Many test patterns exist for TV cameras to judge sharpness by, and this one was designed by myself in order to judge corner to corner sharpness. The aspect ratio of this image is 4:3 as is customary for video, Super8 and 16mm. Each circle contains horizontal and vertical tapering wedges in two resolutions with concentric rings showing their spacings in equivalent lines of resolution. You'll find this test pattern handy for testing your camera and its various zoom settings. Make sure you frame the image without much white surrounding it. Click on the image for an A4 sized printable version (vidsharp.gif, 250KB). The problem with this kind of test pattern is that it does not show where the lens focuses most sharply. Is it in front or behind the test pattern?

Basic optics as discovered by Descartes, can be described in two ways as shown in the diagram here. A lens projects an image located at infinity, like the sun, at a precise spot located at the lens' focal point. Here all sunrays concentrate to project a small image of the sun. Newton discovered the symmetry of behaviour for rays inside and outside the lens and formulated the basic optics with this symmetry in mind as B x V = f x f.

In the test setup we place a 3 diopter closeup lens A in front of the camera C which is set to focus at infinity. The optical illusion scale B is placed in front at a slight angle. The focal point thus lies 1/3 metre or 333mm in front of the lens, and this can be measured. Equally important, the focusing wedge in the viewfinder optics must agree that this is the viewfinder's point of focus too.

It is very important to shoot the film with wide aperture. In the case of movie or video, this means that the intensity of the light needs to be dimmed down. Here is what the camera sees, roughly a broad band of equal width and marks of equal widths. Where the camera was focused, a clear depth of field pattern emerges. Notice how the perception of depth of field narrows towards the finer marks, another optical illusion. The example here was photographed at 15cm f4 and the camera's autofocus was used for focusing. The conclusion is that the autofocus and lens agree.

The graph shown here is from an actual measurement. Horizontally it shows the amount the primary lens was turned, in notches, both clockwise (CW) and counterclockwise (CCW). Vertically the shift in focus as read from the chart. Many points are needed because of the amount of uncertainty in reading the focal point from the depth scale. The average curve drawn through all the measurements shows that the present focal point causes 40mm unsharpness on a distance of 333mmm, and that the exact focal point lies three notches CCW. After making that adjustment, the lens indeed performed optimally.

 

This is what the optical illusion depth scale looks like side-on. This version begins at 5cm from the lens (left) and ends at 34cm distance. Red marks have been placed for 15, 20 and 25cm. The 25cm mark can be used for a camera with a 4 diopter closeup lens at its infinity distance setting. Make sure aperture is maximally open. Note that due to printer, plotter and scanner resolution the finest graticules become messy where lines come close together. However, this does not normally matter as coarser graticules are above these. You can print a ready A4 version (250KB) ilusion3.jpg

The optical illusion scale The Anthoni Optical Illusion Scale overcomes the illusion that what is nearer appears larger and thus sharper, by creating the illusion that all marks on the scale are of equal size. Thus marks become higher and wider as they are located further away. The computer program to achieve this was originally programmed for an Apple computer with Watanabe Miplot plotter for A3 size paper. By running the program four times at different settings, the Optical Illusion Scale is drawn.

Program description The program shown above is a BASIC program for the Apple computer, but in its simplicity can be transcribed for other situations. The variable names used are explained in the diagram above the program. This makes the INPUT section straightforward. In the main program loop the variable D = distance to lens is incremented with half the pen width to achieve a totally black mark. In statement 300, D is incremented by the width of the mark to create a white mark.

The Miplot plotter needed integers in tenth of a millimetre for D, H2 and H3, hence the 10 times multiplication factor. The 'M' command moves the pen in statement 240 without writing. In statements 270, 280 the 'D' (Draw) command draws the lines.
The standard scale was drawn in four runs for L= 4,2,1,0.5  H= 10,10, -10, -10  and Q= 10, 0, 0, -10 (mm) for ever finer blocks, with fixed settings S=100, E=400, F=333, P=0.5
Note that if the marks are too high, they cannot fit inside the film frame.
Tip: when developing your own program, begin with a wide pen of 0.5mm to get quick results. For best resolution choose a fine pen of 0.1mm.

Here is another program version, this time programmed in WANG BASIC2c for a KYOCERA printer. The printer's native PRESCRIBE2 commands were used for plotting, as explained in the program. Because this is a laser printer, all four sectors are drawn, one after another, before printing the page.
 

-- Seafriends home -- UW photography index -- Rev 20051018,