Add Microsoft Azure Cognitive Services Text-to-Speech

[stone-alex]

Description

Add Microsoft Azure Cognitive Services Text-to-Speech as a second TTS backend.
Provider is auto-detected from API key format (same mechanism as STT).

Why

• For Players who want Azure neural voices
• Zero config for users
• Full backwards compatibility with existing GoogleTTS

Acceptance criteria

• Implement elite.intel.ai.mouth.MouthInterface in new class AzureTTS
• Singleton, zero DI – exact same pattern as GoogleTTS
• Dynamically loaded via elite.intel.ai.ApiFactory using KeyDetector.detectProvider(apiKey, "TTS")
• Full support for elite.intel.ai.mouth.AiVoices enum with on-the-fly voice switching
• Properly fire IsSpeakingEvent(true/false) via EventBusManager to block STT while speaking
• No Python, no JNI, no native packages → pure Azure Java SDK only
• Works on Windows and Linux
• Builds and runs with ./gradlew run and fat-jar
• Provide temporary Azure key to maintainer for CI testing

Useful links

• Azure Speech SDK Java: https://central.sonatype.com/artifact/com.microsoft.cognitiveservices.speech/client-sdk
• TTS quickstart: https://github.com/Azure-Samples/cognitive-services-speech-sdk/blob/master/quickstart/java/jre/text-to-speech/src/TextToSpeech.java

Difficulty: Medium (mirror GoogleTTS + voice list handling)