Microsoft's Azure Cognitive Services are a set of machine learning models for problems such as object recognition, sentiment analysis, entity extraction, and speech to text.
This demo uses the Speech API's Speech to Text Batch Transcription service to transcribe audio from an Azure Blob Storage Container
- Data must be uploaded into an Azure Blob Storage Container
- Audio files must be in either WAV or MP3 format (with PCM Codec) or OGG (OPUS Codec), with a bitrate of 16-bit and a sample rate of either 8 or 16 kHz, in either mono or stero.
- A Cognitive Services Speech Subscription will be required, using a standard tier subscription
Further details are available from the documentation
A Swagger is available for API testing and documentation (see "Custom Speech Transcriptions" section). Note you may need to change the region URL from uksouth to where you've provisioned the Speech API in the following section.