In the news
Perceived TTFA: the Coval TTS latency benchmark
Gradium · Published · 3 min read
In 30 seconds
- What happened
- Coval TTS benchmark now measures perceived latency including leading silence, exposing models that emit inaudible audio before speech starts.
- Why it matters
- Matters for engineers building voice applications where user-perceived responsiveness depends on when actual sound reaches the speaker.
- Watch out
- Historical benchmark rows before this methodology change are not comparable; the metric shift revealed that some fast models felt slow due to silence padding.
- latency
- benchmark
The patterns behind this
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.