Skip to main content

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:

ResolutionBase ScoreNotes
2160p (4K)100Highest base score
1080p80Standard HD
720p50Reduced detail
480p20Standard definition
Unknown0No 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:

SourceBonusDescription
BluRay+40Physical disc rip, best quality
WEB-DL+30Direct streaming service rip
HDTV+10Captured from broadcast
DVD+5Standard definition disc
CAM/TS-1000Theater 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:

CodecBonusNotes
AV1+25Latest codec, best efficiency
H.265/HEVC+20Good compression, widely supported
H.264/x264+0Baseline, no bonus
MPEG-2-10Outdated, inefficient
DivX/XviD-15Legacy 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 FormatBonusChannelsDescription
Dolby Atmos+257.1+Object-based 3D audio
DTS:X+257.1+Object-based audio
DTS-HD MA+205.1/7.1Lossless surround
TrueHD+155.1/7.1Dolby lossless
DD+ (E-AC3)+105.1Dolby Digital Plus
DTS+85.1Standard DTS
AC3+55.1Dolby Digital
AAC+32.0/5.1Efficient codec
MP3+02.0Basic 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 FormatBonusDescription
Dolby Vision+30Best HDR, dynamic metadata
HDR10++20Dynamic metadata, open standard
HDR10+15Static metadata, widely supported
HLG+10Broadcast HDR
SDR+0Standard 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 TypeBonusExamples
Tier 1+15SPARKS, DON, HDX, NTb
Tier 2+10EVO, RARBG, YIFY
Tier 3+5Various scene groups
Unknown+0No reputation data
Avoided-20Known 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

  1. Query Indexers - Search all enabled indexers
  2. Parse Results - Extract metadata from release names
  3. Calculate Scores - Apply scoring algorithm
  4. Filter Results - Remove blocklisted/low-quality
  5. Select Best - Choose highest scoring release
  6. Send to Download - Add to download queue

Upgrade Decision

When considering upgrades:

  1. Compare Scores - New vs existing
  2. Check Cutoff - Has cutoff been reached?
  3. Verify Monitoring - Is item monitored?
  4. Apply Delays - Wait for better releases?
  5. Queue Upgrade - Download better version

Profile-Specific Scoring

Quality Profile Cutoffs

Each profile defines when to stop upgrading:

ProfileTargetCutoffExample
Quality4K HDR2160p BluRayUpgrade until 4K BluRay
Balanced1080p1080p BluRayUpgrade until 1080p BluRay
Compact720p720p WEB-DLUpgrade 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:

  1. Higher quality failed custom format rules
  2. Higher quality is blocklisted
  3. Higher quality indexer is down
  4. Delay profile waiting for better releases
  5. Cutoff already reached

Debugging:

  • Check search results manually
  • Review custom format scores
  • Verify indexer status
  • Check delay profile settings

Custom Format Not Applied

Check:

  1. Format is assigned to profile
  2. Condition syntax is correct
  3. Release name matches pattern
  4. 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

See Also