|Modern small frame cameras like Super8 do not only bring a spectacular range of zoom but also a very useful wide angle. But often their wide angle footage shows unsharpness which may be caused by a slight misadjustment of the distance to the focal plane. My suspicion was that the film does not run tight against the film guides but a very small distance behind it. This article describes a method not just to measure wide angle unsharpness but also to adjust the lens precisely with the film running.|
Because the distance from lens to focal plane (emulsion) is very critical, wide angle unsharpness is a common problem, but more so for small format cameras like 16 and 8 mm film and video. I own two Braun NIZO S560 Super8 cameras both of which show unsharpness in wide angle footage, and on further investigation I found that indeed most Super8 cameras suffer from this problem.
|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?|
The diagram of a film camera shows how the subject is projected through a secondary telelens A, through a semi-transparent prism B and through a primary lens C onto the focal plane which is inside the Super8 cartridge. At the point of the prism B, the light runs almost in parallel such that the distance from the secondary telelens to the focal plane is not critical. Because the light here forms a near parallel beam, it can be diverted through a narrow tunnel around the cartridge where it meets another mirror and more optics. In the eyepiece a virtual image is projected, which views the subject at true distance. With the top mirror the position of the image is adjusted with three screws and ultimate wide angle sharpness is adjusted by turning the primary lens C.
|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.
|We will now use Newton's formula to calculate how sensitive a lens
is relative to shifts in its focal plane. We do this by moving the subject
V from 2m to infinity, examining the focal plane shift B for various lenses:
As one can see, a misadjustment of 12 microns for a 5mm lens, shifts the subject from 2m to infinity. It explains why wide angle lenses are often unsharp. The table also shows that the effect proceeds quadratically, and that depth of field which is what the table also shows, increases quadratically with the inverse of focal length. Thus a 5mm lens has four times more depth of field than a 10mm lens and 16 times more than a 20mm lens, and so on.
|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
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.
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.
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.
|0005 REM Optical illusion scale. 20051017
0006 REM for WANG BASIC2C & Kyocera PRESCRIBE2 plot commands - MAP= move to absolute position - DAP draw to absolute position - SPD set pen diameter
0060 REM origin at top of paper: X=10.5: Y=0: REM pen dia in cm: P=.01
0065 SELECT PRINT 215(0): PRINT "!R!UNIT C;SPD";P;";EXIT;";
0070 REM in cm: REM focal pt: F=25: REM start: S=5: REM end: E=34
0071 REM sector 1: L=.4: H=1.0: O=1.0: GOSUB 100
0072 REM sector 2: L=.2: H=1.0: O=.0 : GOSUB 100
0073 REM sector 3: L=.1: H=-1.0: O=.0 : GOSUB 100
0074 REM sector 4: L=.05: H=-1.0: O=-1.0 : GOSUB 100
0090 PRINT "!R!PAGE;EXIT;": SELECT PRINT 005(80): STOP "DONE"
0100 REM subroutine draws one sector: D=S
0110 IF D>E THEN RETURN : REM width bar: W=D*L/F: REM end bar: N=D+W-2*P: REM height bar: H1=D*H/F: REM elevation: H3=D*O/F
0120 PRINT "!R!MAP";X+H3;",";Y+D-S;";EXIT;";
0130 FOR D=D TO N STEP P/2: REM relative D: D1=D-S: H2=H1+H3
0140 PRINT "!R!DAP";X+H3;",";Y+D1;";DAP";X+H2;",";Y+D1+P/2;";EXIT;";
0150 NEXT D: D=D+W: GOTO 110
9988 REM RESAVET "ILLUSION"
-- Seafriends home -- UW
photography index -- Rev 20051018,