What Is Interruption Latency?
Short answer: Interruption latency is the time between detectable user interruption and the avatar’s speech and motion actually stopping.
For engineering teams, interruption latency is a concrete conversation control concern rather than a visual label. Even accurate barge-in feels broken if buffered output continues for too long. The useful engineering question is not merely whether the feature exists, but which component owns it and which event proves it worked.
| Quick reference | Answer |
|---|---|
| Category | Conversation timing |
| Stack boundary | Conversation control |
| Primary concern | Even accurate barge-in feels broken if buffered output continues for too long. |
| Example | A kiosk avatar must stop giving directions almost immediately when the visitor corrects the destination. |
Interruption Latency definition
Interruption latency is the time between detectable user interruption and the avatar’s speech and motion actually stopping. Here the term is scoped to a live AI avatar: a system that listens, generates a response, produces speech and motion, and presents the result while the user remains in the interaction. In that setting, interruption latency must coexist with conversation state, interruption, synchronization, and device constraints.
An implementation definition should name the input, output, owner, and lifecycle. That prevents one team from using “interruption latency” for a local operation while another uses it for the user-visible outcome. Even accurate barge-in feels broken if buffered output continues for too long.
Why Interruption Latency matters in a real-time AI avatar
Even accurate barge-in feels broken if buffered output continues for too long. When this control boundary is wrong, the avatar may answer over the user, wait through an obvious completion, or continue a response after the user has already changed direction. In practice, this makes interruption latency part of the product experience rather than an invisible implementation detail.
The risk is easiest to see in the article’s example: a kiosk avatar must stop giving directions almost immediately when the visitor corrects the destination. The behavior needs to remain correct across the whole turn, including queued work and late events, not only at the instant the primary decision is made.
Where Interruption Latency sits in the avatar stack
How a realtime avatar listens, yields the floor, responds, and stops. Microphone frames, speech-detection events, and partial transcripts enter the conversation controller. The controller combines those signals with the current speaking state, then decides whether to keep listening, yield the floor, dispatch a response, or cancel work already in flight.
For interruption latency, the upstream boundary is user audio and transcript evidence. The downstream boundary is the turn state machine, including LLM dispatch, TTS playback, animation scheduling, and cancellation. Assign one conversation-state owner that can coordinate input events and invalidate every downstream artifact belonging to an obsolete turn. Any later component should consume the resulting state or data without silently redefining what the term means.
How Interruption Latency works
1. Define the input and configuration boundary.
Measure separate timestamps for microphone onset, VAD trigger, cancellation, and audible stop. Document the chosen value or rule alongside the environment in which it was tested; otherwise a change can alter interruption latency without a clear baseline.
2. Make runtime ownership explicit.
Stop local playback in addition to cancelling upstream generation. Make the responsible component visible in logs and cancellation paths so two services do not make conflicting decisions about the same turn.
3. Turn the behavior into an observable contract.
Correlate cancellation with a turn or response identifier to avoid stopping a newer turn. Capture the corresponding event or state in telemetry and test both the expected path and a failure path. This turns interruption latency from an assumption into a verifiable behavior.
Practical example
A kiosk avatar must stop giving directions almost immediately when the visitor corrects the destination. A useful test recreates that moment and follows the term-specific controls in order:
- Measure separate timestamps for microphone onset, VAD trigger, cancellation, and audible stop.
- Stop local playback in addition to cancelling upstream generation.
- Correlate cancellation with a turn or response identifier to avoid stopping a newer turn.
How to test or measure Interruption Latency
Instrument the full event timeline instead of recording one aggregate duration. Capture user-speech onset, detector output, endpoint decision, response dispatch, first playback, cancellation request, and actual audible or visible stop whenever those events apply.
For interruption latency, track false triggers, missed turns, overlap duration, decision delay, cancellation completion, and stale playback. Review distributions and failure counts rather than relying on one successful demo. Segment the result by language, speaking style, room noise, microphone route, speaker route, and device class; a global average can conceal a failure limited to one environment.
Minimum test checklist
- Boundary: Measure separate timestamps for microphone onset, VAD trigger, cancellation, and audible stop.
- Ownership: Stop local playback in addition to cancelling upstream generation.
- Verification: Correlate cancellation with a turn or response identifier to avoid stopping a newer turn.
- Run the same test once on the primary environment and once on a constrained or failure-prone segment.
- Keep start and end events unchanged when comparing releases.
Tradeoffs and failure modes
- Boundary mismatch: If the implementation violates the rule “Measure separate timestamps for microphone onset, VAD trigger, cancellation, and audible stop”, the observed behavior can vary by environment without a trustworthy baseline.
- Ownership conflict: If it violates “Stop local playback in addition to cancelling upstream generation”, two components may act on different assumptions or leave stale work active.
- Invisible regression: If it violates “Correlate cancellation with a turn or response identifier to avoid stopping a newer turn”, a release can change interruption latency without leaving enough evidence to isolate the cause.
Common misconception
This is one control signal inside a conversation loop, not a substitute for measuring the whole end-to-end experience. For interruption latency, the reliable claim is the definition and test boundary documented on this page—not a broader promise about every stage of the avatar pipeline.
Frequently asked questions
Is Interruption Latency the same as Audio Backpressure?
No. The concepts interact, but they describe different boundaries. For interruption latency, the relevant definition is: Interruption latency is the time between detectable user interruption and the avatar’s speech and motion actually stopping. For audio backpressure, it is: Audio backpressure is flow control applied when a producer generates audio faster than downstream components can consume it. Instrumenting them separately makes the root cause of a failure easier to isolate.
What should a team define first for Interruption Latency?
Start with the event or data boundary: measure separate timestamps for microphone onset, VAD trigger, cancellation, and audible stop. Then name the component that owns the rule and the observable result that proves it worked. This prevents two implementations from using the same term for different behavior.
How does Interruption Latency connect to Conversation ID and Audio Backpressure?
Conversation ID covers a neighboring concern: A conversation ID is a stable identifier for one multi-turn interaction between a user and an avatar. Audio Backpressure covers another: Audio backpressure is flow control applied when a producer generates audio faster than downstream components can consume it. Read the three definitions together, but keep their events and ownership separate in telemetry so one metric does not mask another.
Related glossary terms
- Barge-In — Barge-in allows a user to interrupt an avatar’s active response by speaking.
- Conversation ID — A conversation ID is a stable identifier for one multi-turn interaction between a user and an avatar.
- Audio Backpressure — Audio backpressure is flow control applied when a producer generates audio faster than downstream components can consume it.
Continue to implementation and evaluation
- Implementation path: Silero VAD integration
- Evaluation path: Best low-latency AI avatar platforms
- Browse the complete real-time AI avatar glossary
References
Last reviewed: 2026-08-19. Review the linked specifications and current Spatius documentation before using this article as an implementation contract.