Sacpy is a simple python module that deals with binary SAC files.
- python2 or python3
- numpy
- scipy
To use sacpy, the sacpy directory must be placed in a path pointed by the PYTHONPATH environment variable.
To use sacpy, simply import the sacpy module
import sacpy
or for example
from sacpy import Sac
(Sac is the class used to manipulate SAC files)
In the Sac class, attributes have the same name as SAC header variables (see the SAC documentation). For example, the number of data points is given by sacobj.npts. The data points are in sacobj.depvar
You can read binary SAC files using
sacobj = sacpy.Sac()
sacobj.read("SAC_FILENAME")
or
sacobj = sacpy.Sac("SAC_FILENAME")
In the first case, we first instantiate a Sac object and then read the SAC file. In the second case, we instantiate and read the SAC file on the fly.
You can write binary SAC files using
sacobj.write("SAC_FILENAME")
To (deep) copy a Sac object sacobj in a new sacobjcopy, you can use:
sacobjcopy = sacobj.copy()
If you have 2 Sac objects sacobj1 and sacobj2 (including the same number of samples), you can add, substract, multiply waveforms of the 2 files using:
sacobj3=sacobj1+sacobj2
sacobj3=sacobj1-sacobj2
sacobj3=sacobj1*sacobj2
(data points in sacobj3 will be the sum, substraction and multiplication of sacobj1 and sacobj2).
To perform time-integration, you can use:
sacobj.integrate()
To interpolate the data trace to a new sampling step:
sacobj.interpolate(delta)
where delta is the new sampling step (after interpolation).
We use a sinc interpolation to avoid aliasing issues. This result in longer computation time than a linear interpolation.
To decimate the data:
sacobj.decimate(decimation_factor)
Currently, only the following decimation factors are available: 1, 2, 3, 5, 10, 20, 25, 30, 40, 50, 60, 75, 80, 90, 100, 120 Decimation includes a proper anti-aliasing FIR filter.
To filter the data:
sacobj.filter(freq, order, btype)
Applies a butterworth filter to the data. freq is the filter corner frequencie(s) (scalar or list of 2 scalars).
btype can be 'lowpass', 'highpass', 'bandpass' and 'bandstop'. Default values are order=4 and btype='lowpass'.
To deconvolve, the instrument response using a dictionary of poles and zeros (PZ).
The PZ dictionary can be built from a poles and zeros file using the function readPZ as follows:
import sacpy
s = sacpy.Sac('SAC_FILENAME')
PZ = readPZ('SAC_PZs_FILENAME')
filtfreq = [0.001,0.005,0.01 0.1]
s.deconvresp(PZ,filtfreq)
To deal with the zero response at zero frequency, it is usually necessary to filter the signal during deconvolution. filtfreq can be used for that by defining four frequencies (f1<f2<f3<f4). Taper is 1 between f2 and f3 and 0 below f1 and above f4. Frequencies f1 and f2 specify the high-pass filter while f3 and f4 specify the low-pass filter. The filters applied between f1 and f2 and between f3 and f4 are quarter cycles of a cosine wave.
To get the reference datetime, you can use:
sacobj.getnzdatetime()
Similarly, to get the origin, begin, end or arrival datetimes:
otime = sacobj.getodatetime()
btime = sacobj.getbdatetime()
etime = sacobj.getarrivaldatetimes()
To get a time-vector from the reference datetime:
timevec = sacobj.time()
To set the origin time, you can use:
sacobj.setotime(otime)
where otime is a datetime instance
To set arrival times, you can use:
sacobj.setarrivaltimes(phase_dict)
where phase_dict is a phase pick dictionary {phase_name: arrival_datetime)
To do a fourrier transform, you can use:
sacobj.fft()
The Fourier transform is done using the function numpy.fft.rfft the result will be returned and also stored in self.depvar.
If you want to get the corresponding frequency:
freqvec = sacobj.freq()
To plot the seismogram with matplotlib.pyplot:
sacobj.plot()
To plot the amplitude spectum of the seismogram:
sacobj.fft()
sacobj.plot()