Integrate WhisperCpp and VoiceVox for speech recognition and synthesis

[74th]

This pull request adds support for using whisper.cpp as a speech-to-text backend in the StackChan server, alongside the existing Google Cloud integration. The main changes introduce a new WhisperCppSpeechToText class, update the speech recognition module to expose it, and modify the example app to allow dynamic selection of the speech recognizer via environment variables.

New speech-to-text backend integration:

• Added a new WhisperCppSpeechToText class in stackchan_server/speech_recognition/whisper_cpp.py that runs the whisper.cpp CLI for local speech recognition, including support for VAD (voice activity detection) and various configuration options.
• Updated stackchan_server/speech_recognition/__init__.py to import and export WhisperCppSpeechToText, making it available for use throughout the codebase.

Application configuration improvements:

• Modified example_apps/echo.py to allow selecting the speech recognizer backend via the STACKCHAN_WHISPER_MODEL environment variable, enabling the use of whisper.cpp if configured.
• Added imports for WhisperCppSpeechToText and VoiceVoxSpeechSynthesizer in example_apps/echo.py to support the new backend and synthesizer selection.…for enhanced speech recognition and synthesis