how to transcribe a long audio file in swift using ibm cloud?

Question

I have a 10mb and split it into 4mb(frame size limit) 3 parts and then send 3 calls of:

                    speechToText.recognizeUsingWebSocket(audio: chunkData, settings: settings, callback: callback)

but time stamps aren't consistent it gets reset to 0.0 for each call. What's the proper way?

Answer 1

you are making three separate calls to the STT service, that is why the timestamps reset. You need to send the 10MBs within the same call.