Transcriber

Wrapper for Googles Speech to Text API optimzed for transcribing interviews

Requirements

• NodeJS
• Google Cloud Access (free 300$ trial available)

Preperation

• The audio should be a mono FLAC file. Video or Audio can be converted for example with VLC.
• If the audio is lontger than 1 minute it needs to be uploaded to a google cloud storage bucket
  • Create a bucket: https://console.cloud.google.com/storage/
  • Upload file to bucket
• Follow this guide for authentication: https://cloud.google.com/docs/authentication/getting-started
  • set GOOGLE_APPLICATION_CREDENTIALS=C:\Users\ssteinrueck3\Downloads\owner.json

Installation & Usage

Initial install: npm i

Set the URI in analyze.js to your audio file in the bucket and run npm run analyze <name>. You can also change options like language or speaker count here. This operation can take up to half the time of the original audio length.

You can then use npm run format <name> to convert the raw response to a textfile. You can change the output in format.js without the need to run the analysis again.

npm i # Install dependencies
npm run analyze name # Analyse audio on google Cloud and store result in raw.json
npm run format name  # Format response to readable output

Clean up

To save resources you should remove the audio file from the storage bucket if it is no longer needed.