Quality Scoring
This document explains how Cinephage calculates quality scores to select the best releases for your library.
Overview
Unlike simple "highest resolution wins" systems, Cinephage uses a sophisticated scoring algorithm that considers 50+ factors. This approach provides better results by accounting for:
- Resolution (but not resolution alone)
- Source quality (BluRay vs WEB-DL vs HDTV)
- Codec efficiency (H.265 vs H.264)
- Audio quality (Atmos vs DTS-HD vs AAC)
- HDR formats (Dolby Vision vs HDR10)
- Release group reputation
- Your personal preferences (custom formats)
Scoring Formula
The total score is calculated as:
Total Score = Base Score + Source Bonus + Codec Bonus + Audio Bonus + HDR Bonus + Custom Format Scores
Each component contributes to the final decision.
Base Score (Resolution)
Resolution provides the foundation score:
| Resolution | Base Score | Notes |
|---|---|---|
| 2160p (4K) | 100 | Highest base score |
| 1080p | 80 | Standard HD |
| 720p | 50 | Reduced detail |
| 480p | 20 | Standard definition |
| Unknown | 0 | No resolution detected |
Why Not Resolution Alone?
A 1080p BluRay (80 + 40 = 120) scores higher than a 2160p HDTV (100 + 10 = 110) because source quality matters significantly.
Source Quality Bonus
The original source affects quality dramatically:
| Source | Bonus | Description |
|---|---|---|
| BluRay | +40 | Physical disc rip, best quality |
| WEB-DL | +30 | Direct streaming service rip |
| HDTV | +10 | Captured from broadcast |
| DVD | +5 | Standard definition disc |
| CAM/TS | -1000 | Theater recording (effectively blocked) |
Why Source Matters:
- BluRay - Highest bitrate, no compression artifacts
- WEB-DL - Good quality from streaming services
- HDTV - Variable quality, may have logos/commercials
- CAM/TS - Poor quality, handheld camera recordings
Source Detection
Cinephage detects sources from release names:
Movie.2023.1080p.BluRay.x264-SPARKS → BluRay (+40)
Movie.2023.1080p.WEB-DL.DDP5.1-NTb → WEB-DL (+30)
Movie.2023.1080p.HDTV.x264-LOL → HDTV (+10)
Movie.2023.HDTS.x264-EVO → TS (-1000)
Codec Efficiency Bonus
Modern codecs provide better quality at smaller sizes:
| Codec | Bonus | Notes |
|---|---|---|
| AV1 | +25 | Latest codec, best efficiency |
| H.265/HEVC | +20 | Good compression, widely supported |
| H.264/x264 | +0 | Baseline, no bonus |
| MPEG-2 | -10 | Outdated, inefficient |
| DivX/XviD | -15 | Legacy codecs |
Codec Efficiency:
- AV1 - 50% smaller than H.264 at same quality
- H.265 - 40% smaller than H.264
- H.264 - Baseline efficiency
- Older codecs - Larger files, lower quality
Codec Detection
Movie.2023.1080p.BluRay.AV1-GRP → AV1 (+25)
Movie.2023.1080p.BluRay.x265-SPARKS → H.265 (+20)
Movie.2023.1080p.BluRay.x264-EVO → H.264 (+0)
Movie.2023.DVDRip.DivX-OLD → DivX (-15)
Audio Quality Bonus
Audio quality significantly impacts the viewing experience:
| Audio Format | Bonus | Channels | Description |
|---|---|---|---|
| Dolby Atmos | +25 | 7.1+ | Object-based 3D audio |
| DTS:X | +25 | 7.1+ | Object-based audio |
| DTS-HD MA | +20 | 5.1/7.1 | Lossless surround |
| TrueHD | +15 | 5.1/7.1 | Dolby lossless |
| DD+ (E-AC3) | +10 | 5.1 | Dolby Digital Plus |
| DTS | +8 | 5.1 | Standard DTS |
| AC3 | +5 | 5.1 | Dolby Digital |
| AAC | +3 | 2.0/5.1 | Efficient codec |
| MP3 | +0 | 2.0 | Basic audio |
Audio Considerations:
- Lossless (DTS-HD MA, TrueHD) - Perfect quality, large files
- High-quality lossy (DD+, DTS) - Excellent quality, smaller
- Standard (AC3, AAC) - Good quality, efficient
- Stereo (AAC) - Sufficient for dialogue
Audio Detection
Movie.2023.2160p.BluRay.Atmos.TrueHD7.1 → Atmos (+25)
Movie.2023.1080p.BluRay.DTS-HD.MA.5.1 → DTS-HD MA (+20)
Movie.2023.1080p.BluRay.DD5.1-SPARKS → AC3 (+5)
Movie.2023.1080p.WEB-DL.AAC2.0-NTb → AAC (+3)
HDR Format Bonus
High Dynamic Range provides better color and contrast:
| HDR Format | Bonus | Description |
|---|---|---|
| Dolby Vision | +30 | Best HDR, dynamic metadata |
| HDR10+ | +20 | Dynamic metadata, open standard |
| HDR10 | +15 | Static metadata, widely supported |
| HLG | +10 | Broadcast HDR |
| SDR | +0 | Standard dynamic range |
HDR Considerations:
- Dolby Vision - Best quality, requires compatible display
- HDR10+ - Similar to Dolby Vision, Samsung-focused
- HDR10 - Good HDR, static metadata
- SDR - No HDR, compatible with all displays
HDR Detection
Movie.2023.2160p.BluRay.DV.HDR10+.x265 → Dolby Vision + HDR10+ (+50)
Movie.2023.2160p.BluRay.HDR10.x265 → HDR10 (+15)
Movie.2023.1080p.BluRay.x264 → SDR (+0)
Release Group Bonus
Some release groups are known for quality:
| Group Type | Bonus | Examples |
|---|---|---|
| Tier 1 | +15 | SPARKS, DON, HDX, NTb |
| Tier 2 | +10 | EVO, RARBG, YIFY |
| Tier 3 | +5 | Various scene groups |
| Unknown | +0 | No reputation data |
| Avoided | -20 | Known poor quality |
Why Groups Matter:
- Consistent quality
- Proper encoding settings
- Reliable sources
- No fake or mislabeled releases
Note: Group scoring is subjective and configurable via custom formats.
Custom Format Scoring
Custom formats allow you to define personal preferences:
Creating Scoring Rules
Each custom format has a score that adds to or subtracts from the total:
# Boost HEVC releases
Name: 'x265 Boost'
Score: +20
Conditions:
- Contains: 'x265'
- OR Contains: 'HEVC'
# Block certain content
Name: 'Block 3D'
Score: -1000
Conditions:
- Contains: '3D'
- OR Contains: 'HSBS'
Custom Format Examples
Prefer Specific Resolution:
Name: '1080p Preferred'
Score: +10
Conditions:
- Contains: '1080p'
- Does NOT contain: '2160p'
Require Minimum Size:
Name: 'Minimum Quality'
Score: -1000
Conditions:
- Size less than: 1GB
- Contains: '1080p'
Prefer Release Groups:
Name: 'Trusted Groups'
Score: +15
Conditions:
- Contains: '-SPARKS'
- OR Contains: '-DON'
- OR Contains: '-NTb'
Complete Scoring Examples
Example 1: High-Quality Movie
Release: Inception.2010.2160p.UHD.BluRay.DV.HDR10+.Atmos.TrueHD7.1.x265-SPARKS
Score Breakdown:
Base (2160p): 100
Source (BluRay): +40
Codec (x265): +20
Audio (Atmos): +25
HDR (DV + HDR10+): +45
Group (SPARKS): +15
─────────────────────────
Total: 245
Result: Excellent quality, will be selected if available.
Example 2: Balanced Quality
Release: Movie.2023.1080p.BluRay.DTS-HD.MA.5.1.x264-EVO
Score Breakdown:
Base (1080p): 80
Source (BluRay): +40
Codec (x264): +0
Audio (DTS-HD MA): +20
HDR (None): +0
Group (EVO): +10
─────────────────────────
Total: 150
Result: Good quality, balanced file size.
Example 3: Low Quality (Blocked)
Release: Movie.2023.HDTS.x264-Unknown
Score Breakdown:
Base (Unknown): 0
Source (TS): -1000
Codec (x264): +0
Audio (Unknown): +0
HDR (None): +0
Group (Unknown): +0
─────────────────────────
Total: -1000
Result: Blocked by negative score, will not be selected.
Example 4: Custom Format Impact
Release: Movie.2023.1080p.WEB-DL.x265.DD5.1-NTb
Base Score:
Base (1080p): 80
Source (WEB-DL): +30
Codec (x265): +20
Audio (DD): +5
─────────────────────────
Subtotal: 135
With Custom Formats:
Subtotal: 135
"x265 Boost": +20
"WEB-DL Preferred": +10
"NTb Bonus": +5
─────────────────────────
Total: 170
Scoring in Practice
Search Process
- Query Indexers - Search all enabled indexers
- Parse Results - Extract metadata from release names
- Calculate Scores - Apply scoring algorithm
- Filter Results - Remove blocklisted/low-quality
- Select Best - Choose highest scoring release
- Send to Download - Add to download queue
Upgrade Decision
When considering upgrades:
- Compare Scores - New vs existing
- Check Cutoff - Has cutoff been reached?
- Verify Monitoring - Is item monitored?
- Apply Delays - Wait for better releases?
- Queue Upgrade - Download better version
Profile-Specific Scoring
Quality Profile Cutoffs
Each profile defines when to stop upgrading:
| Profile | Target | Cutoff | Example |
|---|---|---|---|
| Quality | 4K HDR | 2160p BluRay | Upgrade until 4K BluRay |
| Balanced | 1080p | 1080p BluRay | Upgrade until 1080p BluRay |
| Compact | 720p | 720p WEB-DL | Upgrade until 720p WEB-DL |
Score Thresholds
Profiles can define minimum scores:
Minimum Score: 100
- Scores below 100 are rejected
- Ensures minimum quality standard
Troubleshooting Scoring
Why Was Lower Quality Selected?
Possible Reasons:
- Higher quality failed custom format rules
- Higher quality is blocklisted
- Higher quality indexer is down
- Delay profile waiting for better releases
- Cutoff already reached
Debugging:
- Check search results manually
- Review custom format scores
- Verify indexer status
- Check delay profile settings
Custom Format Not Applied
Check:
- Format is assigned to profile
- Condition syntax is correct
- Release name matches pattern
- Format is enabled
Unexpected Scores
Verify:
- Quality profile assigned correctly
- Custom format scores are correct
- No conflicting rules
- Release name parsing correctly
Best Practices
Start with Defaults
- Use built-in profiles initially
- Understand scoring before customizing
- Test custom formats thoroughly
Document Your Rules
- Keep notes on why scores are set
- Name formats descriptively
- Review periodically
Test with Real Releases
- Use actual release names
- Verify score calculations
- Check upgrade behavior
Balance Quality and Size
- Higher score ≠ always better
- Consider storage constraints
- Set realistic cutoffs