Guidelines for using the Agfa Horizon Plus
for digitizing Electron Micrographs

This scanner is designed for graphics artists and the Photoshop Plug-in for controlling the instrument is set up to do many things to your image to make it more pleasing to the graphics arts world. Examples are: applying a gamma correction, adjusting contrast, brightness, and sharpness of your image. While these techniques are very effective for most types of images, the subtle information in grayscale EM micrographs is often sacrificed when using the default settings for image acquisition. After wasting a good deal of time trying to get reasonable looking images with this scanner, I have developed the philosophy that all "effects" should be turned off during image acquisition. This is akin to what is done our PDS 1010 does during densitometry: an optical density is recorded at each pixel location and mapped into an image density (12 bit in the case of the PDS). All image processing is can be done later - by Fourier processing in the case of image reconstruction or with Photoshop in the case of making a figure. One could also use the PDS to acquire images for figures, but the Agfa Horizon is much faster and more convenient as long as the very fine image detail is not required (i.e., image detail that is only revealed by Fourier processing and is below the resolution of one's eye).

You might want to sign up for the scanner first. Then . . .

First step: clean the scanner with soft cloth and cleaning solution offered by Jon Weider. Also clean the piece of glass to be layed on top of the negative. Clean the negatives with compressed air to remove lint and dust. Lay the negatives emulsion side down - as many as 6 EM negatives can be scanned at once. Put the large, thick piece of glass from the PDS on top of the negatives and leave the top up.

Start Photoshop and get the acquire menu: this is an example of settings that worked. Given my above-stated philosophy, I would recommend using film type linear and adjusting the gamma later if necessary - or contrast brightness - with Photoshop.

Keep in mind that the finer the sampling, the large the file and the slower it will be to manipulate later. Try scanning at a given setting, increase the magnification of the resulting image until it is 1:1 on the computer screen. Decide if the details you need are present in the image. If not, scan at a finer sampling raster. It is wise to oversample a bit.

Range refers to the range film densities that are mapped to the 8 bits of image densities (0-255). If set to automatic, the program will preview the chosen image area to find the max and min optical densities within this region and map them to 0-255. Other densities will be mapped according to the other settings (gamma, tone curve, mode, original). If you choose "automatic", then be sure not to include any non-image parts in the scan area (e.g., numbers from the film or scratches in the emulsion). If this is unavoidable, then you may have to play with a manual setting for the range. To check that the range is ok, look at the histogram of the image after scanning and see that the full range of densities (0-255) are used for your image. You can select an area of the image and check the histogram just for this area also.

Sharpeness is very bad. This turned out to be the source of my frustration. Because EM film is quite grainy (or can be), the sharpness algorithm was actually sharpening the grain on the film, thus making it look really terrible. You can play with this later using the "unsharpen mask" in Photoshop. In some cases, you may want to actually blur the image in Photoshop rather than sharpen it to reduce the grain.

That's all for scanning. Tiff format for images seems fine. I use LZW compression to reduce file size and network traffic (at the expense of local processor speed). Also, I use IBM PC format, but I don't think this makes much difference.

When putting together a montage figure composed of several micrographs, be sure to balance the gray level (brightness) and contrast level for the group for best results upon reproduction.

This page created with Netscape Navigator Gold

This page last updated on September 20, 1996.