Crate oximedia_caption_gen 

oximedia-caption-gen

Advanced caption and subtitle generation for the OxiMedia Sovereign Media Framework.

This crate provides speech-to-caption alignment with frame-accurate timing, greedy and optimal (Knuth-Plass DP) line-breaking algorithms, WCAG 2.1 accessibility compliance checking, and speaker diarization metadata with crosstalk detection — all in pure Rust.

Modules

• alignment — Word timestamps, transcript segments, segment merging/splitting, frame alignment, and caption block construction.
• autopunct — Deterministic auto-punctuation and sentence capitalisation.
• burn_in — Burned-in subtitle rendering onto raw RGBA video frames using a built-in 8×12 bitmap font.
• caption_diff — Compare two caption tracks and report differences.
• caption_format_adapter — Serialize caption tracks to SRT/VTT/TTML.
• caption_style_guide — Style guide rule enforcement over caption tracks.
• caption_timing_adjuster — Shift, stretch, snap, and EDL-remap caption timestamps.
• diarization — Speaker metadata, turn merging, per-speaker statistics, crosstalk detection, voice activity ratio, and speaker-to-caption assignment.
• forced_narrative — Forced narrative (FN) and SDH subtitle detection and classification.
• language_detect — Byte-trigram language detection for locale-aware line-breaking.
• line_breaking — Greedy and optimal line-breaking, reading-speed helpers (CPS), and line-balance optimisation.
• multi_language — Bilingual caption layout (primary + secondary language).
• multi_language_sync — Anchor-point synchronisation of multi-language caption tracks.
• multilang — Multi-language subtitle support with ISO 639-1 validated language codes, SRT export, and cross-language timing merge.
• phoneme_timing — Phoneme-level timing estimation from word timestamps.
• profanity — Configurable profanity filter for caption text.
• punctuation_restoration — Rule-based punctuation restoration for raw ASR output.
• reading_speed — Caption reading-speed validation (WPS-based).
• style_generator — Font size, position, and colour suggestions based on video frame analysis.
• style_presets — Ready-made caption style configs (Netflix, BBC, WCAG).
• translate — Stub subtitle translation pipeline.
• wcag — WCAG 2.1 compliance checks (1.2.2, 1.2.4, 1.2.6), reading speed validation, minimum display duration, gap detection, and compliance scoring.

Re-exports

pub use alignment::align_to_frames;

pub use alignment::build_caption_blocks;

pub use alignment::merge_short_segments;

pub use alignment::split_long_segments;

pub use alignment::AlignmentError;

pub use alignment::CaptionBlock;

pub use alignment::CaptionPosition;

pub use alignment::TranscriptSegment;

pub use alignment::WordTimestamp;

pub use diarization::assign_speakers_to_blocks;

pub use diarization::dominant_speaker;

pub use diarization::format_speaker_label;

pub use diarization::merge_consecutive_turns;

pub use diarization::speaker_stats;

pub use diarization::voice_activity_ratio;

pub use diarization::CrosstalkDetector;

pub use diarization::DiarizationResult;

pub use diarization::Speaker;

pub use diarization::SpeakerGender;

pub use diarization::SpeakerStats;

pub use diarization::SpeakerTurn;

pub use line_breaking::compute_cps;

pub use line_breaking::greedy_break;

pub use line_breaking::optimal_break;

pub use line_breaking::reading_speed_ok;

pub use line_breaking::rebalance_lines;

pub use line_breaking::LineBalance;

pub use line_breaking::LineBreakAlgorithm;

pub use line_breaking::LineBreakConfig;

pub use wcag::check_caption_coverage;

pub use wcag::check_cps;

pub use wcag::check_live_latency;

pub use wcag::check_min_duration;

pub use wcag::check_sign_language;

pub use wcag::compliance_score;

pub use wcag::run_all_checks;

pub use wcag::WcagChecker;

pub use wcag::WcagLevel;

pub use wcag::WcagViolation;

pub use burn_in::BurnInConfig;

pub use burn_in::SubtitleBurnIn;

pub use burn_in::SubtitlePosition;

pub use multilang::CaptionEntry;

pub use multilang::LanguageCode;

pub use multilang::MultiLangCaption;

pub use multilang::MultiLangCaptionBuilder;

Modules

alignment

Speech-to-caption alignment: word timestamps, segment merging/splitting, frame-accurate caption block construction.

autopunct

Auto-punctuation for caption text.

burn_in

Burned-in subtitle rendering onto raw RGBA video frames.

caption_diff

Caption diff: compare two caption tracks and report differences.

caption_format_adapter

Caption format adapter: serialize CaptionBlock tracks to SRT, WebVTT, and TTML output strings.

caption_style_guide

Caption style guide rule enforcement.

caption_timing_adjuster

Caption timing adjustment: shift and stretch caption timecodes to match edited or re-timed video content.

diarization

Speaker diarization metadata: speaker turns, statistics, crosstalk detection, and assigning speakers to caption blocks.

forced_narrative

Forced narrative (FN) and SDH (Subtitles for the Deaf and Hard-of-hearing) subtitle detection and classification.

language_detect

Language detection for caption transcript text.

line_breaking

Caption line-breaking algorithms: greedy, optimal (Knuth-Plass-inspired DP), reading-speed helpers, and line-balance optimisation.

multi_language

Multi-language / bilingual caption layout.

multi_language_sync

Multi-language caption synchronization.

multilang

Multi-language subtitle support with ISO 639-1 validated language codes.

phoneme_timing

Phoneme-level timing alignment for caption display.

profanity

Profanity filter for caption text.

punctuation_restoration

Punctuation restoration for raw ASR transcript output.

reading_speed

Caption reading-speed validation.

style_generator

Style suggestion engine for caption rendering.

style_presets

Caption style presets for popular broadcast standards.

translate

Subtitle translation pipeline.

wcag

WCAG 2.1 accessibility compliance checks for caption blocks.

Enums

CaptionGenError

Errors produced by caption generation operations.

Type Aliases

CaptionGenResult

Result alias used by caption-generation APIs.