Closed kvssankar closed 1 month ago
In speech to text, When numbers are spoken fast and smart format is true, and endpoint is 1000ms. Even after unlimited silence, is_final or speech_final never becomes true. Utterance end also is never detected.
let request: any = { language: "en", encoding: "mulaw", sample_rate: "8000", model: "enhanced-phonecall", punctuate: true, interim_results: true, endpointing: 1000, utterance_end_ms: 1000, filler_words: true, smart_format: true, };
In speech to text, When numbers are spoken fast and smart format is true, and endpoint is 1000ms. Even after unlimited silence, is_final or speech_final never becomes true. Utterance end also is never detected.
let request: any = { language: "en", encoding: "mulaw", sample_rate: "8000", model: "enhanced-phonecall", punctuate: true, interim_results: true, endpointing: 1000, utterance_end_ms: 1000, filler_words: true, smart_format: true, };