1.10.0 — Resizable plots + Gamut Slice zoom

What's new in 1.10.0

Resizable plot height (aspect ratio)

Both the 3D Plot and Gamut Slice now have a draggable handle on the bottom edge of the graphing area — the same 3-dots grip (with hover/drag highlight) as the File Select pane. Drag it to make a plot taller or shorter to suit your screen and the data. The chosen height is sticky: it's remembered per plot across sessions.

Auto-scaling of the data is unchanged — only the canvas box resizes.

Gamut Slice scroll-wheel zoom

• Scroll to zoom the Gamut Slice in and out with the mouse wheel.
• A new Reset zoom button restores the default auto-scaled view.

Zooming rescales the auto-scaled range about its centre, so the data still drives the bounds — the wheel just lets you get closer.

Localization

The new Reset zoom label is translated across all 12 in-page languages and the translation spreadsheets.

1.9.2 — iccDEV memory-leak & hardening rebuild

Maintenance release — rebuilds the bundled ICC viewer WASM against a clean iccDEV master, picking up the upstream memory-leak and input-validation hardening fixes merged since 1.9.1. No app-facing behaviour change — the profile viewer parses and displays the same data on a hardened IccProfLib underneath.

Upstream IccProfLib fixes now included:

• CIccXform::Create overload leak / double-free via bOwnsProfile (#1308 / #1347)
• CIccCmm::CheckPCSConnections CIccPcsXform leak on Connect failure (#1337 / #1338)
• Spectral-profile transform-creation fixes (#1335)
• CIccCmmSearch::AddXform NamedColor profile leak (#1332 / #1333)
• AddXform ownership / delete-on-error contract (#1327)
• Decode icStatusCMM names in CMM error messages (#1322 / #1323 / #1324)

1.9.1 — CxF/X-4 spot ink characterisation support

CharData 1.9.1

CxF/X-4 spot ink characterisation files

CharData now ingests CxF/X-4 (ISO 17972-4) files — the spot-ink-characterisation variant of the Color Exchange Format. All three conformance levels are detected and labelled in the file panel:

• CxF/X-4 — full characterisation: a tint ramp printed on the substrate and a second ramp printed over black
• CxF/X-4a — single-background characterisation: one tint ramp on the substrate
• CxF/X-4b — single-patch characterisation: just the solid (100 %) and the substrate (0 %)

The file's tint levels load as a device colorant named after the spot ink, and full CxF/X-4 files get a second PROCESS_BLACK colorant (0 = substrate ramp, 100 = over-black ramp). The 3D point cloud, 2D slice, data table, Tone Value chart, Extract/Export, and Compare all work as for any characterisation dataset. Because one or two tint ramps can't support a polynomial gamut model, the gamut shell and Estimate are unavailable for these files — a notice in the file panel explains this, localized in all 12 languages.

Fixes & improvements

• Spectral data in CxF files that declare StartWL only on the ColorSpecification's WavelengthRange (rather than on each ReflectanceSpectrum) is now read correctly — CxF/X-4 files are commonly written this way.
• The CxF type label and the file-panel notices now survive a live language switch.
• The ICC profile viewer's interface text — Tags-table column headers and the inline visualization section titles and evaluator labels — is now localized in all 12 languages.
• Manual and in-app help updated with a CxF/X-4 section.

1.9.0 — Inline ICC tag visualizations & hex/decimal tag numbers

CharData 1.9.0

Inline per-tag visualizations in the ICC profile viewer

Display File → Tags now draws the data behind each tag, inline above its text description — mirroring the arrangement in the companion profiletool:

• Tone curves — TRC tags and the A/B/M curves of LUT tags, colour-coded with a clickable legend to toggle individual channels.
• CIE 1931 chromaticity charts — the white point (wtpt) and RGB colorants (rXYZ/gXYZ/bXYZ), with the relevant point highlighted against the spectral locus and Planckian curve.
• CLUT lattice image and the colour-coded out-of-gamut map (gamt), rendered from the profile's own samples.
• Named-colour / colorant scatters on a*b* and xy.
• Single-point transform evaluator — for A2B* / B2A* / preview tags, apply that tag's transform to one colour set with sliders (float or CLUT grid-node input) and read the result in both human and normalized units. Evaluated through IccProfLib directly.

Each visualization is a collapsible section that reflows for mobile. Graphs are inline SVG and images are drawn to a canvas — all rendered locally, nothing uploaded.

Number format setting

A new Number format control in the Settings panel switches the Tags-tab Offset / Size / Pad columns between Hexadecimal (default — ICC byte offsets are conventionally read in hex) and Decimal. The choice is remembered between sessions and localized across all 12 languages.

Under the hood

• The shared IccVizModel visualization engine is compiled into the existing ICC-viewer WebAssembly module (no separate download); profile-derived text is rendered safely as text, never markup.
• Manual and in-app help updated.

1.8.3

2D gamut slice axis fixes + Tone Value default

2D gamut slice

Fixes axis presentation in the 2D gamut slice view for the a* and b* slice modes.

• Corrected swapped axis labels. In the a* and b* slice views the axis labels were transposed relative to the plotted data, so the L* and chroma axes appeared switched.
• L* is now vertical in both the a* and b* slice views, with the chroma axis (b* for the a*-slice, a* for the b*-slice) on the horizontal axis. The point cloud, gamut-boundary polygon, axis labels, and axis ranges were all aligned so they agree.

The L*-slice (a*–b* plane) is unchanged.

Tone Value

• The Tone Value section now starts collapsed when first created for a dataset (both explore and compare modes). The user's manual open/close is still preserved across result re-renders.

1.8.2 — Spectral→Lab via ICC weighting tables; A, LED-B1, F11 illuminants

CharData 1.8.2

Spectral → Lab: reference weighting tables + more illuminants

• Spectral-to-L*a*b* conversion now uses the authoritative ICC colorimetry-data illuminant × observer weighting tables instead of computing CMF × illuminant SPD on the fly. Each table spans 380–780 nm at 10 nm and is normalised so the perfect reflector has Y = 100, so results match the reference CMM. White-point chromaticities were verified against published values for every table.
• New illuminants: A, LED-B1, and F11, alongside the existing D50 and D65. The Standard Observer control selects the matching table (2° → CIE 1931, 10° → CIE 1964). M-condition behaviour is unchanged (M1 forces the D50 colorimetric illuminant; M2 zeroes the UV band ≤ 400 nm).
• Existing D50/D65 spectral Lab values shift very slightly — they are now more accurate (authoritative CMFs/SPDs with proper bandpass correction).

Docs

• In-app help and MANUAL updated with the expanded illuminant list and a note crediting the ICC colorimetry-data weighting tables.
• Added an architecture diagram (Mermaid + SVG) under docs/.

1.8.1 — ColorSpace ICC profiles & data table for all device spaces

CharData 1.8.1

ColorSpace ICC profiles

• ICC profiles of the ColorSpace class (spac) are now accepted, alongside the existing Output, Input, and Display classes. A ColorSpace profile carries the same A2B (device→colorimetry) transform as an output profile, so it is evaluated through the identical path — its gamut, slices, data table, Compare, and Estimate all work. Abstract, DeviceLink, and NamedColor profiles remain unsupported (none provide a device→Lab mapping).

Data table for every device space

• The Explore data table (generated device-colorant patches alongside their evaluated L*a*b*) previously appeared for CMYK ICC profiles only. It now renders for all device spaces — RGB, Gray, CMY, and N-colorant — using the same pattern: CMYK tabulates the IT8.7/5 patch set; every other space tabulates a generated set of paper white, single-channel ramps, and a two-channel overprint grid. The 3D "show data points" cloud for these profiles now reflects that same device patch set.

Other

• Removed the redundant Regenerate Shell button from the 3D gamut panel — the Density slider is the live resolution control, and every input change already rebuilds the shell automatically.
• Renamed the app title Chardata → CharData; the 3D canvas now slides when the File Select pane is toggled.
• New UI strings localised across all 12 languages (+ translation matrices). Manual and in-app help updated.

1.8.0 — CxF/X-3 support & measurement-only datasets

CharData 1.8.0

CxF/X-3 (Color Exchange Format, ISO 17972-3)

• Read CxF/X-3 as a third characterisation data format alongside CSV and CGATS/IT8. Prefix-agnostic XML parsing maps Object name → sample label, ColorCIELab → L*a*b*, ReflectanceSpectrum → spectral columns, and ColorCMYK/ColorCMYKPlusN/ColorRGB → device colorants.
• Write/Export a full CxF/X-3 document (device colorants + CIELab + spectrum + a ColorSpecification derived from the current illuminant / observer / M-condition).

Measurement-only datasets

Datasets with L*a*b* and/or spectral data but no device colorants (common in CxF, also valid for CGATS/CSV) now load and support:

• the 3D L*a*b* point cloud (sample label on hover),
• the 2D gamut-slice point cloud (projected, no boundary — like image-derived clouds),
• the data table with a sample-label column,
• Compare matched by sample label.

Gamut shell, Extract, Tone Value, Estimate, and G7 validation are unavailable for these (no device→colour relationship to model).

Other

• New UI strings localised across all 12 languages (+ translation matrices); fixed the measurement-only notice so it re-localizes live on language change.
• Refreshed the APTEC_CMYKOGV_Coated_LinearCTV_2025_M1 standard dataset (corrected CGATS header field/set counts).
• Manual and in-app help updated.

1.7.3 — ICC-vs-ICC Compare for matching NCLR profiles

The ICC-vs-ICC branch of Compare was previously hardcoded to 4-channel
CMYK — two NCLR profiles refused with "ICC vs ICC comparison requires
CMYK profiles." That guard also masked a latent alignment bug
(#2): the
source-order remap used a fixed CMYK index, so any future NCLR support
would have silently mis-aligned non-CMYK inks.

This release lifts the restriction for the well-defined case, and
replaces the "CMYK only" error with a targeted message for the cases
that remain ill-defined.

What changed

In runCompareWithIcc, the both-ICC branch now:

• Requires both profiles to expose identical ordered colorant lists
  (e.g. two CMYKOGV proofers). When they don't match, the comparison is
  refused with a message that names both lists.
• For 4-channel CMYK, continues to evaluate against the standard
  IT8.7/5 patch set (unchanged).
• For any other matching colorant list, evaluates against a generated
  N-channel patch set: paper white, single-ink ramps, and a two-ink
  overprint grid that thins with channel count to keep patch counts
  bounded (N=5 → 591 patches, N=7 → 666, N=10 → 821, N=15 → 1831).
• Drops the previously hardcoded ['CYAN','MAGENTA','YELLOW','BLACK']
  source-order constant in favour of a single ICC-slot lookup — the
  deep-equal guard makes that sufficient across every branch.

Wording

• icc_vs_icc_cmyk_only → icc_vs_icc_colorant_mismatch with {a} /
  {b} placeholders for the two colorant lists.
• Translations refreshed in all 12 supported UI languages (en, fr, de,
  it, es, pt-PT, pt-BR, zh-CN, zh-TW, ja, ko, sv) and in the canonical
  translations/Eng-*.xlsx spreadsheets. Eng-Pt.xlsx and
  Eng-Zh.xlsx pick up the previously-empty pt-BR / zh-TW columns for
  this row in passing.

Manual

MANUAL.md §5 (Compare mode) bullet for ICC vs ICC rewritten to
describe the colorant-list requirement and the CMYK / NCLR patch sets.

Closes #2.

1.7.2 — Image gamut in both modes, image-only plotting

The Image tab is now always visible — Explore as well as Compare — and the
3D plot and 2D slice can render the image trace on its own, with no
Dataset A loaded, whenever the image carries an embedded ICC profile or
is a Lab TIFF. Removing the image (or unloading the last dataset) hides
both plots again.

Bind dropdown

The dataset-binding dropdown now adapts to what is actually bindable:

• If the image has an embedded ICC profile, empty dataset slots are
  omitted entirely — only the embedded entry and any loaded datasets
  appear.
• If the image has no embedded profile but at least one dataset is
  loaded, empty slots still appear with a - placeholder so the slot
  positions remain visible.
• If the image has no embedded profile and no datasets are loaded,
  the dropdown is empty and disabled.

Plot defaults

• 2D slicer Show data points now defaults to ON (the prior default
  hid the patch cloud, which is the most common reason people open the
  slicer in the first place).

Wording

• Bind-error message reworded for the new behaviour:
  Dataset {slot} is not loaded. → No color dataset loaded.
• Translations refreshed in all 12 supported UI languages
  (en, fr, de, it, es, pt-PT, pt-BR, zh-CN, zh-TW, ja, ko, sv) and in
  the canonical translations/Eng-*.xlsx spreadsheets.

Manual

MANUAL.md §5.4 (Image gamut) updated to reflect the new always-visible
tab, the dropdown rules, and the image-only plotting case. The §4.4
slicer table notes the new Show data points default.

🤖 Generated with Claude Code