Altairis.Services.AzureSpeechToTextBatch 1.2.0

Azure AI Speech Batch Transcription Client

This library provides a client for performing batch transcription of audio files using Azure AI Speech services in batch mode. The Azure SDK lacks the batch transcription feature, so this client serves as a workaround to enable batch processing of audio files for transcription.

• Installation: use NuGet package Altairis.Services.AzureSpeechToTextBatch.
• API Reference Documentation

Getting started

Prerequisites

• An Azure subscription. If you don't have one, you can create a free account here.
• An Azure Speech resource. You can create one in the Azure portal. Make sure to note down the resource key and region, as you will need them to authenticate with the Azure Speech service.
• Most likely Azure Storage account with a blob container to store the audio files and the transcription results. Make sure to note down the storage account connection string and the container name. You don't have to use Azure Storage, but it is the easiest way to handle batch transcription, as the Azure Speech service requires the audio files to be accessible via a public URL.

Sample code

For more detailed sample code, see the demo CLI application in the Altairis.Services.AzureSpeechToTextBatch.Cli project.

Create a batch transcription

using Altairis.Services.AzureSpeechToTextBatch;

// Create batch data structure
var batch = new TranscriptionRequest {
    DisplayName = "My Transcription",
    ContentUrls = { "https://example.com/audio1.wav", "https://example.com/audio2.aac" },
    Locale = "en-US",
    Properties = new TranscriptionProperties {
        ProfanityFilterMode = ProfanityFilterMode.None,
        PunctuationMode = PunctuationMode.Automatic,
        DiarizationEnabled = false,
        DisplayFormWordLevelTimestampsEnabled = true,
        WordLevelTimestampsEnabled = true
    }
};

// Create instance of the client
var tc = new TranscriptionClient("<your-speech-key>", "<your-speech-region>");

// Submit the batch transcription to API
var response = await tc.CreateTranscription(batch, cancellationToken);
var transcriptionId = response.GetId();
Console.WriteLine($"Transcription created with ID: {transcriptionId}");

Get the status of a batch transcription and download all results

using Altairis.Services.AzureSpeechToTextBatch;

var transcriptionId = "00000000-0000-0000-0000-000000000000";

// Create instance of the client
var tc = new TranscriptionClient("<your-speech-key>", "<your-speech-region>");

// Get the status of the transcription job
var job = await tc.GetTranscription(transcriptionId);
if(job.Status != TranscriptionStatus.Succeeded) {
    Console.WriteLine($"Transcription status: {job.Status}");
    return;
}

// Download the results of the transcription job
Console.WriteLine($"Transcription succeeded. Downloading results...");
var files = await tc.ListFiles(transcriptionId);
using var httpClient = new HttpClient();
foreach(var file in files) {
    // Get SAS download URL for the file
    var downloadUrl = await tc.GetFileDownloadUrl(transcriptionId, file.GetId(), cancellationToken);

    // Download the file using HttpClient
    using var response = await httpClient.GetAsync(downloadUrl, cancellationToken);
    response.EnsureSuccessStatusCode();
    await using var fs = new FileStream(fileName, FileMode.Create, FileAccess.Write, FileShare.None);
    await response.Content.CopyToAsync(fs, cancellationToken);
}

// Delete transcription job
await tc.DeleteTranscription(transcriptionId);