How do FROG and SPIDER compare in accuracy?
While FROG and GRENOUILLE have been shown to be very accurate in numerous tests, SPIDER has rarely been shown to yield precise results. Indeed, it usually yields a shorter pulse than FROG does. Not surprisingly, it has recently been shown that SPIDER is so inaccurate that it shouldnt be trusted for chirp or pulse-length measurements (Birge, J.R., R. Ell, and F.X. Kärtner, Opt. Lett., 2006. 31(13): p. 2063-5). Birge, et al., showed that the error in a SPIDER-measured pulse length is at least 100 times the error in the separation between the internally generated double pulse required for the measurement. This means that this pulse separation must be known with few-attosecond accuracy just to measure a pulse to 10% accuracy—the generally unacceptable accuracy of autocorrelation. Because this pulse separation can easily drift by many attoseconds during a measurement, and there are no checks on such measurements, SPIDER measurements cannot be trusted. A unfortunate problem is that SPIDER us
