The digitization process described below is based on methods that were arrived at through a joint project in which JJT Inc. digitized a small collection of materials for the Unversity of Texas Libraries using the Kontron digital camera. This "test" project was undertaken as a means by which to help the Library learn how the conversion process using a digital camera would compare with the conversion process using a flatbed scanner. It was decided that there were efficiencies to be gained by using the digital camera.
The digital files were created from original media (glass plate negatives, nitrate negatives and positives). To create the digital file a technician placed the print images under the lens of a Kontron ProgRes 3012 digital camera. The camera was mounted on a control stand with computer control of X, Y, height, rotation movements of the camera with respect to the original image, color balance (when appropriate) of the light source and camera focus. The hardware and custom software were used to control black and white levels, gamma, and look-up tables. The combination of controls listed above provides complete control of the composition and the tonal range and values of the image as it is digitized and processed.
The originals were digitized so that the image would fill the maximum area possible within the 3072 x 2320 pixel resolution window of the digital camera. These 3K by 2K images (3K) have been scaled to create images defined by 1280 x 1024 pixel (1.2K) and 192 x 192 pixel windows. The panoramic prints have been scanned in segments with the short side of the print corresponding to the 2320 pixel dimension of the camera, and the segments have been concatenated to create seamless digital images. Scanning by Kontron is done at 12-bit-per-color-per-pixel (36-bit resolution) and the images are output at 8-bit-per-color-per-pixel (24-bit resolution). For black and white, this works out to 8-bit grey-scale resolution, and for color, 24-bit full color resolution.
Each image was viewed in a 1024 x 768 pixel window on a 20-inch monitor calibrated to the proposed International Color Recommendation Standard Default Color Space for Internet (sRGB). Images were evaluated in the recommended sRGB reference viewing environment before and after digitization to ensure proper capture. Also during digitization, images were evaluated for black and white levels, light intensity, gamma, and color (when applicable), and were adjusted, as necessary, to represent the tonal range and values of the original. Extraneous areas in the image were cropped out (but no part of the photograph was lost) and a frame surrounding each image was added so that the viewer would know that the entire image had been scanned.
In post processing, the image was sharpened, scaled, compressed, and dithered to create the various digital imaging files. Vertical images, rotated during scanning to make maximum utilization of the camera's aperture, were rotated back.
The first step in quality control was to display the images on a 20-inch monitor within a 1024 x 768 pixel window and review them to confirm that adjustments made during digitizing had the desired effect. Further adjustments were made as needed. Once written to a file, however, the image was not changed.
There are three (3) images for each of the 8,241 originals:
- an uncompressed, TIFF image
- a compressed JPEG image (with an average compression ratio of 10-15:1) at 1.2K resolution and 24-bit depth is the reference image viewable and downloadable by viewers;
- a dithered GIF image at thumbnail resolution (192K x 192K) and 8-bit depth is the inline image for the HTML/Web page.