After the discussion about using CRT displays as tachistoscopes it is probably worth saying something about when the light is actually emitted. As described earlier in the chapter, the face of the screen is scanned by the electron beam from left to right and top to bottom. As the beam impinges on each small area of phosphor a pulse of light is emitted which typically rises quickly to a peak value before decaying to nothing over a period of about 1ms. Different phosphors have different properties but this represents a typical value. Therefore in a typical display used for vision research the light that is emitted from the top left hand corner of the screen happens about one frame time or 10ms before that emitted from the bottom right hand corner. The implication here is that any experiment that looks at the temporal response to a full screen spatial stimulus is effectively convolving it with a 10ms time window with the effect being to limit the upper frequency response. This is especially obvious when making evoked potential recordings where the electrical potential being recorded is the sum of the contributions from the retinal cells all of which have been stimulated at a slightly different time. Help is at hand however for those interested in the ultimate recordings simply by using the techniques described in the section about tachistoscopes. Consider displaying a chess (checker) board that counter-phases at 1Hz on a 100Hz frame rate monitor so that each of the squares comprises 100 pulses of light followed by 100 no-pulses. In an uncorrected system the white squares at the top and bottom of the screen produce their pulses of light on the same frame and hence suffer a one frame or 10ms phase delay between them. To perform the time correction simply divide the screen into horizontal bands, any convenient number will do but think along the lines of eight, and advance in time the stimuli in the lower bands by reducing the luminance of the last frame in the on sequence and adding a corresponding amount to the last frame in off sequence. Do this in proportion to the distance of the band from the top of the screen. In this way, a square that lies in a band half way up the screen will have a frame sequence consisting of 99 fully on frames followed by one frame at half luminance followed by 99 fully off frames followed by another frame at half luminance. The total energy of the stimulus remains at 100/200 but the time has been shifted by half a frame.

Figure 12 Moving a stimulus closer to a mask

^ Back to top