Lenticular stereogram maker

By Charles Douglas Wehner

Lenticular stereography at its best can be spectacular. It is one of the few systems that allow unaided viewing.

Although as many images as necessary can be taken of a still-life subject, simply by shifting the camera, the problems of combining them into a stereogram can be daunting.

The major problem of creating the striped image is effectively solved by the free availability of high-quality printers, and by the use of the simple LEN DOS command that is described here.

A search of the Internet has revealed that the lenticular plastic is also available.

So if LEN tickles your fancy, you can read here how to make 3D pictures with up to SIX components, and at up to 3072 dots wide.

The first thing to do is create a directory such as C:\WEHNER\LEN into which you put LEN.COM, and right click on it to create a short-cut.


Next, you can change the icon. Three different sizes are provided, as both BMP and ICO.

Then, you need to put into that directory two or more images as BMP files. These images can be L.BMP and R.BMP - for the left-eye and right-eye view, or files 1.BMP (the rightmost), 2.BMP, 3.BMP and so on up to 6.BMP, the leftmost.

It is not simply a choice of two or six. If you have NIMSLO originals, you can have FOUR such files. Three or five are also acceptable.

The files MUST be 16-megacolour types. That is, when you multiply the width by the height by THREE, and add 54, that must agree with the file size.

The pictures MUST be identical in size.

There must be NO LEN.BMP in the directory. The program will not overwrite previous work. So if you have already made a LEN.BMP you must rename it.

Now you go to the START button, or wherever you put the icon, and click on it.

A DOS box should appear showing you any error messages. If all was correct, the program will say "DONE".


You can now close the box by clicking on the X at the top right.

The result of this image-processing is a striped picture that cannot be seen in 3D without the special lenticular plastic. Nevertheless, it is conspicuously the right kind of image. Incidentally, it is Sir Charles Wheatstone, the father of stereoscopy.


It consists of pairs of stripes, where each dot has been made from the AVERAGE of two dots in each picture. So if you want to preserve the horizontal definition, you have to start with a picture that has been doubled in width and in height.

If the proportions are to remain correct, you have no other choice. Either you squash the image laterally or stretch it vertically. The choice is yours.


If you had six images, LEN will create six stripes where each dot is the average of six dots in the original.

It is at this point that we need to do some arithmetic to ensure that the raster is suitable for the available lenticular plastic.

GOEX offer lenticular plastic at 160, 90, 75 and 50 lenticles per inch. The viewing angles are 26, 30, 49 and 50 degrees respectively.

LPC (such as LPC-Europe) offer plastic at 75 and 62.

In reply to a question in the Newsgroups as to where one can obtain small quantities - rather than industrial quantities - Dr. Richard Schubert of Berlin, Germany (http://www.stereoscopicscanning.de) wrote to say that his company supplies it.

Suppose you have a printer that boasts 300 dots per inch. You can see that 300 divided by FOUR gives 75, whilst 300 by SIX gives 50. 300 divided by 5 gives 60 - but the author did not find 60 LPI plastic.

Owners of the four-lens NIMSLO camera would presumably go directly for the 75 lenticle-per-inch plastic. The rightmost image would be 1.BMP, and then you have 2.BMP 3.BMP and the leftmost, 4.BMP.

At 300 dots per inch, the pattern of four stripes will repeat every seventy-fifth of an inch.

Of course, one could have a 5.BMP and a 6.BMP in BLACK - the same size images as the true NIMSLO pictures. Then there will be FIFTY stripes per inch - but this will look better when viewed from further away. The black stripes guard against the "pseudostereo" effect.

Pseudostereo is created when the left eye sees the right eye`s image, and vice-versa. It is interesting to note that Wheatstone dwellt upon this point in his 1838 lecture. He had seen this before anybody else.

Black stripes are particularly effective if you use overhead projection film, and illuminate the image from the rear - for purposes such as shop-windows displays.

If you have only TWO images, and want to use the 75 LPI plastic, you would wish to print at 150 dots per inch - which is usually not possible. Make R.BMP into 1.BMP and 2.BMP (identical), and L.BMP into 3.BMP and 4.BMP (also identical).

Now erase R.BMP and L.BMP from the directory and run LEN. If LEN cannot see R or L, it looks for the numbered files 1 to 6, and finds four of them.

If there was one blank file, 5.BMP, it will give 60 stripes per inch. If there are two, 5.BMP and 6.BMP (or if R becomes 1, 2 and 3 while L becomes 4, 5 and 6), you will get 50 stripes per inch.

There is a question of dot bleed. Is the promised 300 dots per inch actually DELIVERED? In a Nimslo image, where there are four pictures with 18 millimetres between taking lenses, dot bleed may actually reduce the "clicking" of the image. However, if dot bleed is a nuisance, one should examine the advanced printer settings to see whether the printing scans can be interleaved with drying-out pauses.

The question of the 62 lenticle per inch plastic is interesting.

From the diagram it is clear that the printed stripe-pitch should be COARSER than the lenticle pitch. This is due to PARALLAX.


Some stereograms have to be BENT in a curve equidistant from the observer in order to be correctly viewed. Curving in this way eliminates the parallax error. The best stereograms have parallax taken into account.

When a 60 stripe per inch image is positioned behind a 62 lenticle per inch reseau, the parallax compensation is such that it is best viewed from about THIRTY times the plastic thickness. At fifteen inches viewing, the plastic would need to be half an inch thick.

It is difficult, therefore, to see how 62 LPI plastic can be applied to images made on domestic printers, with definitions of 72, 300 and 600 dots per inch.

For those who do not wish to experiment, Dr. Schubert`s stereoscopicscanning.de offers a complete lenticular kit. The author has no connection with that company, and has not tried the kit.

Images can be saved as GIF or JPEG - the two main systems at the time of writing. GIF will reduce the number of separate colours to 256, but preserves detail down to the individual dot. JPEG will degrade the image - more so when the compression is high. It should be used ONLY ONCE - at the time of completion of the image. MINIMUM compression should be used if the stereoscopic effect is not to be degraded by compression artefacts.

The most sensible thing to do is to stick to the BITMAP format. This does not degrade the image. There is a version with LZW compression - Run Length Encoding - and these RLE files sometimes take up much less room, depending on picture content.

Archive copies of finished images should be kept as BMP or RLE. Expanded GIF or JPEG saved as BMP will not do - the information loss CANNOT be reversed by expanding to BMP.

BMP files are BIG - so they should be stored on a Compact Disk (CD).

Here are the hot-links for the ZIP files. Go on the Internet to the WINZIP DOWNLOAD PAGE to install Winzip if you do not have it. Then EXTRACT to a directory such as C:\WEHNER

WINZIP ZIPPED FILE
http://wehner.org/tools/len.zip

Lenticular stereograms can be made from "flattie" images that have been reconstructed by means of MOVER.COM.

WINZIP ZIPPED FILE
http://wehner.org/tools/mover.zip

Go to http://wehner.org/tools/mover to see how to reconstruct images in stereo.

Home

(C) 2003 Charles Douglas Wehner.
Use freely but do not plagiarise.