Re: Camera obscura

[Bernard Cleyet <anngeorg@PACBELL.NET>]

A common student exercise is to find the size of the pinhole which gives
optimum resolution, i.e. when the two sources of "unsharpness" are equal.


For LK's camera I obtain 0.71 mm as the optimum diameter for an object
at "infinity".  For one at one meter, 0.37 mm

(I assume monochromatic light w/ half micron wave length)

Continuing for the "webcam"  w/o lens:      I assume ~one inch "focal
length";   three cm.  using the size of the Airy disk and Raleigh's
criterion, I obtain twelve min. resolution.  For a three cm FL, no point
in having the pixels (much) smaller than 0.1 mm on a side.    A typical
cheapo (PC-164C- SUPERCIRCUITS) has ~ 500 pixels Horiz. in 1/3 inch or ~
10X finer resolution.  I strongly recommend using a lens and a larger
"hole".

bc   who makes, often, mistakes


Ludwik Kowalski wrote:

> On Saturday, Aug 16, 2003, Bob Sciamanda wrote:
>
>
>
> > . . . some of the low priced web cameras on
> > the market today use a pinhole in lieu of a lens.
>
> That is new to me. It can be a good topic to
> discuss in a class. Knowing how small pixels
> are, and other geometric parameters, one
> would suspect this to be impossible. The
> pinhole providing a single-pixel geometrical
> resolution is likely to be so small that the
> diffraction would prevent the expected
> resolution. I am probably wrong; in the past
> I did take good pinhole camera pictures on a
> film. But that film was 35 cm from the pinhole.
> The web cameras are probably several times
> shorter.
>
> The issue is worth addressing numerically, for
> example, as a problem in a textbook, or in class.
> Keep in mind that a pinhole too small for light
> is not necessarily too small for X-rays.
> Ludwik Kowalski