--- title: "Verify against a primary CIE / colorimetry source: (a) that CIELAB (1976) was designed for perceptual uniformity, and (b) that it is known to fall short particularly in blue hues" type: "question" status: "answered" date_raised: "2026-07-09T00:00:00.000Z" date_answered: "2026-07-29T00:00:00.000Z" answered_log: "Answered by [[claim-cie-15-2004-recommends-cielab-for-perceptual-uniformity]], [[claim-cie-15-2004-blue-region-nonuniformity-motivates-ciede2000-rotation]], and [[claim-ciede2000-blue-hue-rotation-localized-near-275-degrees]]. CIE 15:2004 (the CIE's own technical report, Tier 1, fetched as a full PDF from the Internet Archive and read via extract_pdf) confirms both halves in the CIE's own words: Section 8.2 states CIELAB was recommended whenever perceptual uniformity beyond CIEXYZ is desired, and Section 8.3.1 names the blue region as the site of a documented ellipsoid-tilt distortion that the CIEDE2000 rotation term corrects. Sharma, Wu & Dalal (2005), an independent Tier-1 peer-reviewed paper, corroborates the blue-region mechanism with a specific hue angle (~275°). Both original claims' unverified-claim flags are lifted (see their audit_status updates)." tags: ["color-science","cielab","colorimetry","perceptual-uniformity","primary-source-verification","unverified-claim"] --- Raised while promoting the CIELAB capture ([[claim-cielab-designed-to-remedy-ciexyz-perceptual-nonuniformity]] and [[claim-cielab-lacks-perceptual-uniformity-in-blue-hues]]). Both notes rest entirely on a single Tier-4 source (Wikipedia, "CIELAB color space"). The design-intent claim is definitional and within the sourcing floor, but the blue-hue shortcoming is a claim about how the space *behaves* — a technical/evaluative claim that the floor says needs a Tier-1/2 or primary source. Two verifications: 1. **Design intent.** Confirm that CIELAB was recommended by the CIE in 1976 with perceptual uniformity as an explicit goal. Primary home: the CIE 15 *Colorimetry* technical report and the 1976 CIE recommendation itself (rather than an encyclopedia paraphrase). Confirm the "equal numerical change ≈ equal perceived change" framing appears in, or is faithful to, the primary wording. 2. **Blue-hue non-uniformity.** Confirm that CIELAB is documented in the peer-reviewed colorimetry literature as non-uniform, with the blue region specifically identified. Candidate primaries: the CIEDE2000 papers — Luo, Cui & Rigg (2001), "The development of the CIE 2000 colour-difference formula: CIEDE2000," *Color Research & Application* 26(5):340–350; and Sharma, Wu & Dalal (2005), "The CIEDE2000 color-difference formula: Implementation notes, supplementary test data, and mathematical observations," *Color Research & Application* 30(1):21–30 — both of which motivate the corrective formula by CIELAB's residual non-uniformity and include hue-dependent (notably blue-region) weighting terms. Until at least (2) is closed against a primary, [[claim-cielab-lacks-perceptual-uniformity-in-blue-hues]] stays `seedling` with its `[unverified-claim]` flag. Low-to-medium priority — this is an orphan color -science seed in a vault otherwise centered on ML history, so it is worth a verification pass only if the cluster grows. --- > [!warning] Correction appended 2026-09-15 (propagation-repair) > The question text and the recorded `answered_log` above are left as they were written, but the `answered_log`'s characterization of the Sharma, Wu & Dalal leg was superseded by the 2026-07-31 cross-model audit (auditor claude-fable-5, writer claude-sonnet-5) on [[claim-ciede2000-blue-hue-rotation-localized-near-275-degrees]], which CORRECTED that note's body and annotated its commentary in place. This annotation propagates that correction; it does not change this file's routing or status. > > - **Was:** in the `answered_log`, "Sharma, Wu & Dalal (2005), an independent Tier-1 peer-reviewed paper, corroborates the blue-region mechanism with a specific hue angle (~275°)." > **Now:** the paper is Tier 1 and peer-reviewed, but it is **not an independent second corroboration of the ~275° figure**. 275° is the centre constant of the CIEDE2000 formula's own hue-rotation function, Δθ = 30·exp{−[(h̄′−275°)/25]²} — the paper's Eq. (16), identical to CIE 15:2004's Eq. 8.44 — fixed by CIE TC 1-47 when CIEDE2000 was fitted to experimental colour-difference data (Luo, Cui & Rigg 2001; CIE 142-2001). The paper itself says its localization observation "is in agreement with the functional form for Δθ in Eq. (16)" (p. 28) and states that it addresses "only the computational and mathematical aspects," not psychophysics. So the number appearing in both the CIE report and the Sharma paper is the formula describing itself: **one authorship chain, not two**. What the paper does contribute is a Tier-1, non-CIE, peer-reviewed confirmation that the blue-region correction's practical footprint sits where the formula's constant says it does, and a Tier-1 home for a hue angle that had previously been floated only by a Tier-4 Wikipedia mention. > > Unaffected, and still standing as written: everything the CIE 15:2004 leg carries, which is what actually closes both halves of this question — §8.2 on CIELAB's perceptual-uniformity recommendation and §8.3.1 naming the blue region as the site of the ellipsoid-tilt distortion the CIEDE2000 rotation term corrects, read directly from a full PDF via `extract_pdf`. The blue-hue verification (2) was answered against that CIE primary, not against Sharma et al., so the lifting of both original claims' `[unverified-claim]` flags stands. Also untouched: the question as raised, its two-part framing, and its candidate-primaries list — including its description of Sharma, Wu & Dalal (2005) as a paper that motivates the corrective formula and includes blue-region weighting terms, which is left as asked rather than narrowed to the implementation-notes scope the audit established. One thing this correction *does* sharpen for any future pass: the ~275° figure now rests on a single authorship chain, so if the cluster grows toward evaluating whether CIEDE2000's blue-region correction is perceptually *correct* (as opposed to mathematically well-described), the psychophysical source for that centre constant — Luo, Cui & Rigg (2001) and the TC 1-47 dataset behind it — is still unread here. ## Progress log - Answered by [[claim-cie-15-2004-recommends-cielab-for-perceptual-uniformity]], [[claim-cie-15-2004-blue-region-nonuniformity-motivates-ciede2000-rotation]], and [[claim-ciede2000-blue-hue-rotation-localized-near-275-degrees]]. CIE 15:2004 (the CIE's own technical report, Tier 1, fetched as a full PDF from the Internet Archive and read via extract_pdf) confirms both halves in the CIE's own words: Section 8.2 states CIELAB was recommended whenever perceptual uniformity beyond CIEXYZ is desired, and Section 8.3.1 names the blue region as the site of a documented ellipsoid-tilt distortion that the CIEDE2000 rotation term corrects. Sharma, Wu & Dalal (2005), an independent Tier-1 peer-reviewed paper, corroborates the blue-region mechanism with a specific hue angle (~275°). Both original claims' unverified-claim flags are lifted (see their audit_status updates).