- A single complex data sample for each phone and circsum at the offset time is output.
- A single complex data sample at the offset time, corrected for phone position as projected on a vector from the coordinate system origin to the source, is also output.
- A window of data around the offset time is transformed into the frequency domain and the DC bin is output. (Data has been demodulated so the DC bin corresponds to 28 or 84 Hz depending on the input file).
- A window of data around the offset time, again corrected for each phone's mooring motion by projection, is ffted, and the DC bin is returned.
- Hawaii short files, 28Hz
- Hawaii long files, 28Hz
- Hawaii short files, 84Hz
- Hawaii long files, 84Hz
- Kiritimati short files, 28Hz
- Kiritimati long files, 28Hz
- Kiritimati short files, 84Hz
- Kiritimati long files, 84Hz
The first image shows the magnitude and phase of one of the circsums for a Kiritimati file without any clock corrections:
The next image shows the magnitude and phase of one of the circsums for a Kiritimati file after applying clock corrections:
Although the clock corrections did a good job in lining up the magnitude of the data, there was still a very noticeable "seam" in the phase running through the data at the interface of the shallow and deep arrays. The following plot shows the raw phase difference between the bottom phone on the shallow array and the top phone on the deep array for an early file. There is an approximate average phase difference near pi/2 between the two.
After plotting a number of these for a series of files for the days of interest, there is a definite trend of increasing phase difference that can be seen. A straight line was fitted to the approximate phase differences, and interpolated for the short files:
The interpolation was also done for the times of the long files:
Finally, a correction was applied to the the phase of the lower array based on the filetime. Here is the final corrected result for the file shown above:
... and corrected 84Hz data:
- for Hawaii: DoubleSum_AST_Hawaii_Shal.m
- for Kiritimati: DoubleSum_AST_Kiritimati_DeepShal.m.
A finale time for each transmission was interactively picked by clicking on a point on a plot of the doubly-summed (over all phones and circs) via program AST_PlotAndPickFinale3.m. Here is an example of a double sum plot:
Finale times were picked for all eight categories of files mentioned above, and then script ASTMake10x40_MM2.m was run to implement the finale time/ offset processing discussed earlier. An offset of 1 second before finale time was used. A finale script, AST_Package_Results.m was used to massage the data into a format for use in further processing.
We end with examples of the final outputs.
