Cloud Speech-to-Text V2, Chirp 3 (Standard recognition)
Speech-to-text - chirp_3
- Price (provider's unit)
- Not stated
- Converted $/hour (our arithmetic)
- Not convertedNo price stated
- Languages
111 locales (29 GA, 82 Preview) in the transcription language table derived
“Chirp 3 supports transcription in StreamingRecognize, Recognize, and BatchRecognize in the following languages:”
Context - docs.cloud.google.com Note: Rows counted by us in the transcription language table (29 GA, 82 Preview).
- Streaming
Yes
“V2 | Speech.StreamingRecognize (good for streaming and real-time audio) | Supported”
docs.cloud.google.com - Max audio
BatchRecognize: 1 minute to 1 hour (20 minutes with word-level timestamps)
“Speech.BatchRecognize (good for long audio 1 minute to 1 hour in general, but up to 20 minutes with word-level timestamp enabled)”
docs.cloud.google.com - Status
GA (us and eu multi-regions)
“| us (multi-region) | GA | | eu (multi-region) | GA |”
docs.cloud.google.com
Features
speaker diarization (BatchRecognize only)
“Speaker Diarization | Automatically identifies the different speakers in a single-channel audio sample. Available only in Speech.BatchRecognize | GA”
docs.cloud.google.com utterance-level timestamps (streaming)
“Utterance-level timestamps | Automatically generated by the model. Available only in Speech.StreamingRecognize | GA”
docs.cloud.google.com speech adaptation (biasing)
“Speech adaptation (Biasing) | Provides hints to the model in the form of phrases or words to improve recognition accuracy for specific terms or proper nouns. | GA”
docs.cloud.google.com
Language count is ours, from the table rows. The diarization feature row says 'Available only in Speech.BatchRecognize' while the diarization language section says BatchRecognize and Recognize.