Implemented player combat vocals for Blood Elf and Draenei races:
Player vocal types:
- Attack grunts: Multiple variations per race/gender
- Wound sounds: Pain reactions when hit
- Wound crits: Special sounds for critical hits taken
- Death cries: Final sounds when player dies
Race coverage (60+ voice lines):
- Blood Elf Male: 9 attacks, 8 wounds, 3 crit wounds, 2 deaths
- Blood Elf Female: 5 attacks, 7 wounds, 1 death
- Draenei Male: 7 attacks, 8 wounds, 3 crit wounds, 2 deaths
- Draenei Female: 7 attacks, 4 wounds, 3 crit wounds, 1 death
Technical details:
- Loads 60+ vocal sound files from Sound\Character\*PC folders
- Simple API: playPlayerAttackGrunt(race), playPlayerWound(race, crit), playPlayerDeath(race)
- Random variation selection for immersion
- Volume at 0.9-1.1 depending on sound type
- Crit wounds play 1.1x louder for emphasis
- Extensible design ready for other races
- Only BC races have dedicated PC vocal folders in WotLK 3.3.5a
Usage examples:
```cpp
combatSoundManager->playPlayerAttackGrunt(PlayerRace::BLOOD_ELF_MALE);
combatSoundManager->playPlayerWound(PlayerRace::DRAENEI_FEMALE, true); // Crit
combatSoundManager->playPlayerDeath(PlayerRace::BLOOD_ELF_FEMALE);
```
This adds essential combat immersion with player character reactions!
Changed bell tolls from random intervals to proper timekeeping mechanism:
How it works:
- Bells now toll at the top of every hour (minute == 0)
- Number of tolls indicates the hour in 12-hour format:
* 1 AM/PM = 1 toll
* 2 AM/PM = 2 tolls
* 12 AM/PM = 12 tolls
- 1.5 second delay between individual tolls
- Uses system time (server time for single-player mode)
Technical details:
- Detects hour changes via std::chrono and localtime()
- Tracks lastHourTolled_ to prevent duplicate tolling
- remainingTolls_ counter for sequential toll playback
- bellTollDelay_ for 1.5s spacing between tolls
- Only tolls when currentCity_ is set (in a city)
- Resets bell state when changing cities
- Converts 24-hour to 12-hour format (0 and 12 both become 12)
Example: At 3:00 PM in Stormwind, the Alliance bell will toll 3 times
with 1.5s between each toll, marking the hour like a real clock tower.
This makes bells actually useful for tracking real time while playing!
Implemented fountain sounds as positional audio emitters:
Technical details:
- Added fountainSounds_ library (FountainSmallMediumLoop.wav)
- Fountain type already existed in AmbientType enum but was unused
- 6 second loop interval for fountain sounds
- Volume at 0.8x water volume for gentle bubbling effect
- Max distance: 35 units (same as other water sources)
- Counted in activeWaterCount limit (max 3 water sources at once)
- Spatial 3D audio based on fountain position
- Can be placed via addEmitter(position, AmbientType::FOUNTAIN)
Implemented periodic bell chimes that add atmosphere to major cities:
Bell types by faction:
- Alliance bell: Stormwind, Ironforge
- Night Elf bell: Darnassus
- Horde bell: Orgrimmar, Undercity
- Tribal bell: Thunder Bluff
Technical details:
- Loads 4 bell toll sound files from Sound\Doodad
- Plays every 120-180 seconds (2-3 minutes) with random variation
- First bell toll 60-90 seconds after entering city
- Volume at 0.5 for noticeable but not overpowering effect
- Only plays when currentCity_ is set (not in wilderness)
- Each city uses faction-appropriate bell sound
- Separate timer (bellTollTime_) from regular city ambience
- State logging for debugging bell events
Implemented city-specific ambient soundscapes for all six major cities:
Alliance cities:
- Stormwind: day/night crowd and marketplace sounds
- Ironforge: underground forge ambience (no day/night)
- Darnassus: day/night elven city sounds
Horde cities:
- Orgrimmar: day/night orcish city atmosphere
- Undercity: underground undead ambience (no day/night)
- Thunder Bluff: day/night tauren plateau sounds
Technical details:
- Added CityType enum (NONE, STORMWIND, IRONFORGE, DARNASSUS, ORGRIMMAR, UNDERCITY, THUNDERBLUFF)
- Loads 12 city sound files from Sound\Ambience\WMOAmbience
- Underground cities (Ironforge, Undercity) use single sound without day/night variants
- 20s loop interval for city ambience (more frequent than zone ambience)
- Volume at 0.4 for noticeable but not overwhelming urban atmosphere
- Cities take priority over zone ambience to prevent mixing
- updateZoneAmbience() now checks for active city and skips if in city
- State change logging for debugging city transitions
Implemented three new ambient audio systems with automatic day/night transitions:
Weather ambience:
- Rain sounds (light/medium/heavy intensity based on weather system)
- Snow sounds (light/medium/heavy intensity)
- Automatically syncs with visual weather system in renderer
- Different loop intervals based on intensity (18-30s)
- Disabled indoors
Water ambience:
- Underwater swimming sounds (18s loop)
- Ocean surface sounds
- State tracking for entering/exiting water
Zone ambience:
- Forest (normal and snow variants)
- Beach sounds
- Grasslands
- Jungle
- Marsh/swamp
- Desert (canyon and plains variants)
- All zones have separate day/night sound files
- 30s loop interval for subtle background atmosphere
- Disabled indoors
Technical details:
- Added WeatherType enum (NONE, RAIN/SNOW LIGHT/MEDIUM/HEAVY)
- Added ZoneType enum (NONE, FOREST_NORMAL, FOREST_SNOW, BEACH, GRASSLANDS, JUNGLE, MARSH, DESERT_CANYON, DESERT_PLAINS)
- Loads 26 new sound files from Sound\Ambience\Weather and Sound\Ambience\ZoneAmbience
- Weather intensity thresholds: <0.33 = light, 0.33-0.66 = medium, >0.66 = heavy
- Renderer automatically converts Weather::Type + intensity to AmbientSoundManager::WeatherType
- All ambience respects volumeScale_ and indoor state
- State change logging for debugging transitions
Replaces the pitch-shifted weapon hit sound hack with the official
Sound\Ambience\WMOAmbience\BlackSmith.wav ambience loop. This provides
authentic blacksmith atmosphere with natural hammer strikes and anvil sounds.
Changes:
- Use proper blacksmith ambience file instead of modified weapon sounds
- Remove pitch shifting (1.6x) - use natural sound at 1.0x pitch
- Adjust volume to 0.6 (from 0.25) for ambient loop
- Play every 15 seconds (from 2.5s) for natural atmosphere
- Single loop file replaces 3 random weapon hit sounds
Implements full NPC voice interaction system supporting 6 different sound categories
for all playable races/genders. System loads ~450+ voice clips from MPQ archives.
Voice Categories:
- Greeting: Play on NPC right-click interaction
- Farewell: Play when closing gossip/dialog windows
- Vendor: Play when opening merchant/vendor windows
- Pissed: Play after clicking NPC 5+ times (spam protection)
- Aggro: Play when NPC enters combat with player
- Flee: Play when NPC is fleeing (ready for low-health triggers)
Features:
- Race/gender detection from NPC display IDs via CreatureDisplayInfoExtra.dbc
- Intelligent click tracking for pissed sounds
- Combat sounds use player character vocal files for humanoid NPCs
- Cooldown system prevents voice spam (2s default, combat sounds bypass)
- Generic fallback voices for unsupported NPC types
- 3D positional audio support
Voice Support:
- All playable races: Human, Dwarf, Gnome, Night Elf, Orc, Tauren, Troll, Undead
- Male and female variants for each race
- StandardNPC sounds for social interactions
- Character vocal sounds for combat
Technical Changes:
- Refactored NpcVoiceManager to support multiple sound categories
- Added callbacks: NpcFarewell, NpcVendor, NpcAggro
- Extended voice loading to parse both StandardNPC and Character vocal paths
- Integrated with GameHandler for gossip, vendor, and combat events
- Added detailed voice detection logging for debugging
Also includes:
- Sound manifest files added to docs/ for reference
- Blacksmith hammer pitch increased to 1.6x (was 1.4x)
- Blacksmith volume reduced 30% to 0.25 (was 0.35)
WotLK 3.3.5a uses .wav files for NPC voices, not .ogg as shown in retail Wowhead. Fixed audio engine to preserve original sample rate from WAV files (preventing chipmunk playback). Implemented race/gender detection using CreatureDisplayInfo.dbc and CreatureDisplayInfoExtra.dbc to play correct voice types for each NPC.
MPQs may store files internally with forward slashes. Testing both:
- Sound/Creature/... (forward slashes)
- Sound\Creature\... (backslashes)
This should help identify the correct path format for accessing files.
Tests various path structures to find where voice files actually exist:
- Wowhead StandardNPC paths
- Lowercase variations
- PC voice files (HumanVocMale/Female - player emotes)
- VO/Voice directories
- Simpler creature paths
This will help identify which files actually exist in the 3.3.5a MPQs.
Loads speech-enUS.MPQ, expansion-speech-enUS.MPQ, and lichking-speech-enUS.MPQ
which contain NPC voice/creature sound files. These MPQs were present but not
being loaded, causing all .ogg NPC greeting sounds to show as NOT FOUND.
NPC sounds are in Sound\Creature\[npctype]\ not Sound\Character\.
Changed to use correct paths like humanmalestandardnpc/humanfemalestandardnpc
with .ogg extension. This matches actual WoW 3.3.5a structure.
These managers were only being initialized in loadTerrainArea (online mode).
Added initialization in loadTestTerrain as well so they work in both modes.
Now the file probing should actually run and show which sound files exist.
The cachedAssetManager was only set in loadTestTerrain() for single-player mode.
In online mode (loadTerrainArea), it was never set, so NPC voice and mount sound
managers never initialized. Now gets asset manager from Application instance if
not already cached. This will enable file probing and voice/sound loading.
Tests various potential sound file paths at startup and logs which ones exist
in the MPQ. This will show us the correct file paths to use for both NPC
voices and tavern music.
Added extensive logging to diagnose issues:
- Logs WMO model IDs when entering WMOs (to identify correct tavern IDs)
- Logs when tavern music should be playing
- Logs NPC voice playGreeting calls and their results
- Logs which sound files are being played
- Added more emote variations (Hello, Yes) for NPC voices
This will help identify why tavern music and NPC voices aren't working.
The ImGuiWindowFlags_NoInputs flag was blocking all input including mouse clicks.
Removed this flag to restore action bar clickability while keeping keyboard shortcuts functional.
NPC voice fixes:
- Changed sound paths from "Greeting" to "Hello" emote (more reliable)
- Added warning log when no voice samples load
- Voice files should now play when clicking NPCs
Tavern music:
- Detect tavern WMOs by model ID (inn buildings)
- Play tavern-specific music when inside taverns
- Crossfade back to zone music when exiting taverns
- Adds cozy ambient music to inn/tavern buildings
Fixes issue where closing the flight map window would incorrectly mount the
player on a griffin at spawn point. Added taxiLandingCooldown when closing
taxi window to prevent UNIT_FLAG_TAXI_FLIGHT check from re-triggering mount.
Fixes issue where dust particles were angling down into terrain and not visible.
Changes spawn position to fixed 0.2 units above ground, increases upward velocity
to 1.2-2.5 for clearer rise, and uses only horizontal velocity component for drift
to prevent Z component from pushing particles downward.
Added NPC voice manager that plays greeting sounds when clicking on NPCs:
Features:
- Voice line library with multiple race/gender voice types (Human, Dwarf,
Night Elf, etc.)
- 3D positional audio - voice comes from NPC location
- Cooldown system prevents spam clicking same NPC
- Randomized pitch/volume for variety
- Loads greeting sounds from character voice files in MPQ
- Generic fallback voices for NPCs without specific voice types
Voice lines trigger automatically when gossip window opens (SMSG_GOSSIP_MESSAGE).
Uses same audio system as other sound effects with ma_sound_set_position.
The mount dust code was calling camera->getForward() without checking if
camera was null first, causing a crash that prevented mounting/dismounting.
Added camera null check to the condition.
Implemented dust cloud particles that spawn at mount feet when running
on the ground, similar to the original WoW. Features:
- Brownish/tan dust particles using point sprite rendering
- Spawn rate proportional to movement speed
- Particles rise up, drift backward, and fade out naturally
- Only active when mounted, moving, and on ground (not flying)
- Smooth particle animation with size growth and alpha fade
Uses similar particle system architecture as swim effects with
GL_POINTS rendering and custom shaders for soft circular particles.
Removed the "N" indicator that was showing on the minimap when rotation
was enabled. The indicator was confusing and not needed - when rotation
is disabled (default), north is always at the top of the minimap, and
when rotation is enabled, the map itself rotating shows the orientation.
Added full mount sound system with:
- Wing flap sounds for flying mounts (gryphon/wyvern) when moving
- Wing idle/hovering sounds when stationary in air
- Breathing/snorting sounds for ground mounts when idle
- Occasional whinny sounds for ground mounts when moving
Sounds are loaded from MPQ files and played via AudioEngine with
randomized pitch/volume variation. Mount sound manager tracks mount
type, movement state, and flying state to play appropriate ambient
sounds at natural intervals.
Updated setMounted() to accept creature display ID and notify the
mount sound manager, which uses display ID ranges to detect mount
type (flying vs ground).
Reduced mount footsteps from 4 beats to 2 beats per animation cycle
with evenly-spaced timing at 0.25 and 0.75 normalized time. The
previous 4-beat pattern (0.1, 0.35, 0.6, 0.85) was too frequent and
not properly synced with mount animations.
Changed default value from true to false in three places:
- Minimap class (rotateWithCamera member)
- GameScreen settings (minimapRotate_ and pendingMinimapRotate)
- Restore Interface Defaults button
The minimap will now remain fixed/north-up by default, and the settings
checkbox properly controls the rotation behavior.
Added ImGuiWindowFlags_NoInputs to action bar window so it doesn't capture
keyboard input. Previously, hovering over or clicking the action bar would
set WantCaptureKeyboard=true, blocking the 1-9,0,-,= hotkeys from working.
Closing the taxi map without selecting a destination now properly dismounts
the player. Previously, selectTaxiNode() would optimistically mount the
player, then closeTaxi() would only clear taxiWindowOpen_ without cleaning
up the mount state, leaving the player mounted.
Implement framework for playing mount sounds (flapping, galloping) based on mount type and movement state. Actual sound playback to be implemented next.
- Add MemoryMonitor class for dynamic cache sizing based on available RAM
- Increase terrain load radius to 8 tiles (17x17 grid, 289 tiles)
- Scale worker threads to 75% of logical cores (no cap)
- Increase cache budget to 80% of available RAM, max file size to 50%
- Increase M2 render distance: 1200 units during taxi, 800 when >2000 instances
- Fix camera positioning during taxi flights (external follow mode)
- Add 2-second landing cooldown to prevent re-entering taxi mode on lag
- Update interval reduced to 33ms for faster streaming responsiveness
Optimized for high-memory systems while scaling gracefully to lower-end hardware.
Cache and render distances now fully utilize available VRAM on minimum spec GPUs.
Each water tile generates waves independently using local coordinates,
causing height mismatches at tile boundaries. Reduced amplitude minimizes
visible seams.
Problem:
- Waves calculated per-tile using local vertex positions (aPos)
- Adjacent tiles have different local coords at shared boundary
- Wave height discontinuity creates visible gaps between tiles
Solution:
- Reduced ocean wave amplitude: 0.25 → 0.06 (75% reduction)
- Reduced canal wave amplitude: 0.10 → 0.04
- Kept frequency and speed for subtle animation
- Waves still visible but gaps minimized
Technical note: Proper fix would be world-space wave calculation, but
requires passing tile offset to shader. This amplitude reduction is
simpler and effective for smooth ocean appearance.
Taxi flights are externally controlled - no need for collision detection,
movement processing, or input handling. Massive performance improvement.
Changes:
- Early return in CameraController::update() when externalFollow_ is true
- Skips all collision queries (terrain, WMO, M2)
- Skips all movement physics (gravity, swimming, jumping)
- Skips input processing (keyboard, mouse)
- Skips camera collision raycasts
Performance impact:
- 100% elimination of ~17 collision queries per frame during taxi
- No wasted cycles on movement code when position is scripted
- Camera still updates position via setExternalFollow system
- Smooth taxi flights with zero collision overhead
This addresses the core issue: "shouldn't be looking for collision at all
during taxi" - now it doesn't!
Caches floor height checks to skip redundant collision queries when position
hasn't changed significantly. Major performance improvement during movement.
Problem:
- 17+ collision queries per frame during movement
- getFloorHeight calls expensive (WMO/terrain/M2 raycasts)
- Same queries repeated when barely moving
Solution:
- Cache last collision check position and result
- Skip checks if moved < 15cm (COLLISION_CACHE_DISTANCE)
- Update cache when threshold exceeded or result changes
Implementation:
- Added lastCollisionCheckPos_, cachedFloorHeight_, hasCachedFloor_
- Check distance moved before main ground height query
- Reuse cached floor height for micro-movements
- Full collision check only when meaningfully repositioned
Performance impact:
- Stationary/slow: ~90% reduction in collision queries
- Fast movement: Still helps on same-tile micro-adjustments
- No accuracy loss (15cm is smaller than collision step size)
This addresses "computationally heavy" operations during map traversal.
Eliminates visible grid symmetry when viewing ocean from altitude by adding
pseudo-random phase offsets and multiple wave frequencies.
Changes:
- Added position-based hash functions for pseudo-random phase offsets
- Phase-shifted primary waves (w1, w2) by hash * 2π
- Added 2 higher-frequency detail waves (w3, w4) at lower amplitudes
- Detail waves: 1.7x-2.1x frequency, 0.28-0.35x amplitude
- Different speeds and directions for each octave
Result: Natural-looking ocean with broken symmetry visible from altitude.
No more obvious grid pattern or repetitive crests.
The hash function creates consistent randomness per vertex position,
breaking the regular sine/cosine pattern without performance impact.
Fixes visible square gaps in ocean near coastlines by rendering masked tiles
if they have any water neighbors. Creates smoother, more natural coastlines.
Implementation:
- Check 8 neighboring tiles when a tile is masked out
- If any neighbor is water, render the edge tile anyway
- Creates 1-tile overlap at coastline boundaries
- Fills square gaps between water tiles
Before: Hard square edges with missing ocean tiles at coast
After: Continuous water coverage with smooth blended coastlines
The straight-line grid artifacts are reduced by ensuring adjacent water
tiles connect properly instead of leaving rectangular voids.
