PD ISO TS 22028 — Photography Extended Colour Encodings for Digital Image Storage, Manipulation and Interchange (Parts 3 & 4)

🎨 PD ISO/TS 22028 is a BSI Published Document covering extended colour encodings for digital image storage, manipulation and interchange. In modern digital imaging workflows, cameras, monitors and printers differ in colour capability, so a unified colour encoding is needed as an intermediate exchange format to achieve “what you see is what you get” colour reproduction. The series defines several extended colour encodings: Part 3 specifies the reference input medium metric RGB colour image encoding (RIMM RGB), and Part 4 specifies the European Colour Initiative RGB colour image encoding (eciRGB 2008). Both were prepared by BSI Technical Committee CPW/42 (photography) and published in 2012; Part 3 supersedes the withdrawn DD ISO/TS 22028-3:2006.

1. Scope and General Framework 📌

PD ISO/TS 22028 applies to the definition of colour encodings needed for digital image storage, manipulation and interchange. Part 3 (RIMM RGB) defines a metric RGB encoding based on a reference input medium, used to represent image data from various input devices (digital cameras, scanners); it is typically characterised by a linear or near-linear encoding transfer function and a wide gamut. Part 4 (eciRGB 2008) defines the RGB encoding proposed by the European Colour Initiative, widely used in prepress and publishing workflows, adopting a logarithmic or segmented transfer function to improve dark-tone encoding efficiency.

2. Definition and Components of a Colour Encoding 🔧

A complete colour encoding is usually defined by the following elements:

• Chromaticity primaries: define the chromaticity coordinates of the RGB primaries corresponding to the encoding.
• White point: define the reference white (usually D50 or D65).
• Transfer function: define the non-linear mapping between encoding values and actual luminance, often linear, logarithmic or segmented (e.g. the sRGB approximation).
• Encoding range and bit depth: define the minimum/maximum values and bits per channel (e.g. 8, 10, 12, 16 bit).
• Reference medium or output model: define the viewing conditions and reference media characteristics for the encoded image.

RIMM RGB uses a linear transfer function with a large dynamic range, suited to representing input scenes and intermediate processing; eciRGB(2008) uses a segmented transfer function to optimise dark-tone encoding density, suited to cross-media colour management before printing.

3. Requirements and Reference Specifications 🔩

Standard requirements for colour encodings include: clear definition of encoding parameters (primaries, white point, transfer function, range, bit depth), a way of marking the encoding (recording the encoding type in metadata), and a statement of conformance to reference specifications. The standard also references relevant documents (e.g. the ISO 15076 series on ICC colour management and CIE colorimetry documents) as the basis for colour conversion and configuration.

In practice, image files should record the encoding type (e.g. RIMM RGB or eciRGB) in metadata so that downstream colour management systems automatically identify and apply the correct colour conversion, achieving cross-device colour consistency.

4. Engineering Practice Essentials 🔍

• 🎨 Colour management consistency: use ICC colour management consistently across capture, editing, proofing and printing, converting different encodings into a single working colour space.
• 🔧 File annotation: record the colour encoding type in EXIF/IPTC metadata to prevent downstream misinterpretation.
• 📊 Bit-depth selection: for high-dynamic-range or heavily graded images, prefer 16-bit encoding to reduce quantisation error.
• 🖨️ Output adaptation: before printing, convert RIMM RGB/eciRGB data to the printer gamut via soft proofing, and check for out-of-gamut colours.

4b. Practical Workflow Considerations 🔧

Adopting extended colour encodings in a digital photography workflow requires careful handling of a number of practical issues. First, the encoding space must be declared correctly in the image metadata: RIMM RGB linear and eciRGB(2008) use different transfer functions, and mislabelling the space leads to severe colour errors downstream. Second, bit depth matters — linear encodings need more bits per channel than the perceptual eciRGB transfer to maintain the same shadow detail, so choice of depth should balance file size against the need to avoid posterisation in dark tones. Third, processing should remain in a working space of sufficient gamut before output; applying sharpening, tonal adjustments or compression in a small-gamut space can throw away colour information irreversibly.

For long-term storage and interchange, the profile or encoding declaration should be embedded in the file so that the colour appearance can be reconstructed later even if the original software is no longer available. When converting between the two encodings, the transformation should pass through a reference medium or profile connection space rather than a direct one-step table, to preserve neutrality and avoid double transfer-function application. For archival and research use, the measured data in linear space is often preferred, while eciRGB(2008) is common for creative editing and exchange because of its perceptually uniform behaviour.

In summary, extended colour encodings extend the photographer’s working range beyond the limits of conventional sRGB files. With correct encoding identification, sufficient bit depth and a proper colour management pipeline, they preserve both the highlight and shadow detail of high-contrast scenes and provide a reliable basis for professional image storage, manipulation and interchange.

5. Frequently Asked Questions ❓