add: add project specification

This commit is contained in:
Esa Kataja
2026-04-04 18:05:26 +03:00
parent 3273beb574
commit 25ea35fc26
+116
View File
@@ -0,0 +1,116 @@
# videnc-vibe Specification
## Overview
FFmpeg-based video transcoder for DVD/Blu-ray to SVT-AV1 with automatic metadata fetching.
## Configuration
Config path: `-c` flag or `~/.config/videnc-vibe/config.yaml`
```yaml
omdb_api_key: "your-key-here"
encoding:
dvd:
crf: 30
preset: 2
bluray:
crf: 29
preset: 3
paths:
input: "./input"
output: "./output"
originals: "./originals"
failed: "./failed"
```
## Flags
- `-d` : Delete original after successful encode (default: move to originals/)
- `-c` : Config file path (default: ~/.config/videnc-vibe/config.yaml)
## Input Parsing
Regex patterns:
- Movie: `tt\d+` → extract IMDB ID
- Series: `tvm\d+` + `s\d\de\d\d` → extract TVmaze ID + season + episode (both required)
## Media Detection
| Condition | Media Type |
|-----------|------------|
| < 600k pixels/frame | DVD |
| ≥ 600k pixels/frame | Blu-ray |
## Encoding Parameters
### Video
- Codec: SVT-AV1
- Pixel format: yuv420p10le
- Scaling: zscale filter with spline36
- Preset: 2 (DVD) / 3 (Blu-ray) - configurable
- CRF: 30 (DVD) / 29 (Blu-ray) - configurable
- SVT-AV1 params: film-grain=10,film-grain-denoise=1,scd=1,qm-min=4,qm-max=15
- Deinterlace: bwdif mode 0 if interlaced
### Audio
- Pipeline: Extract → WAV (48kHz) → Opus (48kHz)
- FFmpeg libopus is NOT used (poor quality)
## Matroska Tags
### Movie
| Tag | Source |
|-----|--------|
| TITLE | OMDb response |
| DATE_RELEASED | OMDb response (ISO 8601: YYYY-MM-DD) |
| IMDBID | Input filename |
| ORIGINAL_MEDIA_TYPE | DVD or Blu-ray |
### Series
| Tag | Source |
|-----|--------|
| TITLE | TVmaze response (episode name) |
| COLLECTION | TVmaze response (show name) |
| SEASON | Parsed from filename |
| EPISODE | Parsed from filename |
| DATE_RELEASED | TVmaze response (ISO 8601: YYYY-MM-DD) |
| IMDBID | TVmaze response |
| TVMAZE_ID | Custom tag |
| ORIGINAL_MEDIA_TYPE | DVD or Blu-ray |
## Stream Metadata
- Preserve language tags per stream
## Output Naming
| Scenario | Filename |
|----------|----------|
| Movie with metadata | `{Title}.{IMDBID}.mkv` |
| Series with metadata | `{ShowName}.S{season}E{episode}.mkv` |
| No ID match / API fail | `{random}.nometadata.mkv` |
## File Handling
| Outcome | Input file | Output file |
|---------|------------|-------------|
| Success | `-d`: delete, else: move to originals/ | move to output/ |
| Failure | move to failed/ | move to failed/ |
## Logging
- info.log: Human-readable info level
- error.log: Human-readable error level
- structured.json: JSON structured logs
## Polling
- Scan input folder every 10-30 seconds for `.mkv` files
## Commit Message Format
Use Conventional Commits style:
- `add:` for new features
- `fix:` for bug fixes
- `refactor:` for code restructuring
- `chore:` for maintenance tasks
Example: "add: add command line argument support"