Fact · Speech recognition
Sainath et al. define endpointer latency as the time it takes for the microphone to close after the user stops speaking.
Checked against its source on October 17, 2026
In the source's own words
“We define endpointer (EP) latency as the amount of time it takes for the microphone to close after a user stops speaking.”
A Streaming On-Device End-to-End Model Surpassing Server-Side Conventional Model Quality and Latency, Tara N. Sainath et al., ICASSP 2020 (arXiv:2003.12710), 2020, Section 4.
Open the original- What it measures
- The authors use this definition to compare on-device and server endpointers, and they exclude network and computation time from it.
- What it does not show
- Their definition excludes the time for the final text to arrive, so it is one latency among several and not the full wait a user feels.