Lyo.Tts.WindowsSpeech 2.0.0

Lyo.Tts.WindowsSpeech

Text-to-speech through Windows SAPI. WindowsSpeechTtsService uses the built-in Speech API.

Features

• SAPI. Calls the Windows Speech Synthesis API.
• Voices. Pick any installed Windows voice.
• Rate. Speech rate in the range -10 to 10.
• Volume. Volume in the range 0 to 100.
• Bulk. Batch synthesis through the shared Lyo.Tts pipeline.
• Logging. Writes through Microsoft.Extensions.Logging.
• Metrics. Optional instrumentation for TTS calls.
• Concurrency. Safe to use from multiple threads.
• Async. Every method accepts a CancellationToken.

Examples

Listen for events

using Lyo.Tts;
using Lyo.Tts.Models;
using Lyo.Tts.WindowsSpeech;
using Microsoft.Extensions.DependencyInjection;

var services = new ServiceCollection();
services.AddWindowsSpeechTtsService(options =>
{
    options.DefaultVoiceId = "Microsoft Zira Desktop";
    options.DefaultOutputFormat = AudioFormat.Wav;
    options.MaxTextLength = 5000;
    options.EnableMetrics = true;
});

await using var serviceProvider = services.BuildServiceProvider();
var ttsService = serviceProvider.GetRequiredService<ITtsService<WindowsTtsRequest>>();

var result = await ttsService.SynthesizeAsync("Hello, world!");
if (result.IsSuccess && result.AudioData is { Length: > 0 })
    await File.WriteAllBytesAsync("output.wav", result.AudioData);

Ready-made options instance

services.AddWindowsSpeechTtsService(new TtsServiceOptions
{
    DefaultVoiceId = "Microsoft Zira Desktop",
    DefaultOutputFormat = AudioFormat.Wav,
    MaxTextLength = 5000,
    EnableMetrics = true,
});

Enumerate voices

var ttsService = serviceProvider.GetRequiredService<ITtsService<WindowsTtsRequest>>();
var ok = await ttsService.TestConnectionAsync();
// TestConnectionAsync logs every installed SAPI voice via the registered ILogger.

Supported platforms

Windows only. This package builds and runs on Windows. It needs the Windows-specific System.Speech package.

What gets registered

AddWindowsSpeechTtsService (Action<TtsServiceOptions>? or TtsServiceOptions) wires:

• TtsServiceOptions (singleton).
• WindowsSpeechTtsService (singleton; subclass of TtsServiceBase<WindowsTtsRequest>).
• ITtsService<WindowsTtsRequest> resolved from the singleton above.

It does not wire the non-generic ITtsService. No WindowsSpeechTtsAppService adapter in this package today, so callers that depend on ITtsService should depend on ITtsService<WindowsTtsRequest> instead, or wire their own adapter. The other Lyo TTS providers (Lyo.Tts.AwsPolly, Lyo.Tts.Typecast) register both interfaces because they ship an *TtsAppService adapter.

Requirements

• Windows OS
• .NET Standard 2.0 or .NET 10.0
• System.Speech package (pulled in automatically)

Notes

• This library will not compile off Windows
• The library calls Windows SAPI, which exists only on Windows
• Windows Speech Synthesis always emits WAV

Dependencies

Generated from ProjectReference / PackageReference (same model as docs/Lyo.ProjectGraph.html).

• Lyo.Common.Core (direct, lyo)
• Lyo.Exceptions (direct, lyo)
• Lyo.Tts (direct, lyo)
• Lyo.Tts.Models (direct, lyo)
• System.Speech 10.0.5 (direct, microsoft, $([MSBuild]::IsOSPlatform('Windows')))
• Lyo.Metrics (transitive, lyo)
• Lyo.Result (transitive, lyo)
• Microsoft.Bcl.AsyncInterfaces 10.0.5 (transitive, microsoft, netstandard2.0)
• Microsoft.Extensions.Logging.Abstractions 10.0.5 (transitive, microsoft)
• System.Memory 4.6.3 (transitive, microsoft, netstandard2.0)
• System.Text.Json 10.0.5 (transitive, microsoft, netstandard2.0)