Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
PD ISO/TR 17321 — Colour Characterization of Digital Still Cameras Standard Series
Graphic Technology and Photography — Colour Characterization of Digital Still Cameras (DSCs)
📷 PD ISO/TR 17321 is a BSI Published Document series adopting ISO technical reports on graphic technology and photography, focused on colour characterization of digital still cameras (DSCs). The series was prepared by BSI technical committee CPW/42 (Photography) and carries ICS 37.040.10 (photographic equipment). Building on the ISO 17321 colour characterization framework, the series explores how to determine scene analysis transforms, and how user controls and readouts should be designed for scene-referred imaging applications, providing a reference for camera colour science, colour management and digital imaging workflows.
💡 What a camera “sees” often differs from what a standard observer sees. The goal of colour characterization is to quantify and correct that difference; these technical reports focus on two highly practical links: scene analysis transforms and the user interface of scene-referred imaging.
1. Series Structure and Scope 📌
This batch comprises two technical reports that explore colour characterization from two angles — transform determination and user interface:
| Part | Document | Subject | Main Content |
|---|---|---|---|
| Scene analysis transforms | PD ISO/TR 17321-2:2012 (ISO/TR 17321-2:2012) | Considerations for determining scene analysis transforms | Goals, fundamental colour-related DSC characteristics (gain, dark current, etc.), transform determination |
| User controls & readouts | PD ISO/TR 17321-3:2017 (ISO/TR 17321-3:2017) | User controls and readouts for scene-referred imaging applications | Design of scene-referred (SR) capture processing mode, processing aims, colour encoding and file format |
⚠️ Colour characterization quality must be considered across the whole capture-process-output chain. The reports focus on the camera side; output-side characterization of displays and printers still requires the corresponding colour management specifications.
2. Fundamental Colour-Related Camera Characteristics 🔬
Part 2 systematically discusses the camera colour-related characteristics that must be understood before determining scene analysis transforms:
- Camera gain: the signal amplification of each colour channel, directly affecting colour balance and signal-to-noise ratio.
- Camera dark current: the inherent sensor output without illumination, affecting dark-area colour accuracy and needing correction and compensation.
- Spectral response and channel crosstalk: the sensor’s response curve over the visible spectrum and inter-channel crosstalk determine the physical capability of colour reconstruction.
- Colour cast and white balance: colour-balance strategies under different illuminants affect the applicability of scene transforms.
✅ Only by quantifying gain, dark current and spectral response can a reliable scene analysis transform be built; these measurements are also the base inputs for subsequent colour characterization models.
3. Determining Scene Analysis Transforms 🎯
The scene analysis transform maps raw sensor responses into scene-referred colorimetric space — the core of colour characterization:
- Transform goals: balance “colour fidelity” against “visual acceptability” for different uses (scientific measurement, commercial photography, artistic creation).
- Determination method: fit the transform matrix by regression or optimisation against standard test targets (colour charts).
- Evaluation and validation: assess colour accuracy after transformation using independent test images and quantify colour difference.
- Limitations: performance depends on illuminant, exposure and lens spectral transmission; reports should state scope and error.
⚠️ A scene analysis transform is not a magic “colour fixer”. It performs well under its training illuminant and test conditions but can deviate significantly under extreme illuminants or highly saturated scenes — always check how well the application matches the conditions.
4. User Controls and Readouts for Scene-Referred Imaging 🖥️
Part 3 (2017) addresses scene-referred imaging applications and the user controls and readouts a camera should provide:
- Design of SR capture processing mode: define capture modes suited to professional post-production workflows, preserving maximum colour and dynamic-range information.
- Processing aims: strategy of minimal in-camera processing to avoid irreversible image degradation.
- Colour encoding and file format: recommend suitable high bit-depth, wide-gamut colour encodings and file formats to support post colour management.
- User controls and readouts: exposure indicators, colour-temperature readouts, histograms and colour-space settings that help professionals control capture precisely.
💡 The value of scene-referred mode is “leaving the decision to post”: it delivers as-raw data as possible, letting photographers or post engineers perform white balance, tone mapping and output conversion from complete information — ideal for film, advertising and other colour-critical work.
5. Engineering Practice Essentials 🛠️
- 🔍 Camera selection and calibration: applications needing objective colour data (heritage digitisation, industrial inspection) should choose cameras that can be calibrated and output scene-referred data.
- 🔧 Test targets and illuminants: set up standardised test environments (colour charts, controlled light sources) so transform evaluation is reproducible.
- 📊 Colour management workflow: embed scene-referred data in ICC-based colour management for end-to-end colour consistency from capture to output.
- 📋 Documentation: record transform matrices, measurement conditions and evaluated colour differences to form a traceable characterization report.
6. Frequently Asked Questions ❓
❓ Q: What is the difference between scene-referred and output-referred imaging?
A: Scene-referred data represents the scene’s colour and dynamic range and suits post-processing; output-referred data is already encoded for a specific output device (e.g., a display), displaying directly but losing some scene information.
A: Scene-referred data represents the scene’s colour and dynamic range and suits post-processing; output-referred data is already encoded for a specific output device (e.g., a display), displaying directly but losing some scene information.
❓ Q: Why measure camera dark current?
A: Dark current generates spurious signals without illumination, hurting dark-area colour and signal-to-noise ratio; measuring and correcting it improves characterization accuracy at low light.
A: Dark current generates spurious signals without illumination, hurting dark-area colour and signal-to-noise ratio; measuring and correcting it improves characterization accuracy at low light.
❓ Q: Can a scene analysis transform work under any illuminant?
A: Generally no. The larger the difference between training and usage illuminants, the larger the colour error; when necessary, recalibrate under the target illuminant or use multi-illuminant transforms.
A: Generally no. The larger the difference between training and usage illuminants, the larger the colour error; when necessary, recalibrate under the target illuminant or use multi-illuminant transforms.
🔍 Q: Do the technical reports require camera manufacturers to implement specific features?
A: No. They provide methods and considerations; manufacturers choose implementation depth by product positioning — professional products typically offer more scene-referred control.
A: No. They provide methods and considerations; manufacturers choose implementation depth by product positioning — professional products typically offer more scene-referred control.