Probably the atmospheric version of SOFAR; thermal layers in the atmosphere acting as a wave guide for very low frequency sound waves. In the ocean, the sound of explosions can travel thousands of miles in the SOFAR channel.
Weather inversion where the sound the travels slightly faster and farther in different temperature air medium plus reflection of waves that would normally dissipate to upper atmosphere if I read this correctly.
http://spectrums.com/sound-and-the-weather/
Umm...no, you've never heard heavy artillery fire then? It won't be loud, more like a disant thunder but i know for a fact it can be heard in 100-150mile range. No idea how altitude, weather, geography and other factors can extend this range but I wouldn't discount it without knowing more facts.
Shell detonations could definitely be heard at that distance through propagation through the ground/water. Even at sea, I don’t think you’d be able to hear the muzzle blast at anywhere close to 100 miles.
Ceremonial artillery doesn't fire projectiles, just propellant charges, so it seems implausible that it could be heard at those distances.
No kidding! I was on the interstate this last week and a sign showed more than twice that distance just to the next major city in the state. (along that highway, there were closer ones in other directions)
Seems plausible base on my experiences outdoors, but it heavily depends on ambient conditions. Over such large distances I think the unlikely part is you do that trick twice, but even that is plausible merely by looking at seasonal changes.
Some areas are famous for acoustic shadows or anomalies. There’s a place in Lake Tahoe called Granlibakken that used to have one of these during the summer, with the sound of what was occurring at the base bouncing off of the mountain and being blocked in some areas.
This is a not widely documented Civil War phenomenon, but happens probably quite often without people knowing it since by definition, it is the ABSENCE of sound that has to be noticed.
If your network is good enough to gather a microphone's data from each soldier/drone on the battlefield, you could centrally assemble a very accurate location and type on any weapon fired and other ambient sound. That data could be used to evaluate risks and return fire accurately within milliseconds. It could give a general a lack of fog-of-war only available before in simulations.
Possibly with small calibre weapons in urban areas.
Sound ranging is a solved and reliable way of identifying artillery locations and was the gold standard until counter-battery radar was developed during WW2. It's still in service with all major militaries (in the West, largely using BAE HALO) as it's a completely passive sensor system and therefore not subject to attack from HARM.
I think shotspotter's problems are more political than technical.
Not to say they don't have technical challenges too: better sensors than the human ear turn out to be fairly easy, but interpreting that data is still hard. Brains are complicated, "hearing" is a process contributed to by nearly every layer of meat that reacts to sound. The human ability to notice a new noise in the cacophony of everyday life is really an amazing trick if you think about it.
I know a non-shotspotter effort to identify gunshot sound in a very different context which also is having a very hard time; there is a real unsolved technical problem here. And it makes me think that ID'ing gun model from audio - which will be harder than simply identifying the existence of gunfire - is going to be very difficult.
Yes, but with caveats. Disclaimer: just a hobbyist, not my professional area, corrections appreciated
- oldschool radio direction-finding with a couple of directional antennas that are manually positioned and rotated (i.e. based on signal attenuation) doesn't really work for point-in-time things like a single brief signal unless you have massive arrays of well-positioned receivers, which is fine if you're building an HFDF installation but not fine if you're a squad on the move
- multilateration on the basis of signal attenuation, analogous to that sort of oldschool RDF, is possible but IME doesn't perform very well; same problem where you often need a lot of well-positioned points to shrink the possibility envelope down to something useful
- multilateration on the basis of time of arrival, analogous to more modern pseudo-doppler / correlative interferometry RDF techniques, can perform a lot better but requires a very high quality shared time source - if the model is analyzing info from an array of observers, this can be a problem on typical battlefield network designs that prioritize availability over performance - but this approach also enables single site setups where a local time source and a handful of receivers can read a bearing without actually needing to network at all, though this involves a lot of math
I read an ieee journal article over a decade ago claiming success with helmet mounted sensors. The system used the time of the sonic crack and the muzzle report as observed from each helmet. Its like GPS, if you have 5 or more observers you can solve for the time reference as well. Actually, it might be less since they had multiple observations from each helmet. It supposedly could then highlight the location of the shot using AR. It then went on to say they could tell what the firearm was from the muzzle report. Can't find the article now. Doesn't seem that outlandish, but as you say, accuracy would be the real question. The speed of sound can vary so much based on weather.
> It then went on to say they could tell what the firearm was from the muzzle report.
> over 95% caliber estimation accuracy for all shots, and close to 100% weapon estimation accuracy for 4 out of 6 guns tested.
That's pretty cool. Gotta read the whole thing now.
e: kinda relevant to the other subthread about time sync:
> Correlating ToA measurements requires a common time base and precise time synchronization in the sensor network. The Routing Integrated Time Synchronization (RITS) [15] protocol relies on very accurate MAC-layer time-stamping to embed the cumulative delay that a data message accrued since the time of the detection in the message itself. That is, at every node it measures the time the message spent there and adds this to the number in the time delay slot of the message, right before it leaves the current node. Every receiving node can subtract the delay from its current time to obtain the detection time in its local time reference. The service provides very accurate time conversion (few μs per hop error), which is more than adequate for this application. Note, that the motes also need to convert the sensorboard time stamps to mote time as it is described earlier.
Yeah I'm pretty sure that was it. They were getting really good classification results with simple statistical methods. Today I suspect someone would have thrown a hybrid CNN-LSTM at it and gotten similar results but no explainability and 100x the processor requirements. "Try the simplest thing that might possibly work".
A few μs for all that work seems high when gps could get you to ns, or adding extra observers could remove the need for prior synchronization. I guess it is a more robust answer at least, in case GPS isn't there and some of the observers got shot already.
GPS solves the sync issue to the order of nanoseconds. You can also have local TOA determine angles that get passed up to triangulate. The US Army deployed some of these techniques in antisniper work in Iraq.
That's a good point, and required time precision seems like it should be proportional to the velocity of the signal being investigated. I had some bad experiences trying to use gps time for RDF, but that might not apply here.
> Our time synchronization approach yields errors signif-
icantly less than 100 microseconds. As the sound travels
about 3 cm in that time, time synchronization errors have a
negligible effect on the system.
Sounds like it?
The anti-sniper systems I knew about already (Boomerang) are that single-site setup where time sync isn't a problem. The "Individual Gunshot Detector" sounds more like what you're describing.
> The Individual Gunshot Detector (IGD) by Qinetiq consists of a shoulder-mounted unit with four acoustic sensors and a chest display that attaches to body armor.
Still looking for more detail, but that sounds like a local ToA > pass up angles for triangulation type deal.
So there are products on the market that TDoA/Interfereometric emitter localization via ad hoc soldier portable units. ELTA has one product like this for example. They typically sync a decent local oscillator to GPS/GLONASS, and send that data stream along with any sampled signal that the network is communicating to do the computation.
Given that sounds occlude each other as overlapping pressure waves in air, this seems impossible. It's like unblurring a photo - information is lost and can't be recovered.
Similar historic outcome. I was in the room for Howard‘s Dean “scream“. which basically ended his campaign. As far as I can tell the cause was not enough attention paid to the microphone sound reinforcement. He had to push too hard to be heard in the raucous room.
I'd like to find the history of declaring that scream ended his campaign. Since you were there, did anyone in the crowd react in that manner afterwards, or was it really just an excuse the media used to lambast him artificially (which is what it felt like to me)?
I followed the campaign through the media at the time.
His campaign had already started going downhill at the time. The context of the rally where that occurred was that he'd finished 3rd in the Iowa Caucus, which was a bad result.
The scream was something that went viral and made him appear pretty odd which didn't help.
When it happened, by the media coverage, you would have thought he dropped his pants in the middle of a preschool or something. There was this almost uniform narrative that this 'scream' was crazy and it baffled me. These people all know how microphones and live events work, and it really felt like a concerted effort to drag the guy down - particularly given how gaffe-prone Bush was (putting it lightly), and especially given the whole Trump thing. 24-year-old me felt like it was a media drive-by, and I don't subscribe to the "big bad media" narrative that a lot of folks do. (I worked for a newspaper for a while, local journalism is especially important, which is why it's been targeted and gutted, but I digress)
Later - I can't remember where - I saw a series of shorts that broke down these pivotal moments. Dukakis in the tank, Dan Quayle and potato, the Dean Scream, and others. It painted a much more nuanced picture around complications within the organization of the campaign that were playing behind the scenes, and with or without the scream, Kerry was playing a much stronger game, apparently. I feel like Howard Dean might have even contributed to the documentary. It was really interesting, I wish I could remember the context of these shorts. "Did anyone in the crowd react in that manner" - I remember accounts from the time saying that in the room, it wasn't weird at all. Everyone was pumped.
None of that changes my mind that media supersaturated the scream. Late night shows, NPR, cable news, and I'm sure conservative media in particular were going to town.
I am a music recording engineer by hobby, and for me it was obvious that some threshold had broken. I was embarrassed for the sound team. (They had flubbed a Joan Jett appearance the day before.) I did not expect it to play so bad nationally.
A big piece of context missing from the media is that before Dean came on stage, the room was full of the loudest hooting and hollering group of Texans I’ve ever been with. We had all bussed up from Austin to volunteer. So he probably made a tactical error to try to go over the top of the crowd’s energy instead of coming in calming.
It was the broadcast mixing. IIRC his mic was played over the house sound as expected and it sounded fine for the audience. For the broadcast the engineer only used his mic and didnt mix in any of the audience background noise. Thats why it was a normal event for people in the room, but sounded like a lunatic turned up to 11 with no audience to briadcast viewers.
A while back I actually heard the different inputs, and the mix between what was broadcast and what normally would be.
You can hear the crowd as well (maybe through his mic?).
The thing watching this live on TV wasn’t the scream, but the gyrating hand movements and the gradual decent from a politician into a WWE wrestler voice as he’s naming a long list of states.
The scream was actually more of a yippee kinda sound.
But the disassociating thing for me was seeing him land a disappointing finish (3rd in Iowa) and then act like he’s a rock star about to headline a huge nationwide tour.
Sitting at home, it’s like yeah, why don’t you try to win one first before you go bragging about how you’re going to win the next 10.
So, for me it was the content + tone, not just the tone.
I would be very interested to see how this analysis controls for human factors: plenty of generals in the US civil war were, with the evidence of historical analysis, incompetent (think: Buell) or borderline sympathetic to the enemy (McClellan)
Might “I’d’ve attacked had I heard them!” be a convenient after the fact rationalization?
There's a boatload of other staffers nearby, to say nothing of individual soldiers who might have their own account of what they heard. It seems unlikely that a general would really get away with, "oh I didn't hear it" when everyone else there clearly did. I mean, they might escape court martial, but everyone there will know they're full of it.
That is really interesting. Honestly, I had never even thought about it until now, but now it reminded me of how Titanic may have sunk due to fatamorgana.
Weather really does play major role in human history, even when you don't see or expect it.
Wait how is this possible?