Polarising filters come in a range of different types. For stereo the most common type (by far) is the linear polariser. The most desirable features of polarisers for stereo are: high transmission when two polarisers are placed parallel, and low transmission when two polarisers are crossed (at right angles). The graphs below show transmission curves for various filter types available from Polaroid versus light wavelength (red in near 700nm, and blue near 450nm).
One possible type for stereo is the HN38 which I have used successfully, but it is possible that type HN42HE could provide better performance. (It appears that the digit part of the polariser code refers to the average transmission percentage when two polarisers are placed parallel. Also, note that the y-axis is logarithmic).
Polarisers are generally placed in one of two positions in the light-path of a projector, with various advantages and dis-advantges as follows:
Measurements on a modern projector featuring an efficient condensor system, low voltage QH lamps and standard heat filter located in the condensor system showed the following (at ambient temperature of about 16degC) :
Thus the polariser reduces the gate temperature by about 15 degrees C, which is beneficial provided that the polariser life is acceptable, as it is absorbing some of the heat from the lamp. Because of this it is important to provide a cool-air flow over the polariser. A TDC projector modified for use with 250W/24V lamps displayed a gate temperture just under 30degC, and features a polariser place before the slide. A Kodak SAV series projector without polarisers showed a gate temperature of around 44degC.
The HN38 Polaroid material described in the data sheet below was measured by placing a sheet of the material into a projector (just before the slide gate). The light output was measured under two conditions:
The difference of 9 stops is equivalent to 1 part in 512, or roughly 0.2%. In other words, the brightness of a ghost image would be 9 stops duller (or 0.2%) than a high-light in a projected image (assuming no-futher depolarisation occurs from the screen, and that the polariser in the viewers glasses is of the same quality as the HN38 material).
This is a very creditable result, and is consistant with Polaroids published data, showing that the leakage in the "polarisers crossed" condition rising to exceed 0.1% towards the blue end of the spectrum (see below).