Audio file to text conversion in AWS Lambda function

Question

I am trying to convert an audio file into text using speech_regcognition python package. But It works only with the very small audio file (Let's say 10 sec). But I have to convert an audio file which is more than 60 secs. I know that I can use Google speech API to convert an audio file into text but it is paid service. I don't want to use a paid service. So now I have planned to split that original audio file into multiple small audio files. Let's say If I have 60 secs audio file I am converting each part as 10-sec audio files. I am trying to do this part in AWS Lambda service once the audio files are uploaded into S3 bucket. So that I can use this speech_recognition free service, I don't need to use Google API key (Paid).

My question: If I use this methodology, it will create any problem or google service can identify while I am using this in my AWS Lambda service?

Note: This audio file to text file conversion AWS Lambda function will be executed more than millions time.

Answer 1

I would recommend that you save yourself some time and hassle by using Amazon Transcribe. It would be easy to integrate with your AWS Lambda function.

If you are worried about costs, there is a Free Tier:

60 MINUTES PER MONTH FOR 12 MONTHS

The Amazon Transcribe Free Tier is available to you for 12 months, starting from the date on which you create your first transcription request. When your free usage expires, or if your application use exceeds the free usage tier, you simply pay standard, pay-as-you-go service rates.

Thereafter, it is $0.0004 per second. That's 2.4¢ per minute or $1.44 per hour.

The personal time you'll save will be more than worth the cost if you exceed the Free Tier.