| |
Title
Section 3: Evaluation of Image Capture Cards
Introduction
Pick up virtually any PC computer magazine and you will be presented with a bewildering choice of cards offering combined facilities such as video capture and playback, enhanced graphic display capabilities, windows acceleration, TV tuning, and amongst other things, still image capture. This, coupled with the development of other bus architectures, PCI (Peripheral Component Interconnect) and VL (Video/Vesa Local), has led to some confusion for the teacher who simply wants to purchase a card for still image capture.
This section takes a look at a number of cards available for still image capture including the more traditional cards and those where still image capture is one of a number of options offered. There is a concern that the quality of images obtained with the newer, cheaper, multifunctional cards will be inferior to that from the 'standard' image capture cards, many of which will shortly be no longer available.
The cards evaluated are primarily for the PC, as the majority of academic institutions support the PC platform for delivering electronic teaching and learning materials. However, Apple MAC options are detailed where appropriate and the evaluation includes one image capture card for the Apple MAC.
The ISA (Industries Standard Architecture) Bus (also known as the AT bus) is found in any AT type computer. However the data transfer rate across the ISA bus is limited and is restrictive for applications needing high data transfer rates such as those involving moving video. The PCI bus not only provides greater bandwidth but is also processor independent. This is probably one of the PCI bus' greatest assets and because of this is likely to become the future standard for bus architecture. Thus any PCI card can be installed into any PCI based computer whether a PC, PowerMac, DEC, Alpha, etc. At the moment the data bus width of the PCI bus is 32 bits and as long as appropriate drivers are available will probably be developed further to support 64 bits. The VL bus, while offering higher bandwidth, is unfortunately a PC dependent architecture and is therefore less likely to develop much further.
At the time of writing few PCI cards are available supporting still image capture. This situation is likely to change rapidly.
The Vesa Media Channel (easily confused with the VL bus) is a feature built into cards that is independent of bus architecture and can therefore be found on ISA, PCI or VL cards. Vesa Media Channel allows one to link (daisy chain) up to 15 cards together, thus allowing any limitation in bus speeds to be bypassed or enabling the cards to run in parallel with the bus effectively maximising bandwidth.
To evaluate a number of still image capture cards in order to provide recommendations and guidelines on purchase to the academic community.
A number of cards (11 in total) were evaluated. For each card, details including appearance, connectors, installation of hardware and software, capture times and other observations are given. The default settings given for each card for brightness, contrast and colour were used. The cards and any associated software were assessed using the same set of images in each case. The quality of the captured images was compared against a group reference.
The cards can be divided into two categories:
- where the normal PC VGA output is diverted from the monitor onto the card, the card then overlays (in hardware) the incoming video signal onto the VGA signal and the combined output is shown on the monitor. Some systems (for example, the TARGA card) may additionally use a separate monitor, other systems may obtain the VGA signal from the feature connector which is found on a number of VGA cards.
- where software is used to 'overlay' the incoming video signal onto the VGA card's normal output (as in Video for Windows). No re-wiring (other than the connection of inputs) is normally needed.
Equipment set-up
A standard set-up of hardware and software was used throughout the evaluation. Figure 3.1 illustrates the hardware set-up used.
Hardware
Viglen PCI 486 DX2 66Mhz, 32 Mb RAM, Diamond Stealth PCI graphics display card with 4 Mb video RAM, multisync monitor (Idek Ilyama 8217). In addition, to enable a thorough test of the installation procedure for each card, the computer also contained two further cards: a SoundBlaster Pro card and Ethernet card in addition to the on-board SCSI-2.
A Sony DX 3-chip camera giving both composite and RGB outputs and carousel unit for holding the test transparencies. The video signal (composite or RGB as appropriate) was fed into the capture cards and also fed to a second monitor to enable the images to be focused. A Waveform monitor was used to maintain a standard video level for each image for all the cards tested.
Software
Images were captured using the image capture software provided with the card. In the one case where a TWAIN interface was provided, images were captured using Micrografx Photomagic.
Images:
The following images were selected and used to evaluate the cards:
Image Type of Image Colour/greyscale Video Level
number /b&w
1 Microscopy - transmission electron micrograph grey scale 90
2 Microscopy - light micrograph grey scale 100
3 Line Drawing b&w 90
4 Scenic - canal scene colour 100
5 Scenic - buildings colour 100
6 Light microscopy, section - histology colour 100
7 Anatomical colour 100
8 Macro, mosquito and skin tones colour 80
9 Colour/grey-scale test cards colour 100
Table 3.1 giving details of the images used.
The images were chosen as representatives of the types of image (and associated problems) that would be captured in reality.
Images 1-3 are greyscale. These were used to check not only the ability to handle continuous grey tones but also for purposes of seeing overall colour casts. Image #3 is quite detailed with a lot of edges, high in contrast, and is useful for illustrating the problems encountered when handling images with high frequency content.
The remaining images are colour. Image #4 was used for the initial comparison of cards, as a number of elements within the picture illustrated the differences between the cards well: the sharp edges of the houses; the roof tile (pattern); the telephone wire and lamppost against the sky; deep shadows; etc. Being of fairly high contrast, image #5 was used when assessing the amount of detail present in deep shadow. Image #9 provided a more objective measure of colour casts and black/white levels (contrast range); the CMYK levels being used for measuring (supposedly) black and white areas of the scales.
For each card the images were captured at the highest combination of resolution and number of colours offered.
The results are divided into two sections: 1) Details of individual cards and 2) Comparison of cards in relation to quality of image capture.
Title
|
