Andreas,
Good discussion.
I believe the limit is at roughly 100fps. Just set your monitor refresh rate to 60 Hz and compare it with 100 Hz.
The frame rate that guys like Tom and Anand measure is I believe the frame rate of how fast data get's written to the memory buffer of the video card (on average as you point out). On the other side, this data get's written by the RAMDAC to generate analog output of the video card to the analog monitor.
The ideal situation would be if the graphics card can generate data at the same frame rate as the refresh rate of the monitor, never dropping below this floor. You don't really care a frame rate any higher than that, since the information is not going anywhere. It stays in the video card. If the frame rate is 150 fps and monitor refresh rate is only 75 Hz, every other frame goes basically nowhere. It doesn't leave the video card.
Your comparison of different monitor refresh rates and your perception of them is I think a different argument, and it goes into how a human eye reacts to an image being drawn on CRT with various refresh rates. The human eye is not too crazy about anything below 70 Hz. You can have a monitor with refresh rate of 75 Hz (which will not hurt your eyes, while the monitor is displaying something with a frame rate of 25fps, which you may perceive as sluggish, but it will not hurt your eyes.
If you use a TFT display it's even worse (usually they can only be refreshed between 20 and 30 times per second).
I was under the impression that the common refresh rate was 60 Hz. But the way TFT generates image is different, since each pixel generates persistent light, as opposed to CRT, where the spot that represents a pixel gets hit by the ray, is at it's maximum intensity as it is hit, and slowly goes down in the intensity of the light for the next 1/60 of a second (in case of 60 Hz), which basically generates blinking at very fast pace. The faster the blinking is, the more the eye and the brain get fooled into thinking that they are looking at a persistent light.
Joe |
