AGOCG logo
Graphics Multimedia VR Visualisation Contents
Training Reports Workshops Briefings Index
This report is also available as an Acrobat file.
  Back Contents

Appendix Further equipment details

Equipment common to high and low budget scenarios The Computer Terminal used to house the digitiser board and retouching software for both central and departmental resource models was a Panrix 486 PC with a Wizard 9000VL 24 bit (true colour) Video Card, SVGA Monitor and Photo CD-ROM Drive. File sizes tended to be very large for captured images prior to manipulation, so there was also a 1Gb rewritable Optical Disc Drive available. The photo-retouching software package used was Adobe Photoshop 2.5 for Windows, which included a range of particularly subtle enhancements to improve the look of images without making them appear 'unnaturally' enhanced. The SVGA Monitor was a ViewSonic 17in running at 800 x 600 pixels displaying 24 bit colour. Given the trade-off between number of colours, number of bits and number of pixels, this was a comfortable combination with which to work within the hardware package supplied by Panrix, which contained 16 megabytes of RAM and 2 megabytes of VRAM.

Common to both equipment resource models was a standard Kaiser RS1 rostrum copy stand with 1000mm column. A number of alternative light sources were available for attachment to copy stands. Reflected light sources were either 2 x 800watt Rank Strand Redhead focussable spots or 2 x Anglepoise lamps with standard 60 watt domestic frosted bulbs if lower light levels were required. 60 watt domestic bulbs provided a 'warm' light source from the colour temperature point of view, but the Redheads gave a predictable 3,000 - 3,400 degrees Kelvin. Alternatively the Anglepoises were used with small (275 watt) Photoflood lamps, rated at 3,200 degrees Kelvin, for the predictable colour temperature required by slide film, but they ran very hot in this mode which is only recommended for short periods. Transmitted light was provided by a Master 18 x 12in. (45 x 30cm) screen size lightbox. It was suspected that the fluorescent tubes inside had a blue spike in their spectral output but this was not a problem and these particular lightboxes are widely used for photographic display purposes. Both video rostrum cameras were white balanced with each light source as appropriate. If light output needed to be cut down, one thickness of standard Lee ND6 filter was used to reduce output by 2 full stops. Although this filter was not spectrally neutral the rostrum cameras white balanced to it, but slide film exhibited a green cast, which was 'massaged' in Photoshop but not totally cured after digitisation. We used one thickness of ND6 on the lightbox to reduce its brightness for comfortable working when capturing slides.

A high quality video monitor was necessary to maintain standards when working with variable processing paths. Analogue signals were particularly vulnerable to quality loss if manipulated prior to digitisation, and confidence in the quality of the analogue output of the rostrum camera before transfer was essential.

Photographic considerations

The film stock used for slides was Kodachrome 64, giving maximum sharpness (between 63 and 100 lines per millimetre - LPM) coupled with some exposure latitude in case of bad weather conditionsThe film stock used for slides was Kodachrome 64, giving maximum sharpness (between 63 and 100 lines per millimetre - LPM) coupled with some exposure latitude in case of bad weather conditions which in this project were not a problem. While the KY-F30B video rostrum camera (high budget) resolved about 400 horizontal TV lines the usable portion of a 35mm slide can resolve >2016 lines horizontally (32mm x 63LPM) so its spatial resolution is >5 times better than the KY-F30B. In spite of these considerations soon after image capture commenced it became apparent that for some subjects, particularly those viewed through a microscope, a different benchmark from the slide might have been more appropriate.

Although spatial resolution was excellent, light levels for Kodachrome 64 had to be accurate to less than + or - half a stop for correct exposure coupled with optimal contrast range. It is worth considering these conditions in comparison with Kodak Ektar 100 negative stock used for negatives and prints. This film inherently has greater contrast range and an exposure latitude between -1 and +3 full stops either side of correct exposure, within which parameters an acceptable colour print can be obtained. To produce each photographic image a standard Nikon F3 35mm camera was used with appropriate Nikon lens plus Wratten filter where necessary for the particular subject. Exposure was kept between f5.6 and f8 to maximise lens sharpness, with exposure time calculated automatically. This was possible because all the subjects were static and a tripod was used throughout.

Central resource equipment

Rostrum camera.

A major piece of equipment for the Central Resource Model was a vertically mounted JVC KY-F30B 3 chip video Rostrum Camera channel with a JVC HZ-714B 7-98mm/f1.4 (14-1) zoom lens with macro facility. Using an additional 6 diopter close up attachment it was possible to fill the frame with a single frame of 16mm film or any size of image larger than that by using the lens in its zoom and macro modes with or without the close up attachment. This camera package comes with automatic black and white balance, and +9 or +18db of extra light sensitivity allowing it to operate down to 20lux at f1.4 with consequent resolution loss.

There is provision for capturing positive or negative images, although the negative option is not as useful as at first might appear. In spite of the potential gains of working from negative film (greatly improved contrast ratio and exposure latitude), this does not mean that a rostrum camera with a negative switch can be simply used for this purpose. The colour dyes used in negative film have been perfected to provide optimum quality when printing hard copy. Electronic transfer of negative images causes negative colours to fight with each other, exhibiting 'cross talk' (interference), and an elaborate system of filtering, known as 'masking' is necessary to cope with different spectral sensitivities between media to capitalise on potential gains. A colour paint box will not adequately replace masking arrangements and by itself is not a comprehensive solution. The negative switch on a rostrum camera, without additional masking, will correctly replace blacks with whites and vice versa, but other colours will not turn out so well.

A feed of 100% colour bars, and enhanced edge sharpness through the use of a high-resolution shutter or contour correction are also facilities on this camera head. Controls for operating and adjusting enhancement devices are strongly built for intensive use and are a major reason for budgeting a high outlay on this single piece of professional quality equipment. While enhancement may be performed after digitisation faithful colour rendition is better achieved before transfer if possible. Spatial resolution problems (presented for instance by thin lines on maps and drawings) are also better optimised before digitisation if possible.

Output from the camera can be either Composite (1 volt peak to peak, 75 ohm), Y/C or S-Video (1.0 + 0.3 volts peak to peak, 75 ohm) or RGB (3 x 0.7 volts peak to peak, 75 ohm). Standard broadcasting line-up procedure is followed by closing down the iris fully and black-balancing the camera with a single push button control. The iris is then opened up and white balancing is carried out by pointing the lens at a piece of white paper illuminated by the light source of choice or, in the case of transmitted light, by switching on the light box and white balancing against its back-lit opalescent glass. White balance is then carried out using the same push button control; the principle being that if a camera will transmit white correctly it will do the same with the full range of colours. This automatic line-up procedure is sufficient to ensure high quality colour and spatial resolution when used on all reversal images (slides, prints and artefacts), provided the camera does not develop a fault. As with all quality automatic line-up cameras today, if a fault does develop it can be difficult to access it for maintenance.

The KY-F30B camera operates routinely with the lens at about f5.6 - f8 with the Anglepoises plus 60 watt opalescent domestic bulbs.

Video capture board

A Screen Machine II board was chosen for Central Resource operation. This equipment can accept either a Composite video signal (output 1.0 volt) or a Y/C signal (1.3 volts in total). The Y/C signal, being more powerful, provides better spatial resolution, but the Composite signal is adequate for many applications. Images can be captured at full screen, half screen or quarter screen size with this board. Full screen capture (736 x 560 pixels) was preferred on the basis that unless there was a known image size required for output, it was best to start with as much information in store as possible. During processing, file sizes would be shrunk downstream, but with less digital information than eventually needed for display, quality would be lost by trying to insert it later. Preferred capture was to Targa file at 736 x 560 pixels in millions of colours via Photoshop to optical disc, which at that stage meant a file size of about 1.2 megabytes for a full colour photographic image. The Targa file format was selected because Iterated SystemsĘ fractal compression (which we intend to assess in the future) was designed to work with Targa files.

To compare other methods of capture meaningfully (e.g. via scanner) standardisation on 736 x 560 pixels was maintained as specified for the Screen Machine II board. Colour and resolution control settings were standardised on the capture board to allow realistic comparison of the output of the different image capture devices.


The Central Resource scanner was a Hewlett Packard Scanjet IIC with associated Deskscan II software. In common with most scanners today, the Scanjet was TWAIN compliant for use with a photo-retouching package like Photoshop or Picture Publisher. Colour control within the scannerĘs own software was not sophisticated and more subtle results were achieved in Photoshop prior to eventual storage on optical disc. To optimise screen display quality the recommended resolution of 75 dots per inch was selected. Higher resolution was unnecessary in the absence of a requirement for hard copy prints.

Low budget equipment

Rostrum camera

Our choice for a low budget camera was a JVC TK-1280E single CCD channel with Computar 18-108mm/f2.5 (6-1) zoom lens with macro facility. This was a camera capable of giving very acceptable colour and spatial resolution quality for its price. Its design provides for two pre-set colour positions plus a manual adjustment position accessing two colour vectors, green - magenta and blue - red, on screw controls. While this camera can deliver good colour quality it does need adjustment to achieve it, possibly for each individual image, and the controls are not really designed strongly enough to stand up to sustained manipulation on a regular basis. This is a camera which will work best if it is set up and then can deliver without too much adjustment. Used for image capture it is likely to have to cope with images of varying quality in many formats, and its two pre-sets and adjustment provision may soon wear out with use. Its final drawback is a white balance sensor which adjusts continuously to whatever image is presented to it in the automatic adjustment mode. This sensor can only be disabled by switching to one of the pre-set positions. In spite of its drawbacks, the camera does represent value for the amount of money that is spent on it.

Video capture board

The low budget board used was a videologic captivator board which retails for around £300. No particular losses of quality were detected because of it, but it did not have the adjustment possiblities standard on the Screen Machine II board.


The low budget scanner was a Logitech Scanman Colour handheld model which was TWAIN compliant for use with Photoshop or a similar photo-retouching package. Scanning was achieved, depending on the size of the original image, by moving the scanner across it in a number of parallel passes to be stitched together more or less seamlessly in the software. This stitching operation is automatic, but can be undertaken manually in the event of the software not being able to find enough common reference points.


In addition to the observations on use of PhotoCD images in 3.1 and 4.3, the following is a technical description of the PhotoCD process from a Kodak Fact File.
"PhotoCD images are created though the scanning of photographic film - either positive or negative - to create a photographic quality digital image. The image file is compressed from its original file size (18Mb to 72Mb depending on film size, i.e. 35mm to 4x5 inches) down to approximately 5Mb to 18Mb, depending on film size, through a visually lossless compression routine. The resulting image file is converted to the Image Pac format where it is represented in five or six resolutions:-
Base/16 (thumbnail resolution)     128x128 pixels
Base/4                             	256x384 pixels
Base (TV resolution)               	512x768 pixels (in reality 480x640 pixels)
BaseX4 (HDTV)                      	1024x1536 pixels
Basex16 (photographic)             	2048x3072 pixels
Basex64 (photographic)             	4096x6144 pixels (Pro Master PhotoCD only)"

Back Contents