Filter and Transform Data

All data transformations can be previewed before being applied, so you can interactively tune transformation parameters to achieve the desired effect..

Available options include:

• Unitary transformations
  One or more channels can be offset (including de-meaning/medianing), rectified, scaled or inverted.

• Filtering
  Frequency-domain filtering can be performed with a FIR filter of up to 200-order. Time-domain filtering is available with RC-type filters, notch filters or moving-average smoothing (multi-pass boxcar, Gaussian) and robust locally-weighted polynomial smoothing (LOWESS).

• Local energy
  Variance, power and Teager Operator local energy measures are available

• File transformation
  Whole data files can be shortened (by removing data from either end), have their sample rates reduced (by decimation or averaging intervening values), have their sample rates increased (by zero-fill interpolation and low-pass filtering), or be concatenated with other files with similar acquisition parameters.

• Channel arithmetic
  Individual channels can be added together, subtracted from each other, removed from the record altogether, or copied from one file to another. User-defined arithmetic operations involving channel values, the current time, and a wide range of standard functions (sine, tan, cos, log, ln, power etc) can be performed using a built-in equation parser.

• De-buzz
  A sequential data-segment subtraction mode for removing mains interference is provided. This can deal with both 50 and 60 Hz interference, and episodic or continuous data. Limit cursors prevent large data excursions such as artefacts influencing the outcome.

  The top two channels show extracellular recordings of the metachronal swimmeret rhythm in crayfish. The bottom channel shows the same data after rectification and smoothing with 4 iterations of a boxcar moving average (half-window = 20 ms). The traces in this channel are superimposed.

A data record heavily contaminated with mains interference (top trace) has been "de-buzzed" by sequential subtraction of the averaged interference record.