A stretched index database correlation, like any other database correlation, involves multiple input databases where each database potentially has a different number of time states. Like a padded index database correlation, a stretched index database correlation also has as many time states as the input database with the largest number of time states. The difference between the two correlation methods is in how the input databases are mapped to a larger number of time states. The padded index database correlation method simply repeated the last frame of the input databases that needed more time states to be made even with the length of the database correlation. Stretched index database correlations on the other hand do not repeat only the last frame; they repeat frames throughout the middle time states until shorter input databases have the same number of time states as the database correlation. The effect of repeating time states throughout the middle is to evenly spread out the time states over a larger number of time states.
Stretched index database correlations are useful for comparing related related simulation databases where one simulation wrote out data at 2x, 3x, 4x, ... the frequency of another simulation. Stretched index database correlations repeat the data for smaller databases, which makes it easier to compare the databases. figure shows example databases A and B related using a stretched index database correlation. Note how the plots for both databases, even though the databases contain a different number of time states, remain roughly in sync.
![]() |
Figure 8 |