Add AI voice integration using Azure Speech Services (Includes UI rework)

[Tiberius-Sicae]

Microsoft Azure Speech Services Integration

• Neural Voices Support: Integrated Microsoft Cognitive Services Speech SDK and extended SpeechServiceConfiguration to support Azure API Keys and Regions.
• **Works with a free Speech Service subscription (details on how to create in Text-to-Speech section via link)
• Dynamic Synthesis: Built dynamic voice loading and PCM stream audio synthesis, routing Azure Neural voices through EDDI's existing SpeechFx effects processor (radio distortion, chorus, echo).
• High-Readability Dropdown: Formatted Azure and Windows voices in the voice dropdown to put friendly language names first (e.g., English (United States) Emma - Neural or Russian (Russia) Svetlana - Local). Also, filtered out redundant Windows "(Online) - Local" voices and sorted all available options alphabetically to simplify selection.
• Accessible Hyperlink: Added a How to get your Azure API Key hyperlink referencing the official setup guide. The link features WCAG AAA high-contrast colours (link blue in Light Mode, bright sky blue in Dark Mode) and permanent underlines for clear accessibility.
• Layout & Save Adjustments: Stacked the settings controls into three rows (API Key, Region, Save Button) to prevent clipping at the edge of the window, and configured immediate synchronous saves to disk to resolve settings loss.
• Volume & Rate Slider Support: Wrapped Azure speech synthesis requests in SSML <prosody> tags to ensure Azure Neural voices correctly honour the volume and speaking rate sliders configured in the Text-to-Speech section.

[Tiberius-Sicae]

Note: After someone else tested this (Cmdr Darkcyde - aka Moon-on-a-stick guy), the minimum-size window for EDDI cuts off the Azure area, and there are no scroll bars to find it again.

The first commit to this PR was an error (I am very new to GitHub). The second commit is the one that includes the Azure code and the UI update.