Transcribe: Understand/Save Speaker Detection

Question

TL;DR I need to recognize the speaker identification (diarization) of the user's voice; couldn’t find a way to do it.
I am building an application using AWS Transcribe streaming. I am able to get the speaker labels of each word, this is great; but, I need to somehow save and reuse the speaker label for further use with the same user.

Answer

Hi,

In addition to MaxSpeakerLabels, can you also set the ShowSpeakerLabels parameter to true value? Please check the speaker diarization page and the complete set of StartTranscriptionJob parameters. To go faster with the troubleshooting you can also try to test your audio files from the Amazon Transcribe console. Hope this helps.

Speaker diazization- https://docs.aws.amazon.com/transcribe/latest/dg/diarization.html

StartDescriptionJob- https://docs.aws.amazon.com/transcribe/latest/APIReference/API_StartTranscriptionJob.html

Transcribe: Understand/Save Speaker Detection

Conteúdo relevante