The zipped movie shows some lag testing. Dell1913 TN panel, framedelay 9, filmed on LGv10 at 120fps, arduino keyboard encoder lights up the red LEDs when fire is pressed (and yellow for up or down). Note: the LEDs appear across more frames than they should (programmed for 17ms) that may be persistence in the CCD? or maybe the switch off is slow.
I count the response in video frames (= 8.3ms) from 0 when the LED comes on to N when the laser shot appears. Typically I get values of 3,4,5,6 which is in line with what I was expecting and it caused me to assume a value of about 10ms for the lag in the LCD (signal processing +pixel response).
However I got a frame 2 response (at 0.82s in the video) and the only way I can explain that is by cutting my assumed display lag to about 5ms.
Attached is my theoretical explanation diagram.
Assuming my understanding is correct, I'd expect better response at the top of the screen (I'm pretty sure the LCD is lit from top to bottom) but I get about 1 frame count worse on average. I've got a slight suspicion that there could be something dicky with the emulation so that the top action drops into the next frame of output (Defender pcb polls every 8.3ms) but I can't guess why yet, that's work for later.
Any comments? or anyone else interested in LCD performance ?
I've got dreams of chopping 16ms off the average time; 8ms via emulation change (8ms polling), and 8ms via 120fps display.