Reference documentation and code samples for the Cloud Speech-to-Text V2 API module Google::Cloud::Speech::V2::StreamingRecognitionFeatures::EndpointingSensitivity.
Endpointing sensitivity is intended for applications that want to minimize result latency, possibly at the expense of quality. Some utterances may be broken up into multiple fragments.
Constants
ENDPOINTING_SENSITIVITY_UNSPECIFIED
value: 0
If no value is specified, the values for
ENDPOINTING_SENSITIVITY_STANDARD will be used.
ENDPOINTING_SENSITIVITY_STANDARD
value: 1
Standard sensitivity, no optimization for latency.
ENDPOINTING_SENSITIVITY_SUPERSHORT
value: 2
Super short sensitivity, optimized for super short utterances like single
words ("Yes", "No", "Hello", etc.) or a single phrase, command or short
query (e.g. "check balance", "five dollars", "call Mom").
ENDPOINTING_SENSITIVITY_SHORT
value: 3
Short sensitivity, optimized for short utterances like single sentences.
(e.g. "Remind me to call the dentist tomorrow morning at nine",
"Navigate to the nearest coffee shop that is currently open")