Update audio extraction to use file-based streaming
instead of in-memory buffers to support processing
video files larger than 2GB.
**Changes:**
- Modified `IAudioExtractor.ExtractMp3FromMp4Async()`
to accept an output audio stream parameter instead
of creating a MemoryStream internally
- Updated `AudioExtractor` to write directly to the
provided audio stream, eliminating the need to load
entire video into memory
- Enhanced `FFMpegService` to detect FileStream inputs
and use direct file-to-file processing when possible,
falling back to pipe-based streaming for other stream types
- Refactored `DefaultCommand` to create output directory
and audio file stream before extraction, enabling direct
file-based audio extraction
- Updated all unit and integration tests to accommodate the
new audio stream parameter
This change prevents `OutOfMemoryException` errors when processing
large video files by streaming data directly to disk rather than
buffering in memory. Additionally, it removes the possibility of
overflowing the 2GB size limit that `MemoryStream` has, which
would result in `IOException` errors when trying to write or
copy to it.