processing ancient Flexichoc source - is it futile?

[Michael_Seman]

I was looking at an old survey, that used Flexichoc source, and I have not encountered that before. On the SEG Wiki pages, there is a mention of that source. It looks roughly like a minimum-phase wavelet, but delayed about 40 ms. So getting rid of such a delay would be a stumbling block. There is no source signature. The observer reports are so brief as to be useless - it doesn't say what near trace offset was or what model of Flexichoc it was. Maybe one could look at direct arrival, but since those things were nearly 2 meters diameter, the wavelet would surely be different horizontal to nearly vertical raypaths. I don't even have any old stacks. One might use those and the water-bottom times to guess what correction was used in the original processing.
So is is a waste of time to proceed?

[GuyM]

IIRC they recorded the source signature including delay on an auxiliary channel, if any of those survived the transcription process.

Outside of that, if the water is deep enough that you have decent separation between the direct and reflected arrivals, you could take a look at the near trace plot and see how that looks. One trick we used with high-res gear where sea-swell was an issue was to digitise an event on the NTP, then smooth it, then calculate the shifts from the smoothed version to the original.

Best on a direct arrival, but you can do it on the seafloor at a push (but with obvious caveats about removing structure)

[Michael_Seman]

There are auxilliary traces, but they looked like garbage at first glance. One had a little blip that disappears on misfires, so I guess that is the source signature. It also has a spike at the start - maybe crossfeed of the firing pulse. If you mute the first 3 samples, then it looks something like a delayed source wavelet. It doesn't look very consistent though. I suppose if you stack the aux from 100 shots, you might get a useable source signature.

[GuyM]

Hmm - are you expecting the source sig to be stable? Some of those novel sources before airguns became ubiquitous were pretty unstable IIRC.
You'd have to deconvolve the data shot-by-shot reading from the auxiliary traces. It was all about supressing the bubble-pulse you got from explosives, and so variability was deemed worth that price.

[Michael_Seman]

From a description of how it worked, I thought it might be a stable source. I assumed it died out as the energy was only equivalent to 1000 cubic inches of air guns at the most. Anyway, I eventually found a report listing the near offsets, and the direct arrival on near traces seems to correspond to no delay on the source. So one less problem.