The SignalDB is the name of the database that contains the conditions and preliminary characterizations of the raw SETI data. You may use this information to help narrow your search of data to analyze, or use the values as features in a machine-learning context.
This database is generated by SETI with their data acquisition software, which you may have seen mentioned elsewhere on this site, called SonATA, which stands for "SETI On the ATA". A detailed description of how the ATA and SonATA work is found on the SETIQuest Wiki
Besides the descriptions below, to better understand what these fields mean, it useful to read the How Observing Works page on the SETIQuest website. Unfortunately no direct link
to that page exists, so you'll have to find the link on the left panel of that page.
They are delievered to you in SETI@IBMCloud via the /v1/aca/meta/{ra}/{dec} endpoint.
| Field | Description |
|---|---|
| uniqueid | Unique ID |
| time | Date/Time of observation |
| acttyp | Activity Type: initial or followup |
| catalog | Target catalog name |
| tgtid | Target Number in Catalog |
| ra2000hr | Right Ascension (hours) |
| dec2000deg | Delination (degrees) |
| power | Raw Power Measure |
| snr | Signal/Noise |
| freqmhz | Start frequency of observed signal (MHz) |
| drifthzs | Signal Drift (Hz/s) |
| widhz | Signal Bandwidth (Hz) |
| pol | Polarization of Signal |
| sigtyp | Signal Type |
| pperiods | Pulse Period (s) |
| npul | Number of pulses |
| inttimes | Integration Time |
| tscpazdeg | Primary telescope azimuth direction (degrees) |
| tscpeldeg | Primary telescope elevation direction (degrees) |
| beamno | Beam signal number |
| sigclass | Signal Class |
| sigreason | Signal Reason |
| container | The name of the Object Storage container in which the raw data is stored |
| objectname | The name of the object in Object Storage in which the raw data is stored |
The date-time of the observation is recorded in ISO format 'yyyy-mm-ddThh:mm:ssZ'
This is the "Activity Type" of the observation. There are a number of possible values:
targettargetN-ontargetNofftarget5-on-nofollowup- "other"
for N = 1 - 5 and where "other" could be a list of special activity types to designate calibrations and special tests.
When the ATA makes an initial observation of a candidate signal, the SonATA software then
decides to look again for the signal. This second observation is called 'target1-on'. The ATA
then looks elsewhere, called target1off. If a signal is not observed in target1off,
the ATA then again looks at the original location. An observation of a candidate signal would
then be called target2-on. This happens repeatedly. Each time a potential candidate signal is
re-observed, the N is increased, up to a maximum of 5. Search your data for acttyp for "large" N.
More information is found on the "How Observing Works" tab at http://setiquest.info/.
Each observed location in the sky corresponds to a particular known object, or target. Each target has an identification number, which is stored in a particular external database or catalog. T
The total power of the signal (in arbitrary units relative to the average noise)
The signal-to-noise ratio is only reported for candidate events.
This is the measured frequency, in MHz, of the signal observed in the spectrogram at t=0.
This is the Doppler drift of the observed signal, in units of Hz/second. This is due to an acceleration between the ATA and the source of the signal. (And because the Earth is rotating, the ATA is accelerating relative to everything that is not on the Earth or not in geostationary orbit. Thus, signals with zero Doppler Drift are considered RFI.)
This is the size of the signal bandwidth, in Hz -- it's the width of the signal in the spectrogram.
The primary polarity of the observed signal.
Values can be
- left
- right
- both
- mixed
It should be noted that left and right polarizations are misnomers. It was the intention of the SETI group
to observe left- and right-circularly polarized signals. However, that never came to fruition. Instead the
antenna at the ATA observe a long the horizontal and vertical polarizations (relative to the Earth's surface
at the observatory). The left and right direcitons are really the horizontal and vertical polarization, respectively.
The signal type can be
- CwP = carrier wave power
- CwC = carrier wave coherent
- Pul = pulsed signal
These are the three main signal types that SonATA looks for duing data aquisition and signals are assigned
one of these types. A CwP signal is a signal that has strong power at a particular frequency (allowing
for Doppler drift). A CwC signal is a signal that has a coherent signal phase. A Pul is signal
that appears the be a pulse CwP signal; it has strong power at Doppler drifting frequency but it comes
and goes within the observation time (~94 seconds). To put it another way, it looks like a typical
narrow-band signal except that it has a time-dependent amplitude such that it's amplitude drops to
zero multiple times during the observation.
The observed periodicity of the pulsed signal. Only for sigtyp = Pul.
Number of pulses observed. Only for sigtyp = Pul.
Total time of observation that resulted in this signal. This should correspond to the total amount of time in the spectrogram.
These tell the direction of the primary telescope in the ATA array at the time of observation.
The tscpazdeg refers to the azimuthal direction relative to North and tscpeldeg is the elevation
above the local horizon.
The ATA can record signals from three different and precise locations in the sky. First, the ATA array is pointed in a general direction of the sky. Then, by combining the data from the array of telescopes in a particular way, one can observe a signal from just a very precise location. This process is called beamforming, and the resulting data is called a "beam". The ATA can observe three beams simultaneously. This records which beam number (1, 2, or 3) produced the signal in the associated raw data file.
This is the Signal Classification. For achive-compamp files, which are the only types of files we are currently
making available, this value will always be Candidate. If we make other file types available in the future,
they will be labeled RFI (radio frequency interference) or Unkn (unknown).
This is the justification for the classification. The values could be
- PsPwrT. Passed power threshold, passed on to candidate status
- Dft2Hi. Drift too high, rejected
- 2MnyCnd. Too many candidates, the system flooded and aborted, rejected
- ZeroDft. Signal has essentially zero drift, therefore most likely terrestrial, rejected
- SNtInCh. Signal drifted out of channel, could not follow up, unresolved
- RctRFI. Signal was found in recent RFI database, rejected
Since all of the archive-compamp files are Candidate events, the sigreason for those files PsPwrT.
This is sub-categorization for Candidate signals
- PsPwrT. For pulses, signal passes on to next follow up
- PsCohD. For carrier wave, signal passes on to a follow up
- ZeroDft. See above, rejected
- SnMulBm. Seen in multiple beams, rejected
- Confrm. Signal seen in only one beam above threshold, pass on to a follow up
- Dft2Hi. See above, rejected
- SNoSigl. Signal disappeared and was not seen in secondary beam either. No follow up, but not rejected.
- SSawSig. Signal disappeared and was seen in secondary beam, rejected.
- NoSignl. Signal disappeared in followup. No follow up, but not rejected
- RConfrm. Signal was confirmed in followup, passes on to next follow up.
- SeenOff. Sognal was seen in an off observation, rejected
- RctRFI. See above
The raw SETI data for this project are stored in an IBM Object Storage instance on IBM Bluemix. In Object Storage, "objects" are stored in containers. Objects can be arbitrary "bags of bits". In our particular case, however, we store each compamp or archive-compamp file as a separate object. (We do this for a special reason due to the nature of this project, but if you're using Object Storage yourself, you'll probably want to consider storing data inside large objects, since this is a more efficient way for Spark to ingest the data.)