Changelog

v2.8

✅ HFR (High Frame Rate) Management: new --limit-fps / --no-limit-fps options to handle videos >30 fps. In series mode (default), HFR videos are limited to 29.97 fps to optimize size. In film/adaptatif mode, bitrate is proportionally adjusted to framerate.
✅ HFR option display: LIMIT_FPS option appears in active parameters (terminal) and in run_started Discord notification.
✅ Intelligent VMAF skip: VMAF score is automatically ignored when framerate is modified (invalid comparison).
✅ HFR tests: 26 unit tests + 2 e2e tests for HFR management.
✅ Centralized constants: new lib/constants.sh module grouping "magic numbers" (adaptive mode, audio thresholds, Discord parameters, HFR thresholds). All constants are overridable via environment variables.
✅ Documentation robustness: systematic documentation of set -euo pipefail in 38 lib/ modules.
✅ Reinforced tests: new unit tests for VIDEO_EQUIV_QUALITY_CAP, _clamp_min, _clamp_max, _min3.
✅ Bitrate factorization: _translate_bitrate_by_efficiency() function centralized in codec_profiles.sh (audio + video).
✅ Test helpers: refactoring of test_helper.bash (standardized variables, intelligent skip, modularity).
✅ Documentation: CONFIG.md enriched (constants section), SMART_CODEC.md updated (audio ranks, bitrate translation).

✅ Queue / FIFO robustness: avoids blocking when no files are processable (empty entries, missing files, invalid queue) + explicit exit if source is excluded.
✅ Discord notifications (optional): start/pause(off-peak)/end messages, more robust payload + opt-in debug mode, dedicated docs + tests.
✅ Stabilized adaptive mode: calculated budgets are actually applied to encoding (exports) + SVT-AV1 cap in CRF (mbr=) and opt-in SVT debug logs.
✅ Video "smart codec": codec-aware skip threshold + "no downgrade" policy (AV1 source re-encoded in AV1 if necessary) + "equivalent quality" cap on less efficient sources.
✅ UX/CLI: better message consistency (skip/early errors) and --quiet option documented in guides.
✅ Dev tooling: make lint (ShellCheck) + warnings cleanup, improves maintainability.
✅ Tests & docs: Bats additions/reinforcements (queue, conversion, notifs) + troubleshooting and aligned guides.

✅ Lossless/premium audio: --no-lossless option to force conversion of DTS/DTS-HD, TrueHD, FLAC tracks (disables "premium" passthrough)
✅ Multichannel audio: finalized rules (downmix 7.1 → 5.1, EAC3 by default, multichannel Opus via -a opus, multichannel AAC only with --force-audio)
✅ Audio refactor: clear separation between decision (lib/audio_decision.sh) and FFmpeg parameters/layout (lib/audio_params.sh)
✅ "Clean code light" refactor: internal simplification of large functions (audio/video/VMAF/suffix) and FFmpeg command building via argument arrays (no expected UX/CLI change)
✅ Tests & docs: new Bats multichannel tests + aligned docs (README + docs)

✅ Adaptive mode: adaptive bitrate based on complexity analysis
✅ Size filter: --min-size option to filter index/queue (useful in film mode)
✅ Multichannel audio: layout normalization (stereo / 5.1) + preservation logic according to mode
✅ Windows / Git Bash: path/CRLF normalization and better robustness with special characters
✅ VMAF & UX: display adjustments + subsampling parameter, improved messages and counters
✅ Tests: refactor and optimizations for faster and more robust execution
✅ Audio refactor: extraction of "smart codec" logic into lib/audio_decision.sh (and lib/audio_params.sh refocused on FFmpeg/layout)

✅ Multi-codec audio: -a/--audio option to choose AAC, AC3, Opus or copy
✅ Anti-upscaling logic: only converts audio if real gain (>20%)
✅ Optimized bitrates: AAC 160k, AC3 384k, Opus 128k
✅ Audio suffix in filename (_aac, _opus, etc.)
✅ Audio refactoring: new dedicated audio_params.sh module (FFmpeg parameters/layout)
✅ Colored help with highlighted options
✅ Video codec display in active parameters