When a fight breaks out in a live-service game, the soundscape can quickly become cacophonous—abilities overlap, movement stacks, and players still have to act on what they hear. Sandy Zhou is a Riot Games sound designer whose credited work spans major live-service releases. "In a teamfight, sound is a competitive interface," Zhou says. "If the mix doesn't protect what players must track, it doesn't matter how beautiful the ambience is." For Zhou, "adaptive" isn't a vibe—it's an engineering requirement: behavior you can specify, predict, and ship.
A Practical Definition of "Adaptive"
"I want a definition that survives real gameplay," Zhou says. "I should be able to describe what stands out when density rises, what steps back, and what the transition curve is. If I can't explain those three, the result won't hold up in play."
Those questions force real choices: what wins when events stack, what deprioritizes without feeling like it vanished, and how the soundscape moves between states without a hard switch. "Players shouldn't feel a state flip," Zhou says. "They should feel continuity."
Zhou organizes that continuity with two levers: "Mix priorities decide what can lead and what has to step back. Playback logic decides how the world behaves over time—how it changes, how it returns, and how it stays consistent as content grows."
Lever 1: Mix priorities—making space for what players must hear
Zhou treats readability as a hard constraint. "Readability has a ceiling," she says. "You need a priority ladder, a headroom policy, and a clear definition of what's allowed to lead when multiple things fire at once." She adds, "A common failure mode is when ambience starts competing with gameplay-critical cues. When the mix loses its hierarchy, the player loses information."
Her approach is to establish a stable baseline with layered details, which are present when room allows, and can step back when clarity is needed. "The key is that I can name what the mix protects," Zhou says, "and I can predict what will step back when density spikes."
Implementation varies by project, but the mechanisms are modest and repeatable. "In practice, that can mean hierarchy-based gain staging, a controlled surge on the right bus when the game can afford it, light bus-level ducking, or dynamic EQ that clears space without making the world feel like it 'turned off.'"
Transitions are part of the same problem. "Players feel the switch if you hard-cut:f you hard-cut." Zhou says. "A good transition is a gradient." She thinks in curves—how quickly a layer enters, how long it lingers, and how cleanly it exits—so change reads as a world, not a system. "If the curve is wrong," she says, "the player hears mechanics instead of a world."
Here's what that looks like outside a live-service fight.
Zhou points to an open-world problem: making a city read at multiple scales. From a skyline vista, you need a broad tonal bed; closer in, you need local texture—without abrupt swaps. "One approach is region-based ambience that changes with distance," she says, "and a curve that makes the transition feel continuous rather than like a state flip."
Elevation adds a second dimension. "High ground and low ground shouldn't sound identical," Zhou specifies. "If the player climbs, the mix can shift elements such as wind, traffic smear, and density of details without sounding like you have swapped loops." The technical bar is the same: a predictable hierarchy and transitions that don't expose the mechanism.
Here's what that looks like in League of Legends.
The same hierarchical thinking applies when the game shifts into a cinematic celebration. Zhou points to a player-visible example from League of Legends: a deluxe Legendary skin moment for Diana that triggers a brief map takeover after a pentakill—where the environment and music step forward like a short cinematic beat.
Zhou describes, "The takeover is built with progressive cues and then holds for roughly twenty seconds before resolving back." But the point isn't the SFX alone—it's that the mix makes room for the moment. "You're temporarily letting the celebration lead," she explains. "Other elements step back just enough to create space, and then everything returns on a controlled ramp. The goal and the result is that the comeback feels continuous rather than like a reset."
Lever 2: Playback logic—keeping the world consistent at scale
If mixed priorities decide what wins now, then playback logic decides how the world behaves over time and space—especially as content scales. "At world scale, you can't hand-author everything forever," Zhou notes. "Maintainability becomes the real constraint."
"In practice it's state-based with clear inputs, predictable transitions, and hard limits so variation doesn't turn random." Zhou specifies. She breaks it into four parts: inputs (time of day, location tags, pressure states, weather, narrative beats), playback logic (how layers enter, build, and cool down), location selection (what wins when zones overlap), and limits (how you prevent repetition and listener fatigue).
If smoothing is weak, states flutter; if limits are weak, details pile up; if overlap logic is unclear, the world feels inconsistent. "Random isn't natural—consistency is what makes a world believable," Zhou says. "Players can't always describe this but they feel it."
She also points out that density should often track activity, not just location: when nearby events spike, the soundscape doesn't need to get globally louder—it needs to rebalance what's allowed to lead.
Why Environmental Audio Gets Expensive Fast if You Plan It Late
Ask Zhou what teams underestimate about environmental audio, and she doesn't start with aesthetics. "Environmental audio gets expensive fast if you plan it late," she says.
"The hidden costs scale quickly—long-form iteration, variation management, and mix hierarchy maintenance across new content," Zhou explains. This is why she insists on early decisions: what's shared versus bespoke, how the system degrades under load, and how transitions behave in worst cases. "If you wait," Zhou says, "you pay for it everywhere—implementation time, clarity, maintenance, performance."
The Throughline: Clarity and Immersion Aren't Opposites
From the outside, teamfights and exploration can feel like opposites—dense competitive chaos versus spacious atmosphere. Zhou's view is much simpler. Both of these demand stability as conditions change. "In teamfights, you're protecting what players must track," she says, "In exploration, you're protecting how the world feels over time."
For Zhou, that stability is the real definition of adaptive. The answer is not more layers or bigger moments but rather, predictable behavior that survives real play. "The goal is that players never think about the system," she says. "They just trust what they're hearing."
"If you do it right," Zhou adds, "the mix disappears—and the world shows up."
About Sandy Zhou
Sandy Zhou is a game sound designer whose credits include League of Legends, Teamfight Tactics, Legends of Runeterra, Call of Duty: Mobile, Return to Empire, and Ruined King: A League of Legends Story (in-game cinematics). She has delivered game audio work for teams at Riot Games, Tencent Games, and ArenaNet. She is affiliated with MPSE and G.A.N.G.
