Fact · Speech recognition
1real-time factor
Microsoft says a device should have a real-time factor of 1 or lower to support real-time speech experiences.
Checked against its source on October 17, 2026
In the source's own words
“the device should have an RTF of 1 or lower”
Embedded speech performance evaluations (Speech service), Microsoft Learn, n.d., Metrics & terminology.
Open the original- What it measures
- Microsoft's threshold for embedded speech on a device. A value above 1 means the device cannot keep up with the audio as it arrives.
- What it does not show
- A threshold stated for embedded (on-device) speech models. Microsoft also recommends measuring several samples and looking at the distribution.