ASRG members record signals from the remote radiotelescope and transport the data to the lab. The data are then placed on servers as described on Performing SETI Data Analysis. The screening by students for signals involves calculating the FFT (fast Fourier transform) of the data repeatedly over time, which can be done by shareware programs like CoolEdit 2000, Goldwave, or other programs. Using such programs, a file is played and displayed as a spectrogram. It is the task of the students to watch the spectrogram for possible "hits." It might well be easier for students than for a program to detect the totally unexpected. Using the Analysis views in CoolEdit gives exceptional sensitivity to signals that do not shift greatly in frequency with time (even better in many cases than SETIEasy).
These are the basic steps:
Read this page thoroughly, download a spectrogram program like GoldWave if you do not already have one, download a file from the server, post an In Progress (Downloaded) message, screen the file, and post a Completed message.
First, you need to download a copy of GoldWave from www.goldwave.com.
GoldWave computes the Fast Fourier Transform (FFT) of the radiotelescope
data over time, and displays the result graphically so operators can look
for patterns in the output. The FFT procedure is the first and simplest
step in searching for patterns that may indicate intelligence in the recorded
signals. This program is Shareware, and if you use it to screen data for
more than a month, the authors require you to purchase the program. You
can install this program on either your own computer or any on-campus computer.
If you choose to do your SETI data screening on campus PCs, you will probably
need to load a new copy of GoldWave every time you change computers or every
day, because the hard drives on campus public PCs are automatically cleaned
off daily. More information about GoldWave, its data
compressors/decompressors (CODECs), and SETI can be found here.
If you run GoldWave and it does not recognize the file type you download,
you need to put the compressor
ACM file (right-click on this link and select Save As) into WINDOWS\SYSTEM
directory (Win95/98) or the same directory folder as your Goldwave.exe file
(works sometimes on Win95). You may also have to edit
your system.ini file. If so, under the label [drivers32] you'll find several
more acm file statements. Just add the statement:
msacm.msnaudio=Msnaudio.acm
Windows 98 users may also try downloading the latest version of Windows Media Player from Microsoft to solve the codec problem.
The instructions for using GoldWave are included in the file that you download. GoldWave not only allows operators to view the data, it also allows the data to be saved in formats that permit weak-signal analysis by more complex software. Now that Goldwave is installed on your machine, you should change a few settings. First, in the window entitled "GoldWave", select Tools/Device Controls. This will display the Device Control window, from which you will perform most of your data analysis. The Device Controls should look like this:
Click on the button which the red arrow points to above. Then, select the Graph tab, and click on the bubble next to "Spectrogram". Click on "Show Axis". If you have enough RAM, GoldWave will run faster if you click on Options / File and set temporary storage to RAM and clipboard to GoldWave.
Download a file from the server using a discussion group link. Use the Preferences selection in your web browser to instruct the program to Save To Disk any sound files you will screen.
Use the Preferences selection in your web browser to instruct the program to Save To Disk any sound files you will screen, or right-click on the file(s) and select "Save As" from the pop-up menu. Please check the discussion groups for follow-up posts about the files to make sure that no one else has already screened the file you choose. (This is usually not a big problem, however, as files are removed from the server once a report is received on them to make room for additional files.) Try to analyze the most recent files first, as these have the best chance of being confirmed by another station if a signal is detected. Do not bother to run files that have a follow-up post that indicates CoolEdit FFT over time, MathCAD, or "SETI-EZ" is Completed, as they have already been screened by a student. You can also skip files labeled "in progress", as they are being analyzed at that moment by someone else.
Now, you are ready to begin screening signals. Choose "New" in the GoldWave window (either by clicking the tab or by choosing File/Open. Use the browse command to find your file, and double-click on it. In the Device Controls window, click on the play button (little green triangle in the top left corner of the Device Controls window) , and you should see the file start to scroll by. A government radio and time beacon in the figure below provides an example of what signals look like. The green lines represent a large signal, and are perfectly horizontal in this example because they are terrestrial in origin and have no Doppler shift.
For most data from the radiotelescope, the screen should look something like this:
An authentic extraterrestrial signal would exhibit a Doppler shift as the signal source moved toward and away from Earth as the planet rotates and translates relative to the deep-space source. Click here for an Excel spreadsheet to calculate the expected Doppler shift from rotation for any location on earth. For our telescope in its typical orientation, the expected Doppler shift is +/-6.07 Hz/min. This means that any real "signal" would be shifted by the rotation of the earth by 6.07 Hz per minute of recording. You can slow the screen frame rate in Goldwave to exaggerate the shift. Of course, the motion of the earth in orbit about the sun, as well as any rotation or orbital motion of a remote planet, would also cause a Doppler shift in the received signal. Theoretically, a Doppler shift would look something like this:
The angle of the tilt is a little exaggerated above. Like CoolEdit, Goldwave has functions that permit dechirping of signals to improve your chances of detecting a weak signal with a frequency drift (under Effects / Doppler and Effects / Pitch). Many small signals are routinely detected in the collected data files. These are checked against the computers' Known Interferences Logs. We are most interested in having students look for complex patterns that would not integrate over time into a stronger signal, because our data collection software dechirps and integrates signals automatically. Click here for more information on eliminating terrestrial interferences before posting signal reports.
 |
This signal was recorded on another program (FFTDSP, which scans vertically instead of horizontally), but the idea is the same. Real signals will not "wiggle" back and forth. Wiggling is characteristic of interference from nearby electronics. |
 |
This picture represents an occasional SETI problem. Any signal that seems to "curve" is probably caused by an airplane or a satellite orbiting Earth. The ASRG concentrates on frequencies around 1420 MHz, which is an internationally protected frequency. However, interferences can bleed over into this region occasionally. |
Click here to see examples of more interference signals.
The reporting methods are described in Where We Have Been Looking Lately.
Every time you download a file or complete analysis of a data file, please report your work by updating the post with a follow-up post. When you download a file, follow up with the file name and "In Progress". This helps to reduce unnecessary duplication of effort by others who may download the same file. However, please do not download a large number of files at one time unless you can analyze them in fairly short order. Our goal is to keep the time a file waits for analysis to a minimum to increase the chance of another station confirming any signal that might be detected.
If you find any unusual signals, post it in a report of file completion. Report the time the signal occurred in the file and include a screen capture image of the signal as a GIF or JPEG if possible. Remember, in nearly 40 years of SETI, no signals from intelligent extra-terrestrial civilizations have been confirmed, so most suspicious signals turn out to be interference. We will share the results of our signal tests with you, and any credit that may be due based on the results of signal analyses and confirmatory testing by other Project Argus stations.
If you have any problems doing the data screening, please let us know and we can help. We appreciate your interest and time. Thank you, and good luck!