Extracting Speaker Labels, Start/End Times, and Transcript Segments from Amazon Transcribe JSON output (Python)

Question

Hi,

I'm using the Amazon Transcribe service with its Python API to convert audio to text. The transcript results come in JSON format. I'd like to extract specific information from the JSON, including:

Speaker labels (who is speaking)
Start and end times for each speaker segment
Transcript text for each speaker segment
Can you advise on the proper method or Python code format to achieve this extraction?

Didier Durand · Answer

Hi,

The simple and efficient way to parse json to a Python dict is to use json.load() function included in Python runtime: it does all the painful JSON parsing for you. You obtain a nice and clean Python dict in 1 single call.

See code samples at https://www.geeksforgeeks.org/json-load-in-python/

You have a Transcribe parser sample in Python at https://github.com/JeeZeh/winston/blob/main/transcribe.py

Best,

Didier

Extracting Speaker Labels, Start/End Times, and Transcript Segments from Amazon Transcribe JSON output (Python)

Add your answer

Relevant content