This script is designed to be minimalistic. You will run it on the command line.
The output should look something like this:
`spk_1 : 0:03 | Something* speaker 1 said.
spk_0 : 0:10 | Speaker 0 said something
spk_X : 0:15 | ^ ^ | | the time the speaker started | | The speaker_label in AWS's raw output file ` Words followed by an asterisk had a confidence level of less than 0.9.
- Nodejs vs 12+
- Node access on the command line
- Terminal program
This version does not check for the type of file that you output. It assumes you ran the Transcribe program with the settings to identify the speaker on.
Run this by passing the path to the program as follows. Assuming you are in the directory of index.js:
index.js "/path/to/aws.json"
When it runs successfully, you will get a 'completed' remark in the command line.
The file will output to the same path and have a filename of aws.json_out.txt