Getting word timestamps in python

maria.khan · July 8, 2020, 1:01pm

Hi all,

I have been trying to get word timestamps in the python package version of deepspeech. I have been able to get metadata but that only indicates the timestamp of each character.

However you can using --json or -e to get the timestamps of every word when using the command line binaries. Is there anyway to do this for the python version? Any help would be appreciated.

lissyx · July 8, 2020, 3:05pm

Have you tried looking at the source code of the python code?

github.com

mozilla/DeepSpeech/blob/master/native_client/python/client.py#L71-L77


def metadata_json_output(metadata):
    json_result = dict()
    json_result["transcripts"] = [{
        "confidence": transcript.confidence,
        "words": words_from_candidate_transcript(transcript),
    } for transcript in metadata.transcripts]
    return json.dumps(json_result, indent=2)

via · October 26, 2021, 9:36am

Hi,
I am trying to use deepspeech for web application, and I want to get speech to text with timestamp of every word spoken.
Can you please direct me to some working sample to do the same?

Thanks very much!