SigMT Phoenix

SigMT Phoenix User Guide

Install Anaconda

I recommend installing Anaconda for using SigMT. Download Anaconda here.

Install SigMT

Option 1: Install from pip

Open Anaconda Prompt and run:

        > pip install sigmt

Option 2: Install from GitHub

Open Anaconda Prompt, create a new environment, and clone the repository:

        > conda create --name sigmt
        > conda activate sigmt
        > git clone https://github.com/ajithabhks/sigmt.git
        > cd sigmt
        > pip install -e .

Update SigMT

Option 1: Update pip installation

Open Anaconda Prompt and run:

      > pip install --upgrade sigmt

Option 2: Update GitHub installation

If SigMT was installed by cloning the GitHub repository, open Anaconda Prompt, activate the environment, navigate to the SigMT repository, and pull the latest changes:

      > conda activate sigmt
      > cd sigmt
      > git pull
      > pip install -e .

Opening SigMT

Open Anaconda Prompt.
If Option 1 (Install from pip) is used for installation. Type:

        > sigmt

Open Anaconda Prompt.
If Option 2 (Install from GitHub) is used for installation. Type:

        > conda activate sigmt
        > sigmt

SigMT Main Window

Choose Phoenix (MTU-5C) from the main window.

Phoenix Processing Window

This is the main processing window for Phoenix.

Create Project

First, create a project to store time-series data and calibration files required for processing.

Click File → Create a Project.
Choose a folder where all project-related files will be stored.

A new pop-up window will appear. Fill in the required fields for the project, then click OK.

Project Folder

Go to the project folder. Keep the calibration files (.json) files in the 'calibration_files' folder.
Please make sure that the file name should be the serial number of coil.

Keep time series data in time_series folder. It should be following order.
time_series → Station1 (station name) → 12345_2016-04-29-202957 (recording)
Please let me know if you need any changes in this structure.
If everything in order, you will the station list in Phoenix window as following.

Processing Explained

Click on 'Load Sites' to get list of sites in the 'time_series' folder.
Choose 'Local' and/or 'Remote' station.
Select Sampling Frequency from drop down.
Click 'Read time series' button.
If you need Decimation, select the decimation (5) and click 'Decimate'. New sampling frequency will be old/decimation_value (For eg: 150/5 = 30 Hz).
Choose required parameters such as 'Notch filter off or on', 'FFT Length', 'Parzen window radius', 'Mahalanobis Distance threshold' and click 'Save Parameters'.
⚠ Parameter Selection Tips
SigMT gives you full control over these processing parameters. Use this flexibility to test different settings rather than relying on a single combination of parameters.
- FFT Length: If you are struggling to obtain sounding curve points at lower frequencies, try increasing the FFT Length.
- Parzen Window Radius: A larger Parzen window radius may produce a smoother-looking sounding curve, but excessive smoothing can introduce a downward bias in the result. Start with a smaller Parzen window radius and increase it gradually, and make sure you are not biasing the results.
Click 'Perform Band Averaging'.
Choose coherency threshold and 'Min % of Data' (Minimum percentage of data) if required. Min % of Data makes sure that atleast that much percentage data will be available after coherency based rejection. It helps to prevent crashing of robust regression.
Choose polarization direction thresholding to reject data based on it. Read SigMT paper for more information. If PD min = -10 and PD max = 10, then the data between this range will be rejected.
Click 'Perform Robust Estimation' to perform robust regression.
You can see plots by clicking button in 'Plots section'.

Save EDI

Click on 'Save as EDI' button to save EDI files. Save it in 'edi' folder in project folder or wherever you like. Keep it a folder with site name (in below example: KL33A). In SigMT, the idea is to create seperate EDI files for each sampling frequency. For example: 2048 Hz, 512 Hz, etc. Then later joins it using the 'EDI Merger'.

EDI Merger

Click 'Open EDI merger' to open EDI merger.
Click on 'Select EDI Files' and go to the folder where EDIs are stored.
Select the EDI files that need to be merged and click 'Open'. Click on 'Plot Data' to manually delete bad data points and duplicate data points.
IMPORTANT: PLEASE SURE THAT THERE EXIST NO TWO DATA POINTS FOR SAME FREQUENCY
The plot will appear as below. Data points will have different colors based on the sampling frequency. USE ONE LEFT CLICK to delete a point.
USE ONE RIGHT CLICK to undo once (only once!!).
After cleaning, it may look like: Keep the window plot open and click 'Save as EDI' to save final merged EDI'. You can use site name for the final EDI. DONE!!