IF an object is to be photographed it must be luminous —that is to say, it must either be a source of light or must reflect light emanating from some other source. The surface of such an object may then be considered to consist of an infinite number of luminous points, each emitting light in straight lines in all possible directions.
If in front of a luminous object, o in Fig. i, we place a screen a, perforated at h with a small pin-hole, then a certain portion of the light emitted from each luminous point of the object will find its way through the pin-hole and produce a spot of light upon a second screen b. Thus a point p will be represented by a spot at q, and another object point p by a spot at q, so that from q to q we shall have an image of the object from p to p. But these spots will be appreciably large, however small the pin-hole may be, hence two very slightly separated object points will be represented by two overlapping spots between which it will be impossible to distinguish. Therefore very fine detail cannot be sharply defined with a pin-hole, though finer definition is possible than is generally supposed by those who have had no experience of pin-hole work.
The pin-hole has also the defect of transmitting such a small quantity of light that very long exposure is necessary to produce any developable effect on a sensitive plate, but it has the virtue of always drawing form correctly and in true perspective. Fig. i shows that all the little beams of light that reach the image screen must pass through the pin-hole and intersect one another at the point h, and, this being the case, the image is a true geometrical representation of the object, whether the latter is a plane or a solid. If the object be solid the image is in " plane perspective," while h is called its " station point," and perspective is always true if a true station point exists.