I use this example from Azure - Continuous Recognition. However, I need to wait for a long time for the recognition to complete, for example for a 1 hour voice file.
Is it possible to have Azure to complete the recognition immediately? Currently, the rate of the recognition is around 1:1, meaning taking 1 hour to recognize a file of 1 hour. I hope it can be done in 1 mitune (1:60).
Hi this is Darren from Microsoft's Speech SDK team. If you are doing recognition from a WAV file, we attempt to upload audio at twice the "real-time" rate. Therefore, on a good network connection, and if the Azure region of the Speech Service you are using is geographically close to you, the fastest you will be able to transcribe one hour of speech is in half an hour.
What programing language are you using? Your link points to Python. Please have a look at our Python sample code for continuous recognition from a WAV file: https://github.com/Azure-Samples/cognitive-services-speech-sdk/blob/master/samples/python/console/speech_sample.py#L257
Note that if you need to do off-line process of long-duration speech files, a better solution for you would be to use "batch processing" and the REST APIs. SpeechSDK is more targeted for real-time transcription. Read about batch transcription here: https://docs.microsoft.com/azure/cognitive-services/speech-service/batch-transcription Samples for batch transcription can be found here: https://github.com/Azure-Samples/cognitive-services-speech-sdk/tree/master/samples/batch/
Darren
