Magnetocardiographic (MCG) measurements in unshielded environment require e
fficient noise cancellation techniques. We have applied two software gradio
metry methods to analyze the time series of signal and reference data recor
ded outside magnetic shielding with high temperature superconducting quantu
m interference device (HTS SQUID) based gradiometers. One method uses adapt
ive frequency dependent gradiometer coefficients determined in the Fourier
domain to subtract the reference from the signal data. The other method com
bines recently developed techniques for nonlinear projection with propertie
s of the wavelet transform to extract noise in state space. The analyzed MC
G data sets showed improved signal-to-noise ratios for both methods as comp
ared to the data recorded with the electronic gradiometer. In this way, it
is possible to increase the bandwidth from 130 Hz for our electronic gradio
meter to 250 Hz without using any additional filtering.