findBadData automatically detects noise or bad data across selected channels.
findBadData is a sophisticated Shogun Post function that provides automatic noise identification. In automated script functions, you can call out a series of separate
findBadData operations, each with varying parameters.
Each operation will in its turn identify different types (higher frequency vs. lower frequency, for instance) of "bad" data that may then be handled separately with the appropriate applications of filter, cutKeys, and/or fillGaps commands.
The command uses the same data identification routines found in the filter command coupled with a
findBadData threshold parameter to adjust the effect. As you decrease or tighten the
findBadData threshold, the greater the number of "bad" keys will be selected, all other parameters being equal.
|Useful range 0.01 to 0.45|
|This value is expressed in millimeters|
|Name||Flag arguments||Argument type||Exclusive to||Comments|
|0||—||ranges||Select all keys in the play range that are potentially "bad".|
|0||—||all||Select all keys in selected time ranges that are potentially "bad".|
|2||integer||—||Select the indicated [int] number of keys on either side of each bad key found.|
findBadData -all -select 1 1 5 0.1 35;
// This execution of the findBadData command will effectively
// use the filter settings 0.1 [cutoff frequency] and 35 [threshold]
// to identify noisy segments within the entire frame range of the
// translation properties of the LFWT Marker node. Further, a
// findBadData threshold of 5 (a value that expresses a variance
// allowance from the provisionally filtered curve) will refine
// the filter selection. The '-select 1 1' option adds 1 keyframe on
// the front and 1 on the end of each noisy segment that is identified.
// To increase the number of keys selected in this instance, decrease
// the value of the findBadData threshold. Lowering the threshold will
// loosen the tolerance and increasing the value will tighten the