Latency measurements using Motmot. Latency measurements show that latency with the motmot software is close to the theoretical minimum possible given a particular camera interface. A) Latency measured as time of arrival of an image after a single trigger pulse was delivered to a camera. The camera (Prosilica GE-680) was set to 10 microsecond shutter integration interval and is expected to have a 5 millisecond transfer time to the computer. B) Latency estimated by correlating frame trigger time with computer clock time using motmot.fview_ext_trig. C) Latency measured by pulsing an LED at random times and calculating the delay until the luminance change was measured on the computer (blue histogram). The external trigger pulse given at 120 frames per second. Note that latency in this case includes the effects of the variable delay between the onset of the LED and the image acquisition as described in the 'Latency measurements' section. The red line is the theoretically predicted uniform distribution, with equal probability from 5 msec to 5 msec plus the inter-frame interval. D) as in C, but with external trigger pulses at 60 frames per second. E) as in C, but with external trigger pulses at 30 frames per second.