Hacker News Reader

Comment by tyingq | Hacker News Reader

Also mentions the "Shot Spotter" technology, which has some controversy associated with it:

https://www.aclu.org/news/privacy-technology/four-problems-w...

A UK research company developed this kind of microphone array tech in the 90s. It first came to fame in the Bosnian war and was used by UK peacekeeper troops to enforce ceasefires. In that case it was most useful for detecting artillery fire and could instantly say what kind of weapon was fired, where it was, and when it fired. This could be used to instantly work out who broke the ceasefire, and peacekeepers could even fire back to destroy the rogue artillery piece and restore order.

I'd really like to know how this technology differentiates a gunshot from a car backfiring.

How often do cars backfire? The internet is telling me that fuel injection and computer controlled mixtures have made it rare in modern vehicles. Carburetors were largely gone from new cars by around 1990.

Construction sounds might be a more common issue. Lots of nail guns use cartridge loads, for example. Or fireworks. Or impact hammers like you would use in a muffler shop, street jackhammers, etc.

It's relatively uncommon nowadays compared to the number of cars that don't backfire. But I certainly hear cars backfiring more frequently than I hear gunshots.

And then some hobbyist decides that backfiring is really cool, and tunes their vehicle to get that effect.

It’s almost a daily occurrence around where I live in the San Francisco Outer Sunset suburbs.

I went to college in Flint. Hearing gunshots were not uncommon. You can tell the difference from a car backfire. At it's core the "chamber" size is different. This creates a different reverb pattern, sound pattern, and frequency.

I'm guessing you could imagine the difference between slapping a surface of water vs concrete. The over noise will be dominated by the sound of your hand, but you can hear specifics for each.

----

Whether or not these systems can tell the difference, I'm not sure.

i hear gunshots every.single.night ( live in Oak Cliff which is on the South side of downtown Dallas ). It's so annoying and then maybe 10% of the time you get the helicopter about 5min later. You can tell gunshots from car backfires because there's 4 or 5 of them in row. pop-pop-pop-pop-pop vs one low-rolling bang

Edit: usually what this is is people leaving a bar/club and sticking their pistol out of their window on the way out of the parking lot. On a nice day with the windows open you'll hear the same when the Cowboys score a touchdown. basically impossible to catch/stop

It likely doesn't. In the article it mentions that it sends the Chicago PD on 60+ wild goose chases every day.

From my understanding, looking into making an open source version years ago, a human will actually listen to the audio clip and verify, plus there is some triangulation going on where they can determine the speed of the bullet and direction for at least high powered bullets using multiple microphones.

I’d guess a backfire (or afterfire) is significantly subsonic and therefore probably has a different (less sharp) profile than gunshot report of typical ammunition.

.45 ACP, which is probably the second most common pistol round, is subsonic by default.

It's definitely not the second-most common pistol round, especially in crime. It's way below .22LR and 9mm. Might even be below .380 ACP, .38 Special, and .25 ACP too. The guns used in crime are quite different than the guns preferred by the enthusiast community -- they're generally older, smaller, and cheaper.

.380 ACP, .38 special, and .25 ACP are also subsonic in their most commonly available rounds. As is .22LR as long as it is being shot out of a pistol (longer barrels will accelerate it past supersonic velocities, but almost all ammo is subsonic out of a pistol like a mark IV).

According to this survey on guns used in crime, it is 9mm, followed by .40 S&W (a typically transonic round), and then .45ACP.

https://oag.ca.gov/sites/all/files/agweb/pdfs/publications/f...

It's easy in a lab, in a city with potential for fairly arbitrary reverberations hitting the microphone at fairly arbitrary timings it's right up there with forensic fire analysis and other crap like that.

I wish that article focused solely on whether or not the technology actually works, rather than leaning heavily on the “problematic” aspect that forces people to maybe face facts that run counter to their dishonest or at least misleading narratives.

It's the ACLU, so their focus seems unsurprising. The company was already caught falsifying data, so it's not made up concerns.

https://www.vice.com/en/article/qj8xbq/police-are-telling-sh...

That article is mentioned in the ACLU piece, as Shot Spotter is trying a $300 million defamation suit over it. They claim they always do a secondary "detailed forensic report" and their communication with police about that is inconsequential.

It’s reasonable to define some high level goals when determining whether a technology like this “works.”

Reply on news.ycombinator.com