There is no doubt about the usefulness of an NBTV EPROM image source. If you are working on any type of televisor project, the EPROM image source will give you a source of test images at any time and for as long as needed. There are two basic circuit options available - my ROMScanner from the Image Communications Handbook and the latest version of  Grant Dixon's* NBTV EPROM generator. These two circuits have much in common:

• Both now feature a crystal-controlled timebase for accurate image display rates without the need to make any adjustments. The advantage of this is that you can refine your televisor using either image generator as a timing standard.

• Both are about equally simple/complex, involving five active integrated circuits and a small number of regulators and transistors.

The ROMScanner is designed around the fact that each of the 32 image lines is digitized with 256 pixels/line (240 in the active image area). This was done to provide the close approximation of continuous analog image lines. Assuming you do a good job choosing and setting-up your images, the ROMScanner should never be the limiting factor in how good the picture looks when displayed on a televisor! The down-side of this approach is that the EPROM will only hold four images (8196 bytes/image) and it is thus not practical to use the unit to display moving images.

In contrast, the Dixon* circuit digitizes each line to 64 pixels, so that each image requires only 2048 bytes for storage. Using the same EPROM as the ROMScanner, Grant's* unit will hold 16 images. With a 27C512 EPROM, the unit will hold 32 images and, with an appropriate image sequence programmed into the EPROM, it can produce an "endless loop" of ~2.6 seconds of full-motion video.

Given that both units are a source of stable NBTV video, which one should you build? Well, if optimum video quality is the issue, the ROMScanner has a slight edge, as you can see by examining Figure 3.8 on page 3.8 of the Image Communications Handbook.  While the difference is real, you would have to have a very good televisor to see it and you might not be able to see the difference in the case of moving video images. If the slight quality difference is not important and it is more useful to have a larger number of images per chip and/or the option for full-motion test images, Grant's* unit would be the project of choice. The point is, both units work very well and you can hardly go wrong picking one over the other.

* I am attributing the circuit to primarily to Grant Dixon as he was the principal designer of the prototype from which, with modifications, the ROMScanner was derived. The current version of this circuit is published in the on-line NBTV Handbook, which you can access via the link at the top of the page. This Handbook is edited by Chris Lewis, Grant Dixon, and Klaas Robers. The chapter does not indicate the relative contribution of the other editors to the current version, and I am hoping that nobody will be too upset is I continue to attribute the design to Grant.