syncfusion-angular-speech-to-text（syncfusion Angular speech TO text 音频）

Syncfusion Angular SpeechToText Component

Component Overview

The Syncfusion Angular SpeechToText component is a lightweight, accessible control that converts spoken words into text using the browser's native Speech Recognition API.

Key Capabilities:

• Real-time Transcription - Convert speech to text with interim and final results
• Button Customization - Configure icon, text, positioning, and styling for start/stop buttons
• Tooltip Support - Add tooltips with custom content and positioning for better UX
• Events & Interactions - Handle listening state changes, errors, and transcript updates
• Methods - Programmatic control via startListening() and stopListening() methods
• Language Support - Multi-language recognition with locale configuration
• Globalization - RTL support and localization for international applications
• Styling Options - CSS classes, HTML attributes, and theme integration
• Security Features - HTTPS enforcement and explicit microphone permissions
• Browser Compatibility - Chrome 25+, Edge 79+, Safari 12+, Opera 30+

Documentation and Navigation Guide

Getting Started

📄 Read: references/getting-started.md

• Package installation (Ivy vs ngcc compatibility)
• Angular project setup and configuration
• CSS imports and theme management
• Module imports and basic component rendering
• Button content customization
• Complete working example with TextArea integration

Button Customization

📄 Read: references/button-customization.md

• ButtonSettingsModel configuration
• Start and stop button content labels
• Icon customization with CSS classes
• Icon positioning (top, bottom, left, right)
• Primary button styling for emphasis
• Visual style variations and best practices

Tooltip Configuration

📄 Read: references/tooltip-configuration.md

• TooltipSettingsModel setup
• Tooltip content for start and stop states
• Tooltip positioning relative to button
• Enabling and disabling tooltips
• Practical tooltip usage patterns

Styling and Appearance

📄 Read: references/styling-and-appearance.md

• CSS class system (e-primary, e-outline, e-info, e-success, e-warning, e-danger)
• Custom styling with cssClass property
• Theme configuration and integration
• HTML attributes customization
• Custom CSS examples and advanced styling

Events and Methods

📄 Read: references/events-and-methods.md

• Event types (created, onStart, onStop, onError, transcriptChanged)
• Event handler implementation and event arguments (including event, name properties)
• Programmatic control with startListening(), stopListening(), and destroy() methods
• Controlling component state with listeningState property (Inactive, Listening, Stopped)
• Canceling listening with the cancel property in onStart event
• Distinguishing user vs programmatic actions with isInteracted property
• Event handling patterns for different scenarios
• Managing listening lifecycle programmatically
• State transition management and workflow control
• Component cleanup and memory management

Speech Recognition Features

📄 Read: references/speech-recognition-features.md

• Retrieving transcript from speech input
• Setting recognition language (lang property)
• Interim results configuration (allowInterimResults)
• Listening state management (Inactive, Listening, Stopped)
• Tooltip visibility control (showTooltip)
• Disabled state configuration
• State persistence with enablePersistence property
• State monitoring and status management
• Custom state management with localStorage

Internationalization

📄 Read: references/internationalization.md

• Localization setup using L10n.load()
• Translation key mapping for error messages and labels
• Multi-language implementation examples
• RTL (Right-to-Left) support enablement
• Language-specific configuration best practices

Security and Error Handling

📄 Read: references/security-and-error-handling.md

• Security risks and mitigation strategies
• Data transmission and privacy concerns
• HTTPS enforcement and MITM prevention
• Microphone permission management
• Error type handling (no-speech, aborted, audio-capture, not-allowed, service-not-allowed, network, unsupported-browser, default)
• Browser support matrix and compatibility

Quick Start Example

Here's a minimal working SpeechToText component with transcript output:

import { Component, ViewChild } from '@angular/core';
import { SpeechToTextModule, TextAreaComponent, TextAreaModule, TranscriptChangedEventArgs } from '@syncfusion/ej2-angular-inputs';

@Component({
    imports: [SpeechToTextModule, TextAreaModule],
    standalone: true,
    selector: 'app-root',
    template: `
    <div class="speechText-container">
        <button ejs-speechtotext
                (transcriptChanged)="onTranscriptChange($event)">
        </button>
        <ejs-textarea #outputTextarea
                      id="textareaInst"
                      value=""
                      rows="5"
                      cols="50"
                      resizeMode="None"
                      placeholder="Transcribed text will be shown here...">
        </ejs-textarea>
    </div>`
})
export class AppComponent {
    @ViewChild('outputTextarea') outputTextarea!: TextAreaComponent;

    onTranscriptChange(args: TranscriptChangedEventArgs): void {
        this.outputTextarea.value = args.transcript;
    }
}

Common Patterns

Pattern 1: Customized Button with Events

Implement a styled button with custom labels and event handlers:

public buttonSettings: ButtonSettingsModel = {
    content: 'Start Listening',
    stopContent: 'Stop Listening',
    iconCss: 'e-icons e-play',
    stopIconCss: 'e-icons e-pause',
    isPrimary: true
};

onListeningStart(args: StartListeningEventArgs): void {
    console.log('Listening started');
}

onListeningStop(args: StopListeningEventArgs): void {
    console.log('Listening stopped');
}

Pattern 2: Multi-Language Support

Configure localization for international applications:

L10n.load({
    'de': {
        "speech-to-text": {
            "startAriaLabel": "Drücken Sie, um zu sprechen",
            "stopAriaLabel": "Drücken Sie, um zu stoppen",
            "noSpeechError": "Keine Sprache erkannt"
        }
    }
});

Pattern 3: Programmatic Control

Manage listening state programmatically without user button clicks:

@ViewChild('speechtotext') speechToText!: SpeechToTextComponent;

public startListening(): void {
    this.speechToText.startListening();
}

public stopListening(): void {
    this.speechToText.stopListening();
}

Pattern 4: Error Handling

Handle speech recognition errors gracefully:

onErrorHandler(args: ErrorEventArgs): void {
    if (args.error === 'no-speech') {
        console.log('No speech detected. Please try again.');
    } else if (args.error === 'not-allowed') {
        console.log('Microphone permission denied.');
    }
}

Pattern 5: Conditional Listening with Cancel

Control when listening can start using the cancel property:

onListeningStart(args: StartListeningEventArgs): void {
    // Cancel listening if conditions aren't met
    if (!this.hasPermission || !navigator.onLine) {
        args.cancel = true;
        alert('Cannot start listening. Check permissions and connection.');
        return;
    }

    // Track if user or system triggered the action
    if (args.isInteracted) {
        console.log('User clicked the button');
    } else {
        console.log('Started programmatically');
    }
}

Pattern 6: Distinguishing Interim vs Final Results

Handle interim and final transcripts differently:

onTranscriptChange(args: TranscriptChangedEventArgs): void {
    if (args.isInterimResult) {
        // Show interim results in real-time (lighter styling)
        this.interimText = args.transcript;
    } else {
        // Process final results (save, send to API, etc.)
        this.finalText = args.transcript;
        this.saveTranscript(args.transcript);
    }
}

Pattern 7: Controlling Component State

Programmatically control the component's operational state:

import { SpeechToTextState } from '@syncfusion/ej2-angular-inputs';

export class AppComponent {
    public currentState: SpeechToTextState = SpeechToTextState.Inactive;

    // Set component to listening state
    startListening(): void {
        this.currentState = SpeechToTextState.Listening;
    }

    // Set component to stopped state
    stopListening(): void {
        this.currentState = SpeechToTextState.Stopped;
    }

    // Reset to inactive state
    resetState(): void {
        this.currentState = SpeechToTextState.Inactive;
    }
}

Pattern 8: Proper Component Cleanup

Ensure proper cleanup to prevent memory leaks:

import { OnDestroy } from '@angular/core';

export class AppComponent implements OnDestroy {
    @ViewChild('speechtotext') speechToText!: SpeechToTextComponent;

    ngOnDestroy(): void {
        if (this.speechToText) {
            this.speechToText.stopListening();
            this.speechToText.destroy();
        }
    }
}

Key Props

Browser Requirements

The SpeechToText component requires:

• Active internet connection for speech recognition processing
• Browser Speech Recognition API support:

- Chrome 25+ - Edge 79+ - Safari 12+ - Opera 30+ - Firefox: Not Supported

• Microphone hardware connected to the system
• User microphone permission granted in browser settings
• HTTPS context for secure microphone access (recommended)