Skip to main content
ZombieMutex

30 results · ↑ ↓ navigate · Enter open · Esc close

← All projects

Project notebook / Audio ML

Implemented

Streaming Speech Inference & Endpointing

I implemented frame-based streaming speech activity detection with hysteresis and endpoint hangover, then exercised it against synthetic telephony filtering, noise, packet loss and packet-arrival jitter.

Discuss this project

What I built

  • Implemented a 20-ms streaming VAD state machine with configurable energy threshold, speech confirmation and silence hangover.
  • Built deterministic packet arrival jitter/loss schedules, simple telephony-style band filtering, and audio frame processing.
  • Measured per-frame events, algorithmic hangover and local compute real-time factor, separating them from actual network and model latency.

Experiments & checks

  • On the included 1.25-second synthetic sample, the detector processed 63 20-ms frames and emitted a speech-start and speech-end event. The configured 8-frame hangover adds 160 ms of algorithmic trailing-silence delay. These are local signal-path observations, not end-to-end live-call latency or streaming-ASR results.

Notes & limitations

In a live conversation, the agent needs to know when to listen, when to respond and when to stop talking. Endpointing too early cuts a person off; too late makes the entire system feel sluggish.

The detector uses energy thresholds rather than a trained neural speech activity model. Packet schedules are simulated; there is no PSTN/SIP/RTP connection, ASR inference engine or GPU benchmark.

A voice assistant with a great language model and terrible turn-taking is just a very articulate interrupter.

Screenshots & media

Screenshot of the public ZombieMutex page for Streaming Speech Inference & Endpointing
Preview of the public project page or browser lab. This is not a screenshot of a private customer system. View full size ↗

Project updates

Dated notes tied to actual revisions of the site, not a simulated activity feed.

  1. Audio experiments published

    Added five implemented audio-ML projects with local experiments and measurement limits.

    Site revision 8d75d02

These entries reflect published changes to this website. The website repository is private; revision IDs are not public source links.

Discuss this project

Questions, ideas, minimal reproductions and corrections are welcome. Each project has its own public GitHub Discussion.

You can read without an account; posting requires GitHub sign-in. Keep private customer information, credentials and restricted material out of public threads.

Open this project's discussion