dfx
Contributor
Me and my wife both have Mares Puck computers and I'm fairly certain that they can't do anything other than plain beep. Somebody else's computer or camera was my first guess too, and I was hoping somebody could identify the make or model from the sound ("oh, that's a $X $Y!"). But I guess it's not so easyLooking at this there are exactly 56 seconds between each of sounds 1 to 4 so they plot linearly - with a change then to exactly 32 seconds between sounds from that point onwards - again plotting linearly, so there is precise timing, not a random effect. The sound also appears electronically produced to me, rather than mechanical. The timing will rule out regulator noise - no-ones breathing timing is that precise.
This re-enforces my feeling that it is like some sort of watch or dive computer alarm, which when not acknowledged or cancelled changed it repetition rate from once a minute to once every 30 seconds. OK 56 seconds and 32 are not quite once a minute and once very 30 seconds but I think with the underwater effects that is what you are hearing.
In the past I have heard sounds from dive computers and wondered whose it was before realising it was actually my own, because sound does funny things underwater. The sound you are hearing is fairly high frequency, and high frequency attenuates quickly underwater (and in air), which is why whales and other marine mammals use a very low ultra sonic signals or clicks, they travel further, so I think the source of the sound was probably not very far away from you when you were recording it.
I took a second video on the same dive (I had turned the camera off in the meantime) in a different part of the reef, and while it does contain the same sound a few times, it's very faint and barely audible. However, proximity to other divers didn't seem to matter, and neither did proximity to our dive boat, especially since we were diving with the same people and the same boat all week and never heard it again. So it did seem to be something localized, but I was unable to pinpoint any particular source.
Somebody on FB suggested that it could be a locator beacon, like one used to find black boxes. That would make sense I guess, but it's still only a wild guess.
Well, how original do you want it?I have done a fair bit of sound analysis of the calls of birds, bats, crickets and cicadas (as an ecologist - don't ask!) and what could tell you more about the source would be running the sound through a spectrographic analysis in a software package such as Batsound from Pettersen Elektronik or Raven or Raven-Lite from Cornell University - these would allow you to see if each 'signal' was identical or whether there was some variance, other than the time intervals.
I've got both packages on my laptop so if there is some way to get the original files to me I will happily have a look.
I've put the original file and some versions of the audio track here: Index of /tmp
The original file is the 500 MB MP4 video file: http://dfx.at/tmp/FILE1201.MP4
Its audio track is in AAC format. I could dump the raw audio track into a file and make it available, but that creates a file without any headers, which makes it highly unlikely that any software can open it. I know I can't.
But I could dump the audio track into a PCM WAV file. It comes out as a 150 MB file in 32-bit floating point PCM format: http://dfx.at/tmp/32-bit.wav
Since this is a bit of a silly format, I've converted to more conventional 16-bit PCM at 70 MB: http://dfx.at/tmp/16-bit.wav
Which then can be compressed down to 30 MB using FLAC: http://dfx.at/tmp/16-bit.flac
Can't say that you don't have choices