Here are a few suggestions for making your images look better
and easier to recognize.
Size and format
Images of the LED Toy are black&white (that is, black or white, no
grey), 24 x 24 pixels in size, and are distributed in the
Images can also be made wider than 24 pixels, up to 24 x 255 pixels,
but such "wide" images cannot be rotated, so you should strive to make
images not wider than 24 pixels. Images with a width of less than 24
pixels are automatically padded to 24 pixels, to allow for rotation.
If using the bitmap program, you may find the optional preview
you can enable with Meta-I or Edit>Image useful.
Since the LED Toy does not know in which direction it or the viewer is
moving, images can appear from left to right, or from right to left.
In some cases, this effect can be desirable, e.g., a question mark
expressing confusion will look even more confused if it changes
direction all the time. On the other hand, an arrow intended to point
to the left will just as often point to the right.
Normally, the speed of movement is highly variable, which will
cause variations in how thin or thick the image appears. Images
that are only recognizable if very precise proportions are maintained,
are therefore not very well suited for this application.
Example: a smiley is basically a face, and can be recognized
even if very much out of proportion, while recognizing a cross
hair is much more difficult.
Horizontal lines are always solid, while vertical lines are always
dotted. This means that a vertical line that is one pixel wide will
be much harder to recognize as such than a horizontal line of the
In general, lines that are two pixels wide look much better than
lines that are only one pixel wide.
Bear in mind that the whole concept is based on the complex pattern
recognition mechanisms in the human brain. Some effects may not be
easily predicted, so you should always verify what your new creations
look like under different conditions.
In particular, very complex images, i.e., those with a large number
of light/dark changes, are harder to see than simple ones.