NSF Collaborative Grant (JHU/Purdue/ASU) - sole PI at ASU site : A. Spanias): J-DSP in  Astronomical Time-Scale Measurements "An Astronomical-Calibrated Time Scale for the  Mesozoic Era," $184k (ASU portion - 36 months), NSF Award 0719714.

Description of the developed functions

Currently available functions are the Earth Signal Generator, Depth-Time Transformation, Interpolation & Re-Sampling and Window.

Other functions that will soon become a part of J-DSP/ESE are filters (including Taner and Gaussian filters), that form a major part of time series analysis, spectral functions (including multi-taper method) and time-frequency analysis (spectrogram). Please continue checking the page for updates.

   Earth Signal Generator

      The J-DSP Earth Signal Generator has in-built earth system signals that correspond to samples collected from the natural phenomena. Three signals incorporated into the Earth  Signal Generator correspond to Gaussian (SPECMAP), non-Gaussian (Lake Baikal) and Trubi Marls data sets. The Earth Signal Generator has the capability to output the signals frame-by-frame at the user-defined frame length or as a whole. When using frame-by-frame data, the user has the option to move to any frame using the buttons for playback control. There is also an intuitive panel that shows a preview of the data of the current frame.

Figure 1. Earth Signal Generator

   Depth-Time Transformation

      Typically the data obtained from the signal generator are in the depth scale and each depth corresponds to a particular time in the actual time series. Depth-time transformation uses age models that relate the depth to a particular time of occurrence. Currently, the J-DSP/ESE applies the age model tables and performs the transformation. The depth scale is converted to the time scale and the output of the block carries the data in time scale.

Figure 2. Depth-Time Transformation using Age Models

   Interpolation & Re-sampling

      Most of the times, the data is uniformly sampled with respect to depth whereas the time scale obtained after the transformation is not uniform. But to apply the signal processing techniques, we need the time series to be uniformly sampled. So, we perform interpolation and re-sampling in this block to obtain uniformly sampled time series. Right now, the block can perform linear or quadratic interpolation. The block also re-samples the data uniformly to the number of samples specified by the user.

Figure 3. Interpolation & Re-sampling of time series

   Window

       After the data is re-sampled it is passed through the window block where it gets multiplied by the appropriate window. Apart from the windows that already exist in the standard J-DSP version, two new windows, Gaussian and Tukey, which are extensively used by geoscientists, have been incorporated in the window block.

Figure 4. Windowing the data