[android] Implement proper TTS voice selection

There are several issues with the existing TTS implementation on Android:

• The API is old and doesn't leverage more than one installed TTS engine properly (e.g. Samsung + Google TTS)
• Doesn't always work reliably (e.g. allows to select voices that don't work, for example, when offline/voices were not downloaded yet)
• Doesn't support subvoice selection
• All TTS engine system changes should be detected in OM without restarting the app
• Review and retest TTS engines mentioned https://organicmaps.app/faq/voice/text-to-speech-tts-and-voice-directions-on-android/ and update the FAQ if necessary with the most actual information

It would be great to review the existing code, update it to use a more modern API available from Android 5+ if necessary, and properly support more than one TTS engine.

Example:

• Scan voices/TTS translations available in OM
• Scan system TTS engines, their languages and subvoices
• Match both and show to a user a list of system voices, mentioning their TTS engine, their language (and subtype if necessary)

Something like:

• English (female II, Google TTS)
• English (UK Alex, Samsung TTS)
• …