Scores are drawn from seven trusted technical review sources: AVSForum, Blu-ray.com, Blu-Ray Authority, CriterionForum, DoBlu.com, High-Def Digest, The Digital Bits. A disc requires at least two independent sources before it is published. Titles with a single review so far appear as In Pipeline, provisional and clearly marked, until that second independent source confirms them.

Every source uses a different scale. We convert all ratings to a 0–10 baseline: stars ×2, letter grades mapped (A+=10 down to C+=7), percentages ÷10. Scores are then averaged across all available sources. A +0.2 DI bonus is applied when the 4K master is independently confirmed as native resolution. Scores are updated when new reviews are published. The detail panel shows the update date and reason when this occurs.

Reference ≥ 9.5	Gold standard. Demo-worthy transfers the enthusiast community regards as the best the format offers.
Excellent ≥ 9.0	Exceptional transfer. Buy with full confidence.
Recommended ≥ 8.0	Good transfer, worth owning. No meaningful reservations.
Notable ≥ 7.0	Competent transfer. Hidden by default, accessible via filter.
Below Threshold	Discs scoring below 7.0 are tracked but not published by default, accessible via the Below Threshold filter.
In Pipeline	Reviewed once but not yet confirmed by an independent second source. Score is provisional and shown muted, not counted in the published catalogue. Browsable via the In Pipeline filter.
Fake HDR	Discs with peak brightness below 400 nits and average below 120 nits. HDR grading does not use available headroom but score unaffected.

A disc's source sets a ceiling on how it can look, so a film scan and a native-digital master are not competing on the same terms. Each published disc is ranked only against others of its own source, film against film and digital against digital and its expanded panel shows where it lands as a percentile of that band plus its nearest-scoring peers. The headline score is unchanged; this is fair comparison within a source, not a re-grade. Source is set by capture format (photochemical or digital), with native large-format and IMAX film noted where applicable. Within an equal score, peers are ordered by review coverage then tighter consensus, so a better-supported disc sits above a thinly-reviewed one at the same number.

Consensus measures reviewer agreement via standard deviation of normalised scores: Strong Consensus (SD ≤ 0.5, 4+ sources) → Consensus → Soft Consensus → Disputed → Contested (SD > 1.6). Confidence measures coverage: how many of the possible sources have reviewed the disc. Both are shown independently because a high-confidence disputed score tells a different story than a provisional consensus one. Confidence is represented by the pip indicators (●) shown on each disc card.

The same film can appear as more than one edition, for example a studio release and a later boutique or Dolby Vision edition. Each disc is scored independently on its own transfer and review sources, so scores can differ between editions of the same title. A higher score reflects a better-reviewed transfer, not a different film.

When a disc has a matching IMDb ID in the streaming database, a Streaming Comparison panel appears in the expanded detail view. Data is sourced from the streaming quality database.

Streaming availability is not actively tracked. The Streaming panel appears only where data has been captured. Its absence does not indicate that a title is unavailable on streaming services.

For the complete streaming quality database with PRO metrics, visit Streaming Reference

Backfilling high-demand titles not yet in the database: Other major franchise releases are being evaluated and will be added as reviews are verified. Priority is based on community requests and transfer significance.

Expanding source coverage for low and moderate confidence discs: titles currently scored from 2–3 sources will be revisited as additional reviewer coverage becomes available, improving consensus accuracy across the database.

An attestation of how a film was authored to look and whether a given 4K UHD disc is faithful to that. It is a separate axis from the quality score. A disc can be reference quality and still not earn a Reference Print, because looking spectacular and being faithful to the authored work are different questions. The two verdicts are shown side by side and can disagree.

Every Reference Print names the reference it is judged against and that anchor sets the highest tier a disc can reach. Film print or photochemical timing: the authored look is a colour-timed film print, so the disc grade is a translation to a digital target and HDR is a target the film was never graded for. Theatrical digital intermediate or DCP: the authored look exists as a concrete digital master and that file is the reference. Creative-supervised restoration: a restoration approved by the original cinematographer or director stands as an attested reconstruction of intent. Lost or reconstructed: no clean authored reference survives and these titles cannot exceed Provisional because faithfulness cannot be proven.

Intent is established from primary creative sources, ranked closest to intent first: statements by the cinematographer, director or colourist on the intended look, documentation of a creative-supervised restoration, trade record such as American Cinematographer, then commentary tracks and release booklets. Every claim of intent carries a resolvable source and a claim that cannot be sourced does not count toward a higher tier. Research may be assisted by automated tools, but the attestation is made and signed by a named, ISF-certified human and no tier is assigned on machine judgement alone.

Attested	Intended look documented from named creative sources and the disc judged a faithful representation of it. Disclosed minor caveats are permitted and stated. The highest tier, because the original master is not held: documented intent and judged fidelity is the standard.
Inferred	Provenance clean and uncontested with no competing version, but no direct statement of intended look on record. Fidelity is reasoned from pedigree and named as an inference.
Provisional	Intended look cannot be established, or fidelity cannot be confirmed. The title stays visible and flagged but is not eligible for Reference Print status. Declining to certify is a normal outcome of the method, not a failure of it.

Faithful	The disc follows the authored grade. The default for Attested and Inferred titles.
Revised	The disc follows a later grade the original creative themselves sanctioned. A legitimate alternate, flagged so the difference from the original is visible.
Divergent	The disc departs from a known intent with no creative behind the change. Recorded as a fault.

Only a curated subset of titles is reviewed for a Reference Print. Most titles remain at quality-verdict level. This is deliberate. The value of the standard is trust, not coverage.