Google’s Speech Recognition combined with LCAT API for open-ended assessment
- Create a Google service account and enable “Google Speech To Text API”.
- If you don’t have a service account, please follow the steps to create one https://cloud.google.com/iam/docs/creating-managing-service-accounts
- As LCAT API V3 now accepts audio of 45 seconds duration, any audio less than 45 seconds can be passed straight to “Google Speech To Text API” to get the transcript.
- Once transcript is obtained, that content needs to be passed to “LCAT API” along with the audio to get the result, as per a normal API call.
- If the audio is above 45 seconds, it needs to be divided into smaller audio chunks that are less than 45s. If you require further assistance, contact us.
- Each audio chunk then should be passed to “Google Speech To Text” to get its transcript.
- Once transcripts are obtained, each audio chunk should be passed to LCAT API along with its corresponding transcript, as in step #4.
- Each chunk score obtained from LCAT API should be added up and divided by the number of chunks to get the average score.
- You can use the repository here to cut the audio as needed.