Seismic Toolkit

Plot SAC ASCII and SAC_BIN data format, with zoom,unzoom, plot channel by channel, or plot all channel (until 2000, tested with 125).

Filtering the data: all filters are causal recursive IFR (Infinite Impulse Response) written using the bilinear Z-transform in the time domain. Their conception using a few number of coefficents gives them very fast with a low memory cost.

The adaptation factor of frequency warants no deformation in the frequency domain of the transfert function. The main filters used are the following: Butterworth High-Pass and Low-Pass (n order), Farrer 10s-6s Low-Pass (a combination of rejector and Low-Pass specially designed for removing oceanic noise), Integrator, Derivator, Integrator with cut-off frequency, Derivator with cut-off frequency, Trend removing, Rejector (n-order), Envelop with Hilbert (not recursive at all), compensator of (n-order), Polynomial filter (n-order, not recursive at all ).

Data plotting : channel by channel, all channels, zoom, unzoom, unfilter, instantaneous time and amplitude informations with mouse pointer.

Fourier domain: Power Spectral Density (PSD) in linear-linear, log-log axes; independant windows for each channel, instantaneous frequency and amplitude informations with mouse pointer, zoom, unzoom of spectra. Dirac, Hilbert transform, Time-Frequency representation (tested until 1 million of points per channel on 3 channels.

Polarization : easy and fast particule motion representation in both horizontal plane and incidence plane, with automatic computation of best direction with eigen vectors of the covariance matrix. Display of linearity and planearity coefficient.

What is new in this release:

• This release corrects a major bug in freeing memory.
• The Win32 version was especially sensitive to this bug.
• This release also includes two new features in the utility programs, in the MLINA (Minimum LINear Algebra library) section.
• The first feature is the computation of the rank of a matrix (in the igenVal_imTQL1() function).
• The second feature is a program for fitting a curve with a polynomial of any order.

What is new in version 0.69a:

• new package (.deb) for Debian - Ubuntu
• new package (.dmg) for MAC OSX 10.4 (Tiger)
• Add function Derivator _Fc (a derivator with a cutting frequency
• impose : setlocale(LC_ALL,"C") to avoid regional setting as decimal separator;

What is new in version 0.69:

• new package (.deb) for Debian - Ubuntu
• new package (.dmg) for MAC OSX 10.4 (Tiger)
• Add function Derivator _Fc (a derivator with a cutting frequency
• impose : setlocale(LC_ALL,"C") to avoid regional setting as decimal separator;
• in Utilities prog...
• tridiagonalization with implicit shift (tred2 tqli2 )for computing eigen vaued and eigen vectors
• the functions tred2 tqli2 are translated in C from Algol (procedures given in Wilkinson & Reinsch "handbook of automatic computation", vol 2, Springer Verlag 1971

What is new in version 0.68:

• Major bug corrected for reading SAC_BIN
• Minor bug corrected in MM_MAG_1_17
• Utility program : TIME_FREQX for computing time_frequency of an entire directory of files

What is new in version 0.67:

• Secure reading of wrong FORMAT to avoid crash when reading files in the wrong format
• New version of MM_MAG (1_16):
• takes into account signal/noise ratio to avoid over-estimation of magnitudes for small S/N ratio
• the validity of Mm is checked with the criterium of a minimum signal/noise ratio
• the minimum valid signal/noise ratio is defined in the configuration file MM_mag_x_xx.ini

What is new in version 0.66:

• Elementary matrix operations (product, transpose, inverse, norm), QR, LU decomposition.
• Hessenberg and tridiagonal reduction (either with Householder transforms, either with Givens rotations).
• Solving largely over-determinedlinear system of equations with QR or LU methods.
• A method for determining the rank of matrix (thus the number of independent unknowns) is proposed (based on an acceptable threshold of spectral radius).
• Eigen vectors and values (that take a fundamental place in polarization analysis) are computed with iterative QR-shift algorithm.

What is new in version 0.65:

• New version of PPOL_MTPA: 1.1
• New version of POLAR: 1.1
• Major bug of compilation fixed in PPOL_MTPA
• Furthermore, it has been discovered a few days ago, that 4 (four) functions included in POLAR_MTPA and POLAR 0.0.x used are coming from a code under very restrictive license, and thus must not to be distributed, of course (it is not GNU License). Consequently previous versions to STK_0.65 have been removed from the repository for downloading. Also, starting from STK_0.65, the code of the POLAR_MTPA_1.1 and POLAR_1.1 has been cleaned. Users must also delete the previous versions of POLAR_MTPA and POLAR; Users must use only versions equal or greater to POLAR_MTPA_1.1 and POLAR_1.1. The WIN32 binary versions does NOT contains the POLAR_MTPA and POLAR programs,thus there is NO problem.

What is new in version 0.64:

• The SAC_BIN2ASCII program was added for converting an entire directory of files in SAC_BIN format into SAC_ASCII.
• The MATH package was added, which contains minimum linear algebra (LU, QR, eigen values, eigen vectors, linear solutions).
• It is still under construction and not appropriate for production.

What is new in version 0.63:

• A new function was added for removing estimated signal from a Fourier Series.
• This functionnality is useful for removing periodic noise without filtering, thus without phase shift.
• New menu buttons were added in the ALL Channel and NZE window.