Data Retrieval Technology - How to Capture Images from piqlFilm?

Table of ContentsCopy Anchor Link

1. Introduction
2. Components
3. Requirements

APPENDIX A. Requirement specification for image output from a reading device

1. IntroductionCopy Anchor Link

Retrieving digital data from film requires two processes. The first process is the image capturing from the film and the second is the decoding of that captured image file. This document covers the first process: the capturing. This document describes the minimum optical parameters required for building an image capturing device optimize for digital data exposed on 35 mm film. The document does not describe the film transport mechanism as this is implemented according to the image capturing device requirements.

2. ComponentsCopy Anchor Link

There are three main components to capture an image from the film; the sensor, the optics, and the light source. Also, the film needs to be advanced frame by frame or continuously, either manually or through a motorized transport system.

Image capturing device (Film scanner) The capturing device converts a photographic image into an array of 8-bit values in the physical memory. Each element if this array represents a specific area on film. The size of this area is the size of the printed pixel (6,08 microns) divided by the oversampling factor. At the time of creation of this document, the film content was captured using a Eliixa e2v line sensor, continuously capturing lines while the film was moving past the sensor. The total number of captured pixels per recorded image was: 12288 x 6879. However, the image can be captured by a digital camera with a CMOS or CCD sensors. See in Appendix A, the requirement specifications for image output from a reading device.

3. RequirementsCopy Anchor Link

3.1. The SensorCopy Anchor Link

Oversampling more than x2.8 is recommended to achieve sufficient image quality for data restoration. The film frames are exposed using a 4K image device, so the minimum recommended sensor size is 11,2K

3.2. The opticsCopy Anchor Link

The frame's dimension on the film is 24,92 mm width by 13.87 ± 0.05 mm height (Defined in SMPTE-59) (1) The lens shall allow to capture the entire printed frame with approximately 200 μm margins*.* The pixel size in the frame is 6,08 μm x 6,08 μm, Modulation Transfer Function (MTF) describes the lens performance by measuring the transfer of modulation (or contrast) from the film to the image sensor When choosing the lenses together with the image sensor, the MTF should be higher than 75% at 83 lp/mm.

1 SMPTE 59-1998 Motion-Picture Film (35-mm) – Camera Aperture Images and Usage (R2003)

3.3. The light sourceCopy Anchor Link

The light source should match the sensor and the optics in terms of wavelength, intensity and spatial distribution. Maximum radiant exposure in the film plane should not exceed 0,2J/cm2.

APPENDIX A: Requirement specification for image output from a reading deviceCopy Anchor Link

List of requirements for output from an image capturing device (film scanner) that shall be fulfilled so that the digital data (2-bit depth) can successfully be decoded.

Specification description

• The output data shall be stored in an array of 8-bit values located in physical memory.
• The oversampling factor shall be more than 2.8, i.e. for a 4k printing the scanned images shall be in 12k resolution.
• Every scanned image shall contain one printed frame.
• The scanned images shall be inverted, means that the film base shall be dark and the high-density patches shall be bright.
• The orientation shall match the orientation given by the Frame Format specification. (see document Generic 4K Frame Format)
• For Frame Format:
• The calibration bar appears on top
• The high-density part of the calibration bar is on the right-hand side
• The printed area shall appear in the middle of the scanned image. Some space must surround the printed frame;
• Distance from a border of the frame to the closest edge of the scanned image shall be more than 20 and less than 500 scan pixels.
• Image rotation should be minimized. The rotation can be controlled by measuring Y coordinates of top left and top right corner. The angle shall not exceed 0.14 degrees. That means for 12 k image, the difference between Y coordinate of the top left and top right corners shall be less than 30 scanner pixels.
• The empty area surrounding the printed frame shall be free from artifacts and bright objects, such as edges of perforation holes, lines, etc.
• The scanned image shall look as uniform as possible regarding brightness. The non-uniformity of brightness shall not exceed 40%. There shall not be visible artifacts like scratches, splashes, bright dots, etc.
• The dynamic range (contrast) shall be more than 150.
• The image must not be saturated.
• The scanner LUT shall be calibrated in such a way that the dependence between density on film and brightness on scans is linear.
• The MTF values taken on vertical and horizontal line pairs shall be more than 75%. The vertical and horizontal line pairs can be found around the frame data container (see document Generic 4K Frame Format).
• A histogram taken over entire digital data frame should show four distinct peaks.