpenAL Soft is what makes audio work on a modern machine.

Your AudioProfile and AudioDescription datablocks need no changes. The alx* API is the same — the naming was always OpenAL-shaped, and the patch simply makes it OpenAL underneath.

Driver switching is handled by an override [patch-script]: audioSetDriver("none"), then audioSetDriver(newDriver), then the shell hum is re-armed via alxPlay(). Environmental audio is gated on alxIsExtensionPresent(EAX), which now reports the OpenAL backend’s capability rather than the hardware’s.

New alx* functions registered by IFC22.dll [binary]:

alxGetContexti(ALC_PROVIDER_COUNT | ALC_PROVIDER | ALC_SPEAKER_COUNT | ALC_SPEAKER)
alxGetContextstr(ALC_PROVIDER_NAME | ALC_SPEAKER_NAME, idx)
alxIsExtensionPresent(name)
alxEnableEnvironmental(bool)

Music gains FLAC

MP3Audio::playTrack is overridden to play FLAC instead of MP3 when the OpenAL provider is active and a matching FLAC file exists [patch-script]. getRandomTrack still enumerates music/*.mp3 for selection, so a mod shipping music should ship MP3 and may optionally ship FLAC alongside.

Force feedback is dead

EffectProfile datablocks — 140 of them in the shipped scripts — still parse, and AudioProfile.effect still accepts them. But TribesNEXT’s IFC22.dll stubs the 11 Immersion exports it replaces [binary], so the .ifr effects never play.

This changes nothing about how you write audio. Declaring EffectProfile blocks remains harmless, and omitting them was always fine. RC2a keeps the vendor DLL, so force feedback still works there.

Sound files still need to reach the client

Unchanged, with one caveat: enableAssetDownloads may deliver a missing .wav on connect. Do not rely on it — see Packaging.

Related

• Weapons — stateSound[n] in the image state machine
• Particles, explosions, and effects — soundProfile on ExplosionData
• Text and messaging — the ~w marker
• Armors — the footstep and impact sound tables
• TribesNEXT QoL patch — the audio overrides in full