Skip to main content
Production use of this feature is available for specific editions only. Contact our sales team for more information.
Azure Speech Transcribe is an orchestration component that uses the Azure Speech service to transcribe speech to text from an audio file in your specified Azure Blob location and Speech resource endpoint. Speaker diarization clarifies which speaker spoke and when, increasing the accuracy of your transcriptions. You do not need to set up a Create Table component before using this component. Read What is the Speech service? to learn more. To stage data to Azure Blob Storage, the Azure credentials associated with your environment must be assigned the Storage Blob Data Contributor role. For more information, read User assigned with the Storage Blob Data Contributor role.

​
Prerequisites

​
Cloud credentials and authentication

Before you use the Azure Speech Transcribe component, you’ll need to add Azure cloud credentials to . This requires registering an application. You’ll also need an application for your Azure Speech service. The cloud credentials application and the Azure Speech applications require the following Azure roles:

​
Properties

Reference material is provided below for the Configure, Destination, and Advanced Settings properties.
​
Name
string
required
A human-readable name for the component.

​
Configure

​
Azure Blob Location
string
required
The URL prefix of the Azure Blob storage location where your audio file is stored. All files found under the prefix will be matched.
​
Azure Speech Endpoint
string
required
A unique endpoint to the Azure Speech resource. This endpoint is generated in Azure when you create an Azure Speech resource. To learn more, read Quickstart: Create an Azure AI services resource.
​
Source Audio Language Code
drop-down
required
Select the audio language code.
Locale (BCP-47)Language
en-USEnglish (United States)
es-ESSpanish (Spain)
es-MXSpanish (Mexico)
fr-FRFrench (France)
hi-INHindi (India)
it-ITItalian (Italy)
ja-JPJapanese (Japan)
ko-KRKorean (Korea)
pt-BRPortuguese (Brazil)
zh-CNChinese (Mandarin, Simplified)
​
Enable Speaker Diarization
boolean
required
Set to Yes if you wish to enable speaker diarization. According to the Azure documentation:
Diarization answers the question of who spoke and when. It differentiates speakers in an audio input based on their voice characteristics. Both real-time and batch APIs support diarization and are capable of differentiating speakers’ voices on monochannel recordings. Diarization is combined with speech to text functionality to provide transcription outputs that contain a speaker entry for each transcribed segment. The transcription output is tagged as GUEST1, GUEST2, GUEST3, etc. based on the number of speakers in the audio conversation.
This setting only works on monochannel recordings.
​
Min number of speakers
integer
required
The minimum number of speakers. This parameter is only available when speaker diarization is set to Yes.
​
Max number of speakers
integer
required
The maximum number of speakers. This parameter is only available when speaker diarization is set to Yes.

​
Destination

Select your cloud data warehouse.
​
Destination
drop-down
required
Select the destination for your data. This is either in Snowflake as a table or as files in cloud storage.
  • Snowflake: Load your data into a table in Snowflake. The data must first be staged via Snowflake or a cloud storage solution.
  • Cloud Storage: Load your data directly into files in your preferred cloud storage location. The format of these files can differ between source systems and will not have a file extension so we suggest inspecting the output to determine the format of the data.
​
Warehouse
drop-down
required
The Snowflake warehouse used to run the queries. The special value [Environment Default] uses the warehouse defined in the environment. Read Overview of Warehouses to learn more.
​
Database
drop-down
required
The Snowflake database to access. The special value [Environment Default] uses the database defined in the environment. Read Databases, Tables and Views - Overview to learn more.
​
Schema
drop-down
required
The Snowflake schema. The special value [Environment Default] uses the schema defined in the environment. Read Database, Schema, and Share DDL to learn more.
​
Table Name
string
required
The name of the table to be created in your Snowflake database. You can use a Table Input component in a transformation pipeline to access and transform this data after it has been loaded.
​
Load Strategy
drop-down
required
Define what happens if the table name already exists in the specified Snowflake database and schema.
  • Replace: If the specified table name already exists, that table will be destroyed and replaced by the table created during this pipeline run.
  • Truncate and Insert: If the specified table name already exists, all rows within the table will be removed and new rows will be inserted per the next run of this pipeline.
  • Fail if Exists: If the specified table name already exists, this pipeline will fail to run.
  • Append: If the specified table name already exists, then the data is inserted without altering or deleting the existing data in the table. It’s appended onto the end of the existing data in the table. If the specified table name doesn’t exist, then the table will be created, and your data will be inserted into the table.
​
Clean Staged files
boolean
required
  • Yes: Staged files will be destroyed after data is loaded. This is the default setting.
  • No: Staged files are retained in the staging area after data is loaded.
​
Stage Access Strategy
drop-down
Select the stage access strategy. The strategies available depend on the cloud platform you select in Stage Platform.
  • Credentials: Connects to the external stage (AWS, Azure) using your configured cloud provider credentials. Not available for Google Cloud Storage.
  • Storage Integration: Use a Snowflake storage integration to grant access to Snowflake to read data from and write to a cloud storage location. This will reveal the Storage Integration property, through which you can select any of your existing Snowflake storage integrations.
​
Stage Platform
drop-down
required
Use the drop-down menu to choose where the data is staged before being loaded into your Snowflake table.
  • Amazon S3: Stage your data on an AWS S3 bucket.
  • Snowflake: Stage your data on a Snowflake internal stage.
  • Azure Storage: Stage your data in an Azure Blob Storage container.
  • Google Cloud Storage: Stage your data in a Google Cloud Storage bucket.

​
Advanced Settings

​
Parse 'Null' & Empty Strings as NULL
boolean
required
Converts common strings that represent null into a null value. This is case-sensitive and works with the following strings: "", “NULL”, “NUL”, “Null”, “null”. The default is No.
Currently, this property is only applicable when using Snowflake as your destination.
​
Trim String Columns
boolean
required
When Yes, remove leading and trailing characters from a string column. The default is No.

​
Deactivate soft delete for Azure blobs (Databricks)

If you intend to set your destination as Databricks and your stage platform as Azure Storage, you must turn off the “Enable soft delete for blobs” setting in your Azure account for your pipeline to run successfully. To do this:
  1. In the Azure portal, navigate to your storage account.
  2. In the menu, under Data management, click Data protection.
  3. Clear the Enable soft delete for blobs checkbox. For more information, read Soft delete for blobs.