/// The supported output formats for encoded transcription segments.
///
/// Why this exists:
/// - A single, strongly-typed representation of output formats across the CLI and library code.
/// - Using an enum avoids stringly-typed conditionals and keeps format
/// selection explicit and discoverable.
///
/// Integration notes:
/// - When the `cli` feature is enabled, derives `clap::ValueEnum` so the enum can be used
/// directly as a CLI flag.
/// - Each variant maps to a concrete `SegmentEncoder` implementation.