35
8.2.2 Should I use a digital or an analog camera?
This is application dependent. If the camera is used as a source of good, clean video that will not
be processed in any way (viewed directly on a monitor), a DVC-0A analog-only camera will
suffice. If the video output from the camera will be processed (as in a frame grabber / image
processor), you have two possible solutions:
1. Use a DVC-0A analog-only camera and feed the output to the Image Processor. With this
solution, there is always the possibility that the "super-quiet" video of the DVC-0A gets noise
added to it within the Image Processor before it is digitized. This may lead to non-optimum
results even with a good, "high-end" image processor. That's due to the fact that the Image
Processor typically resides in a computer that is a noisy, digital environment !
2. Use a DVC-08 or DVC-10 camera with direct digital video output and an Image Processor
with a digital video input. Such a system is considerably more immune to noise and is likely
to deliver better performance.
8.2.3 What are the benefits of using a digital camera?
8.2.3.1 Effective Number of Bits:
In a DVC digital video camera, the signal-to-noise ratio of the video signal that is being digitized
is preserved by taking extreme care in the video processing circuits. The digitization of the video
to 10 bits (in a DVC-10) ensures that all 10 bits truly represent data. In a conventional analog
video system, the video signal from a high signal-to-noise source (such as the DVC-0A) could
have noise mixed into it prior to digitization in the Image Processor's typically noisy, digital
environment. Even if the Image Processor is well designed, it derives its power from the
computer bus which could be a significant source of noise. The benefits of using a high-end
camera and a well designed Image Processor board could easily be lost by adding noise from
the computer bus or from another peripheral component (disk drives, video cards etc.) that may
be plugged into the same bus ! The addition of this noise could very easily degrade the system
performance to that of a 6 or 7 bit system. (See ENOBs)
8.2.3.2 Elimination of pixel jitter:
In an analog video system, the pixel clock is derived from the incoming video via a Phase
Locked Loop (PLL). By the very nature of the pixel clock recovery process in a PLL, a certain
amount of pixel jitter or uncertainty in the clock transitions is generated. This uncertainty is
harmless in applications where the video data is used merely to create pictures. In a high-end
application however, where the data is used for measurement, the pixel jitter of a conventional
analog video system can lead to inaccurate, non-repeatable results. Note: this would happen with
all analog video camera systems, including those using the DVC-0A ! (The fault lies with the PLL
action on the Image Processor boards).
8.2.3.3 Sub-pixel accuracy:
In some applications, a Sub-pixel interpolation scheme is employed; this involves using a
mathematical algorithm to compute numerous pseudo-pixel values between two real pixel
values. Needless to say, the more the accuracy and stability in the two "real" pixel values, the
better the response of the sub-pixel interpolation algorithm.
To Next Page