×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Engineers Use Electrical Hum To Fight Crime

Unknown Lamer posted about a year ago | from the nowhere-to-hide dept.

United Kingdom 167

Hugh Pickens writes writes "A suspected terrorist has been taped planning a deadly attack and the police want to use this evidence in court, or someone has been captured on CCTV threatening an assault. Increasingly, recordings like these are playing a role in criminal investigations, but how can the police be sure that the audio evidence is genuine and has not been cleverly edited? Now Rebecca Morelle writes on BBC that a technique known as Electric Network Frequency (ENF) analysis is helping forensic scientists separate genuine, unedited recordings from those that have been tampered with and the technique has already been used in court. Any digital recording made near an electrical power source will pick up noise from the grid that will be embedded throughout the audio. This buzz is an annoyance for sound engineers trying to make the highest quality recordings, but for forensic experts, it has turned out to be an invaluable tool in the fight against crime. Due to unbalances in production and consumption of electrical energy, the ENF is known to fluctuate slightly over time rather than being stuck to its exact set point so if you look at the frequency over time, you can see minute fluctuations and the pattern of these random changes in frequency is unique over time providing a digital watermark on every recording. Forensic Scientist Philip Harrison has been logging the hum on the national grid in the UK for several years. 'Even if [the hum] is picked up at a very low level that you cannot hear, we can extract this information,' says Dr. Harrison. 'If we have we can extract [the hum] and compare it with the database, if it is a continuous recording, it will all match up nicely.'"

cancel ×
This is a preview of your comment

No Comment Title Entered

Anonymous Coward 1 minute ago

No Comment Entered

167 comments

Thanks! (2, Funny)

Anonymous Coward | about a year ago | (#42259893)

Thanks forensic scientists for telling us how to fix this!

Yours,

The Villains

Re:Thanks! (1)

Anonymous Coward | about a year ago | (#42259937)

The challenges of re-generating electrical hum are obscene, even if you have a recorded source to match up you'll have to purge all the original and bring your own in perfectly, without leaving any evidence of tampering with the sound.

Re:Thanks! (0)

Anonymous Coward | about a year ago | (#42260385)

You don't have to regenerate it. If you have an original you have the original hum. You remove hum from the stuff you're patching in. Now, if only we had some idea of what frequency the electric hum was at ...

Re:Thanks! (1)

Anonymous Coward | about a year ago | (#42260669)

Now, if only we had some idea of what frequency the electric hum was at ...

The point is that it's not 60Hz. It's 59.991234Hz or 60.006789Hz. It's almost never 60.000000Hz. If you can make even the slightest guess as to where the recording was taken, you just look up the utility company's records as to how things were shifting over time, and cross-reference. There's probably only one time in all of history in which the frequency shifted from precisely 60.004567Hz to 60.006789Hz in the space of four minutes with a 30-second pause at precisely 60.005959Hz, and the utility knows which four minutes those were.

Very clever idea, and congratulations to the authors.

Re:Thanks! (4, Funny)

Hognoxious (631665) | about a year ago | (#42261435)

The point is that it's not 60Hz.

No, it's not 50 Hz.

It's because the UK has metric electricity, or something.

Re:Thanks! (1)

TheLink (130905) | about a year ago | (#42260989)

You don't have to power your stuff from the mains while you create your recording. You'll have your own unique hum, but they won't be able to match that with anything - especially if you get rid of the stuff after that.

You could also create your recording purely digitally- no microphones - then there would be no hum except what you choose to artificially add. For example you could record just the hum at a completely different period to throw them off, then to that you add a vocaloid or other voice synthesis stuff. If you're going to do bad stuff you might as well not use your real voice in the recording.

The issue with video is- video stuff might record the hum as part of the video signal (noise in the video). So if you're going to show your kidnapped captive in video it might be tricky to get rid of the hum.

Re:Thanks! (1)

TheLink (130905) | about a year ago | (#42261009)

it might be tricky to get rid of the hum.

Unless as I said you cut off the mains and power everything off something else.

I suspect this tech might also be used to fake evidence against people they "know" are guilty.

Police (1)

noh8rz10 (2716597) | about a year ago | (#42259947)

Ftfs: "how can the police be sure the tape hasn't been edited?"
more like how can the defense lawyers know the police rent railroading them! Or how can the police successfully fake an edit! (or to be fair,m how can the police prove a tape is genuine when the defense lawyer throws up somew fud).

Re:Police (4, Interesting)

Anonymous Coward | about a year ago | (#42260025)

more like how can the defense lawyers know the police rent railroading them!

Maybe it's just my bias from having worked with a lot of forensic analysts, but usually it seems that it's the prosecutors trying to railroad the police... not into fabricating evidence, per se, but "looking real hard" for "anything they can come up with". The police don't much like it, unless it's a "known bad guy" they're trying to pin *something* on, because searching for trivial things to make the DA look good wastes time that they don't have (not with one-year case backlogs).

or to be fair,m how can the police prove a tape is genuine when the defense lawyer throws up somew fud

This is the most common problem, but "throwing up some FUD" is called "providing a rigorous defense". After all, if you don't periodically make the police prove they didn't fabricate evidence, then there's less incentive to not fabricate evidence.

Re:Police (2)

gr8_phk (621180) | about a year ago | (#42260553)

usually it seems that it's the prosecutors trying to railroad the police... not into fabricating evidence, per se, but "looking real hard" for "anything they can come up with".

I always thought that's what happened with George Bush and the search for WMDs in Iraq. The intelligence guys tried really really hard to provide evidence for what he wanted them to find. After the invasion, all the evidence was deemed to be crap (or "faulty" or "less reliable than thought").

Re:Police (1)

RoccamOccam (953524) | about a year ago | (#42260607)

If they were just trying to do this to satisfy George Bush, why were the same conclusions drawn during the Clinton administration?

Re:Police (2)

Sulphur (1548251) | about a year ago | (#42260921)

After all, if you don't periodically make the police prove they didn't fabricate evidence, then there's less incentive to not fabricate evidence.

Use of excessive force against that sentence has been approved.

Re:Police (2)

fuzzyfuzzyfungus (1223518) | about a year ago | (#42260149)

Ftfs: "how can the police be sure the tape hasn't been edited?"
  more like how can the defense lawyers know the police rent railroading them! Or how can the police successfully fake an edit! (or to be fair,m how can the police prove a tape is genuine when the defense lawyer throws up somew fud).

The only real defense against accusations of railroading is context. If this technique works, it should at least be possible to demonstrate that chunks of context haven't been removed(ie. the 'ask question', 'stop tape', '6 D-cell maglite', 'start tape', 'tearful confession' use case); but it obviously doesn't magically add recordings that were never taken, that's more of a procedural problem.

Re:Thanks! (-1, Troll)

Anonymous Coward | about a year ago | (#42259971)

Thanks forensic scientists for telling us how to fix this!

Yours,

The Villains

i wouldnt worry about that.

this is for fighting crime. and where there's crime there's niggers.

niggers don't engineer background noise to edit audio recordings. dat requires mathematics and junks yo, and dat be actin white.

no it is not because of their skin color. but american blacks are the most anti-achievement culture to be found anywhere. failing at life and being a menace is some kind of noble goal to them. cant be skin color because the majority of thug gangsta blacks will harass and threaten and beat up the few who want to get ahead. so sorry if it's not PC to notice.

Re:Thanks! (1, Informative)

Jeremiah Cornelius (137) | about a year ago | (#42260309)

"suspected terrorist has been taped planning a deadly attack "

When? Where? This has never happened. There are statistically, NO terrorists. The FBI is so desperate to locate any, that they manipulate informants and patsies into hopeless and ineffectual plotting. These are "terrorists" who'd otherwise be confined to grumbling on the short-bus!

I think this whole topic is one generated by the anti-terrorism industry. Resources would be better spent on pool covers, if you want to protect against the unexpected loss of innocent lives.

Re:Thanks! (0)

Anonymous Coward | about a year ago | (#42260525)

If you can be "manipulated" into trying to set off a bomb in public, you need to fucking be locked up.

Re:Thanks! (0)

Anonymous Coward | about a year ago | (#42261243)

There's a big difference between "I am pissed off, but have no idea where to even get a bomb, let alone the follow-through to actually set one off", and "I'm able to build and set a bomb I just need someone to piss me off a little more."

Law enforcement has a tenancy to find the first kind of person, give them fake bombs, plan the attack for them, and then arrest them when they try to use their government supplied 'bomb' to execute a government planned 'attack'", while pretending they walked up to someone of the second kind and said "hey wanna blow stuff up?" then made the arrest when he said "yes".

Re:Thanks! (1)

fustakrakich (1673220) | about a year ago | (#42260809)

I whipped off her bloomers 'n stiffened my thumb
An applied rotation on her sugar plum

I poked 'n stroked till my wrist got numb
But I still didn't hear no dinah-moe humm...

Re:Thanks! (2)

neyla (2455118) | about a year ago | (#42261021)

People worry about the spectacular but unlikely rather than the common but nonremarkable. They worry about events that kill many people at once, in one location, especially if it involves explosions or words like "bioweapon" "chemical" or "radiation". (they aren't aware that water is a chemical or that light is "radiation")

Sitting on your coach. Eating too much sugar. Getting too little exercice. Smoking. These things are dangerous.

Terrorists, on the other hand, are essentially noise - unless you're living in something close to a war-zone, but even in Afghanistan or places like that, terrorism likely isn't among the top cauces of death.

Re:Thanks! (1)

mcgrew (92797) | about a year ago | (#42261187)

When? Where?

New York in October.

Federal Reserve Bank Plot [heritage.org]

According to reports by the FBI, 21-year-old Bangladeshi citizen Quazi Mohammad Rezwanul Ahsan Nafis came to the U.S. in January 2012 on a student visa with the explicit goal of carrying out a terrorist attack on U.S. soil. Upon arriving in the U.S., Nafis actively sought out al-Qaeda contacts within the U.S. to assist him in carrying out an attack. Unbeknownst to Nafis, one of the individuals he sought to recruit for his terrorist cell turned out to be an FBI informant.

The FBI is so desperate to locate any, that they manipulate informants and patsies into hopeless and ineffectual plotting.

Yes, almost all of them. But this guy came here with the intention of attacking.

However, I agree with you -- Resources would be better spent on other things, like guard rails on a lot of highways. Ten times as many people die on the highways here every single year than died on 9-11.

doesn't work in most cases (4, Interesting)

slashmydots (2189826) | about a year ago | (#42259905)

Most respectable recordings eliminate these hums by adding a realtime filter called a "noise gate" that eliminates all sounds of any frequency under XX decibels (usually 20-40). That and most eliminate all noises under XX Hertz (usually also 40) which gets rid of some hums without ruining human voices.

Re:doesn't work in most cases (2)

lxs (131946) | about a year ago | (#42260085)

A noise gate only shuts off the sound when the total input level falls below a certain threshold. The noise is still there in the parts that aren't quiet but you can't hear it behind the intentional sound.

Re:doesn't work in most cases (1)

slashmydots (2189826) | about a year ago | (#42261279)

Oh yeah, I totally knew that and forgot. That's why I hate noise gate filters and set them at like 20dB because they always tend to cut off the end of what anyone says and their voice fades out and makes the recording sound noisy and low quality when compared repeatedly to the perfectly blank quiet areas.

Re:doesn't work in most cases (1)

westlake (615356) | about a year ago | (#42260231)

Most respectable recordings eliminate these hums by adding a realtime filter called a "noise gate" that eliminates all sounds of any frequency under XX decibels

Inaudible to the human ear isn't the same thing as undetectable.

A security system may be designed to capture and record all potentially relevant sounds. Filters can be applied in playback.

Re:doesn't work in most cases (4, Informative)

Anonymous Coward | about a year ago | (#42260233)

Most (technically) respectable recordings transfer sound in the range 20Hz ~ 20kHz. Most shoddy recordings are not made using devices that employ digital FFT to get rid of low frequencies entirely. And, filtered analogue signal can still be susceptible to acquiring the hum again. From the space around, from the power supply, from your hand holding the device.

By the way, detecting the utility frequency is a standard technique used in Poland (and, I thus guess, in many other coutries) for years. And it works. Forging the frequency in software is possible (Audacity is mostly enough), if you are a profi that has access to respective databases (or has logged the uf by himself).

Re:doesn't work in most cases (1)

Anonymous Coward | about a year ago | (#42260505)

High pass filtering is used to preserve headroom, however, any signal content removed will have overtones extending throughout the remainder of the frequency range.

Noise gates are only used in difficult recording situations... or as effects. The ENF would be available for analysis on any part of the recording where the expander was not fully closed.

A little knowledge... makes people appear foolish!

Re:doesn't work in most cases (1)

dintech (998802) | about a year ago | (#42260735)

The noise gate would only kick in with the sound levels in the recording drop below that 20-40db threshold. You could still overlay the historical 50Hz hum over the rest of the recording where the gate was open. These should still match up. Where it might would not match would be if is 50Hz + harmonics notch filters were applied to the recording specifically to reduce mains hum. This is possible too.

Re:doesn't work in most cases (1)

deimtee (762122) | about a year ago | (#42261385)

It would still be detectable even with a notch filter.
Back when it was still worthwhile to do, I built several stereo amplifiers, either from kits or designs in magazines. Getting rid of the 50hz hum was always the hardest part.
Earth loops, wires passing close to the power, you name it, it would make the buggers hum.
50 hz notch filters were common in the cheaper designs, but they would only reduce it. You might gain 30 or 40 decibels with a good one, but modern signal processing will still pull the hum out of that no problem.

Re:doesn't work in most cases (0)

Anonymous Coward | about a year ago | (#42260741)

The electrical grid in the UK runs at 50 Hz, 60 in the US. So I'm not sure how that would matter.

Workaround (4, Interesting)

tibit (1762298) | about a year ago | (#42259915)

So, if I apply a tracking notch to the doctored recordings to remove the original hum and its harmonics, and then superimpose data from the database, I'm all set? Well, thank you!

Re:Workaround (5, Interesting)

serviscope_minor (664417) | about a year ago | (#42259993)

That was my first thought.

Though you don't even need access to the database. If you know in advance that you're going to need to doctor recordings, you could just record continuously with a modest amount of data, and filter out all but the hum using the inverse of the notch filter.

You can then so more or less the same process without needing a big database.

I strongly suspect that you wouldn't need much more than sox and octave (using fft/ifft) to do this.

If they start believing this, then you could set someone up by making it appear that a recording has happened at a particular time, when it has not.

Re:Workaround (0)

Anonymous Coward | about a year ago | (#42260291)

I'm certainly no pro but if you look at a recording in, say, Adobe Soundbooth you can easily spot the difference between a bit of audio that is organic and one that's been altered. The spectrogram looks decidedly different.

Re:Workaround (4, Insightful)

Maximum Prophet (716608) | about a year ago | (#42260117)

It's almost like "if there's a way to automatically determine X, there's an automatic way to fake X.

Wait, it *is* that way.

In the future there'll be no way to determine reality, because we'll all just be disembodied heads in vats, networked into vaste reality generating machines.

Re:Workaround (0)

Anonymous Coward | about a year ago | (#42260333)

In the future there'll be no way to determine reality, because we'll all just be disembodied heads in vats, networked into vaste reality generating machines.

Future? What makes you think you aren't already a brain in a jar?

Re:Workaround (1)

blueg3 (192743) | about a year ago | (#42261119)

It's almost as if there isn't a whole field of cryptography dedicated to signals that are trivial to verify and nearly impossible to fake.

Wait, there is.

Re:Workaround (0)

Anonymous Coward | about a year ago | (#42260301)

You could just run it through audio enhancement to make a gooey interface in visual basic to track the killer's IP.

Re:Workaround (0)

Anonymous Coward | about a year ago | (#42260795)

I find that just removing the audio track entirely and adding my own subtitles works better.

Re:Workaround (2)

VortexCortex (1117377) | about a year ago | (#42261011)

So, if I apply a tracking notch to the doctored recordings to remove the original hum and its harmonics, and then superimpose data from the database, I'm all set? Well, thank you!

That was my fist thought too, but my 2nd thought was: So, if I use a hum generator to subtly but abruptly switch between hums then anything they record me saying can be proven fake in court? It's a poor substitute for privacy, but I'll take what I can get; Thanks!

You really are a bunch of dumb fucktards (-1)

Anonymous Coward | about a year ago | (#42259919)

'Unknown Lamer', 'Hugh Pickens writes writes', 'Due to unbalances in production'.

Good grief, what are you people twelve? You really are a bunch of dumb fucktards aren't you. Figures this place is lousy with socialists doesn't it?

Re:You really are a bunch of dumb fucktards (0)

Anonymous Coward | about a year ago | (#42260023)

ITT: An AC bitches about peoples names.

Re:You really are a bunch of dumb fucktards (-1)

Anonymous Coward | about a year ago | (#42260111)

Good grief, what are you people twelve?

No, Leftists are only children emotionally. They think if you just wish hard enough suddenly failed ideas will start working. Eventually they run out of other peoples' money and when they do, they don't grasp the principle. They didn't listen to people who knew it would fail and knew how it would fail. Those people are The Enemy(tm) and must be insulted with as much venom as possible. Leftists didn't listen, they don't listen, and they aren't going to start listening because that would mean not wishing hard enough. Instead they say oh well let's try the same thing, again, but now we'll try it even harder! It'll work this time really!!

Federal government involvement and micromanagement of daily life is NOT how the US got to be such a powerhouse in the world. Going down that path is how the US started to decline. Isn't that simple? Like I said the big-government Leftists are emotional children. They want to go down that path some more because it'll really work this time golly honest injun it will! Hey this thing is failing so let's do it more, yeah, what great logic, oh wait they don't use logic they use their feelings. That's a good way to do art, and a horrible way to run a nation.

But somehow this is considered controversial. Lots of idiots who don't understand it gives the appearance that there is a legitimate debate. We've seen that before.

Re:You really are a bunch of dumb fucktards (0, Troll)

Anonymous Coward | about a year ago | (#42260447)

Yes and no. The real truth is that socialism is for the people, *not* the socialists.

In the end they just want to steal your shit, it's that simple. And yes, childlike.

Oh and for the dumbass who wrote "ITT: An AC bitches about peoples names."

I have an account here but whenever I use it I get modded down to the point I am only allowed to post like once per week, and believe me this is not due to me using abusive language or insults. You fools shutdown any debate that does not qualify as party approved, another sign of the true believers, you cannot succeed in a real debate so you shut the opposition down.

Oh and I wasn't really bitching about names, I was pointing out the absurdly poor grammar, please try and keep up with the adults.

Hence I just post AC. So go fuck yourself.

Re:You really are a bunch of dumb fucktards (0)

Anonymous Coward | about a year ago | (#42260531)

Wow, typical moron.

says doesn't insult, provides insults in next 4 sentences.

And tries to say it's someone else's fault.

I'm guessing your actual arguments aren't any better.

Re:You really are a bunch of dumb fucktards (0)

Anonymous Coward | about a year ago | (#42260873)

You clearly are not paying attention. I will use polite arguments but when I do so here I am shut down.

It is you who cannot handle rational debate. I can prove this. Go ahead and make your argument for forced unionization of the labor force or whatever your leftist pet cause is. Use logic and reason and I will show you how you are wrong.

Go ahead, make my day.

Re:You really are a bunch of dumb fucktards (0)

Anonymous Coward | about a year ago | (#42260697)

they aren't going to start listening because that would mean not wishing hard enough.

As opposed to Republicans who don't listen because that would mean not faithing hard enough.

Federal government involvement and micromanagement of daily life is NOT how the US got to be such a powerhouse in the world.

Are you telling me that if the federal government didn't regulate backyard pot growers as interstate trade, we wouldn't turn into a nation of pot zombie slackers with reefer madness? I thought the war on drugs is what made us all upright and productive citizens.

it'll really work this time golly honest injun it will! Hey this thing is failing so let's do it more, yeah, what great logic

And any second now, Fidel Castro will see the light and convert to capitalism. Aaaaaannnnyy second now...

Wake me up when we have an actual small government conservative choice that has a chance of winning an election. Until then I'll keep voting Libertarian and blaming everyone else for how the country is going to hell in a handbasket.

Re:You really are a bunch of dumb fucktards (0)

Anonymous Coward | about a year ago | (#42261219)

I am not the AC however I would observe, speaking against leftists does not equal supporting Republicans.

In many ways the Republicans weakness and failure to protect individual liberties and the constitution make them more guilty that the Democrats in driving this nation 'to hell in a handbasket'.

The unfortunate truth is that the Democrats and Republicans are actively colluding to work against the citizens liberties. Can I prove this? No. But I believe it.

Don't see it myself (4, Interesting)

ledow (319597) | about a year ago | (#42259967)

Granted, to catch bad fakes and casual deception, this probably works

But what's stopping someone now knowing this, and supplanting their own "hum" back into the right place from any recording make anywhere during the time they want to "pretend" to be recording in?

What's actually stopping them ripping out the hum entirely and replacing it with the hum of any arbitrary period of time? If it's in every recording, and you can spot it in, say, CCTV recordings (that use quite primitive recording equipment and compression methods), then it's also incredibly easy to detect and "fake" yourself.

Also, I don't believe it's as reliable as they make out. We'll find out as more cases use it and it will have to be challenged at some point but even speed-cameras that weren't entirely accurate got a lot of people in trouble and then had to be rescinded years later.

And, totally off-topic, it reminds me of my German teacher in school (I'm English, and we were taught German as an optional foreign language). He basically begged me to take his class instead of Double Science, many moons ago, and even called me a "rat" at a parent's evening in front of my dad when I said I wasn't going to study German any more. Turns out he was kinda hoping I'd take German so he could up the average result a bit.

Two years later, he was sacked because the audio recordings of oral work that he sent off for the final exam had "clicks" on the tape (yes, tape!) in between every question where he'd obviously paused it and briefed the children on the right answer.

Wonder how they submit oral coursework now, with MP3's and things? It would be the work of a second to get a perfectly smooth recording of the same thing happening nowadays.

Re:Don't see it myself (1)

fuzzyfuzzyfungus (1223518) | about a year ago | (#42260105)

Wonder how they submit oral coursework now, with MP3's and things? It would be the work of a second to get a perfectly smooth recording of the same thing happening nowadays.

At least with spoken-language AP exams, you are allowed to record digitally(if memory serves, the AP people actually tell you to use Audacity); but you are then forced to burn the result to CD, just to remind you that we don't actually live in the future yet...

Re:Don't see it myself (1)

Maximum Prophet (716608) | about a year ago | (#42260139)

We're finding out that many forensic "experts" are faking their expertise, and essentially lying on the witness stand.

No need to fake evidence, just find a friendly "expert" to tell the jury what you want them to hear.

Re:Don't see it myself (1)

dgatwood (11270) | about a year ago | (#42260569)

That was my thought as well. This is somewhat plausible for 24-bit recordings, but that isn't exactly something you see a lot of in the real world (outside of studios). Most recordings are 16-bit. The digital noise floor for 16-bit digital recordings is audible. If you have a 60 Hz hum that is loud enough to really be detected, that is also audible.

I suppose if the bottom bit is not dithered, you might—might—be able to apply some sort of statistical analysis to the data over time and arrive at a probability that the 60 Hz hum is likely not discontinuous, but being able to reproducibly detect individual waves with enough accuracy to use it as a time stamp sounds more like CSI physics than reality.

I could, however, believe that you might be able to decode WWV/WWVB from the signal, given a large enough set of samples, 192 kHz sampling rate, and no low-pass filtering. Of course, that isn't commonly seen even *in* recording studios....

Re:Don't see it myself (4, Informative)

queazocotal (915608) | about a year ago | (#42260999)

The noise floor is not that simple.
If you analyse the noise, you find it's broadly spread over a wide variety of frequencies.
If you analyse 10s or so slices of signal, over the range 47-53Hz, things get considerably easier.
10 seconds means that your effective number of signal levels is not 2^16 (65536)
It's sqrt(44100*10)*65535, about 25 bits.
Throwing out the out of band noise means you lose >99.5% or so of it.

As a practical measure.
Go to pretty much any CD you have, and do a FFT in the range 47-53 or 57-63hz.
In the vast majority of cases, you will find a signal.

And it's not quite a timestamp, unless the recording is quite long.
http://www.nationalgrid.com/uk/Electricity/Data/Realtime/Frequency/Freq60.htm [nationalgrid.com] - for example - this is the UKs last hours frequency.
The graph is clearly enough to show that even if you can 'only' resolve in 10 second intervals the frequency, it's quite plausible to say if a 10 minute video can be one specific timeframe, or not.

Re:Don't see it myself (1)

deimtee (762122) | about a year ago | (#42261511)

That's interesting. It's varying from 49.9 to 50.1
2% is a lot more variation than I thought they would have.

Re:Don't see it myself (0)

Anonymous Coward | about a year ago | (#42261071)

Isn't there a 20+ dB difference in human hearing response between 60 Hz and 1000 Hz on the quieter end? I thought my hearing was ok and I've seen 60 Hz and a few harmonics show up on spectrograms of crappy recordings where the hum at least doesn't seem audible, but other components are.

Re:Don't see it myself (1)

neyla (2455118) | about a year ago | (#42261227)

You don't need to detect individual waves. The thing is, electrical hum is supposed to be spot-on 50hz or 60hz (depending on your location), but in reality it drifts around slightly with 50Hz or 60Hz being the *average* frequency. You might have several seconds of 50.01 hz followed by a minute of 49.998 hz and so on.

In other words, if you plot frequency versus time you don't get a straight line at 50 hz, but a line that rise and fall over time while staying on the average very near the goal-frequency.

Re:Don't see it myself (0)

Anonymous Coward | about a year ago | (#42260227)

Most people committing these crimes aren't going to stop and worry, "I wonder if there is an audio recording device near by that happens to be able to pick up the hum of the electrical grid that will later be used against me in a court of law?" Honestly, with how dumb some criminals are, it's a miracle they don't more often leave their cell phone and/or ID at the scene of the crime more often. I imagine this would only rear its technical head in very limited circumstances, before being kicked from the courtroom altogether when it turns out some kind of process in common recording devices interferes with it's 100% accuracy (noise cancellation,for example, converters on the circuits, etc).

Re:Don't see it myself (2)

westlake (615356) | about a year ago | (#42260317)

I imagine this would only rear its technical head in very limited circumstances, before being kicked from the courtroom altogether when it turns out some kind of process in common recording devices interferes with it's 100% accuracy (noise cancellation,for example, converters on the circuits, etc).

Forensic evidence is admissible even when you cannot guarantee 100% accuracy.

The goal line for conviction is set at " beyond a reasonable doubt" not a distant - impossible - metaphysical - certainty,

Re:Don't see it myself (2)

coldsalmon (946941) | about a year ago | (#42260619)

What's actually stopping them ripping out the hum entirely and replacing it with the hum of any arbitrary period of time?

It's not quite so easy to remove noise of this nature, especially if it exists across a broad frequency range. Professional recording engineers have been trying to remove noise from old recordings for decades, and it always has a noticeable effect on the music. I am no sound engineer, but I suspect that this is for two reasons: 1) the presence of the noise destroys some of the musical content, so removing the noise will leave a sonic hole in the recording which cannot be filled because the original musical content has been lost, and 2) our methods are simply not sophisticated enough to remove noise without also removing some music at the same time. I am using "noise" and "music" in this example, but this applies equally well to "hum" and "non-hum" in these police recordings. It is likely that removing a hum would leave a "ghost" fingerprint on the recording, especially if the hum exists at the same frequencies as human speech. TFA states that the frequency is 50Hz which is quite low (less than one octave above the lowest note on a piano), but there may be overtones that extend into the range of the human voice. Just as a trained musical recording engineer can tell if noise reduction techniques have been used on a recording, so a trained forensic expert could most likely tell that noise reduction techniques have been used on a police recording.

Now, if you could make a recording without ANY noise whatsoever, it would be trivial to add falsified noise in later. It might also be possible to add a loud falsified noise that would completely cover up a softer authentic noise. However, it would probably be easy to establish the expected noise level based on the known location of the recording.

Re:Don't see it myself (1)

VortexCortex (1117377) | about a year ago | (#42261171)

What's actually stopping them ripping out the hum entirely and replacing it with the hum of any arbitrary period of time?

Nothing. In court the question will be, "what's the likelihood that it's been faked by sophisticated AV specialists?". It's just one point of data that will be used to determine the ultimate verdict. With enough evidence, say like some planted DNA with your fingerprint, etc. the Jury may put to bed any "reasonable doubt", and find guilty. Think of this next time you get up to leave the coffee house with a strand of your hair on the chair / floor, and toss that almost empty yet DNA laden Starbucks cup, complete with fingerprints, into the trash. Look around and see if there's someone about your height, build and skin / eye color. Perhaps they're planning to snag some evidence for planting at their next crime. They could follow you home, figure out your schedule, and time their crime such to ensure you have no alibi.

Once the detectives "like you" for a crime, they'll find more (circumstantial) evidence that you're guilty, and put less effort into tracking down the real criminal. "Oh, Come On! What are the chances of that happening?!" -- That's exactly what a smart criminal is counting on, i.e. that the jury assumes criminals are dumb.

Where does the 'hum' enter the recording? (1)

fuzzyfuzzyfungus (1223518) | about a year ago | (#42260083)

I can imagine that some mic designs pick up mains hum particularly well; but others(albeit probably the much less common ones, optical ones, say) should be essentially immune. The power supply also wouldn't be an issue on battery-powered devices, though I'm sure most AC adapters are pretty lousy at filtering.

Is the hum all picked up by the mic, in most commonly used designs, or are there other parts of a recording system that are sensitive to picking up noise?

Re:Where does the 'hum' enter the recording? (0)

Anonymous Coward | about a year ago | (#42260165)

I can imagine that some mic designs pick up mains hum particularly well

Yo mama's mic gave ME a hummer.

Re:Where does the 'hum' enter the recording? (2)

Bill Currie (487) | about a year ago | (#42260253)

It's the wires into the high-gain mic amp that pick up the hum.

Re:Where does the 'hum' enter the recording? (1)

fuzzyfuzzyfungus (1223518) | about a year ago | (#42261271)

Do you know if the signal path between mic and amp in the teeny new digital MEMS mics are long and unshielded enough for it to show up on devices using those?

Something like an ADMP441 is a 4.72 mm × 3.76 mm × 1 mm little can with pads, and both the mic and all the analog wrangling are handled inside, external interface is I2S. How much 'we-shielded-it-just-enough-that-the-FCC-doesn't-care' analog signal path would you need before the amplifier?

Re:Where does the 'hum' enter the recording? (0)

Anonymous Coward | about a year ago | (#42260995)

Dynamo hum?

Electrical Fingerprint (1)

Wolfraider (1065360) | about a year ago | (#42260225)

Form what it looks like to me, they are using the random increases and decreases to create a sort of "fingerprint" of the recording. This fingerprint could then be used to verify that the recording is genuine at any other time. It would be very hard to duplicate the random noise in a recording exactly if you tried to tamer with the recording.

Re:Electrical Fingerprint (1)

Maximum Prophet (716608) | about a year ago | (#42260311)

You don't have to tamper with the analog sound around the microphone, you just have to tamper with the digital recording enough to fool the algorithm the determines the recording is "correct". If you have access to the original recording, the algorithm and the database, you'd be able to do this.

Studies have shown that even innocent suspects will confess when presented with overwhelming (fake) evidence.

Re:Electrical Fingerprint (1)

K. S. Kyosuke (729550) | about a year ago | (#42260869)

Form what it looks like to me, they are using the random increases and decreases to create a sort of "fingerprint" of the recording.

Actually, the description makes it sound more like dendrochronology and tree rings.

not to mention.... (0)

Anonymous Coward | about a year ago | (#42260271)

...any editing would likely cause a big phase hit in the stray line frequency pickup, at least before this technique was made public. Really a brilliant idea.

Uniformity across the country (1)

jickerson (2714793) | about a year ago | (#42260273)

If the hum is generated from imbalances between the generation and consumption of power, is there any difference in the hum across the entire grid? Or do the transmission lines transmit the current evenly across the entire grid to the point where the hum in one area is indistinguishable from a spot clear across the country?

Re:Uniformity across the country (2, Interesting)

Anonymous Coward | about a year ago | (#42260423)

The short answer is: yes, at least under steady-state conditions. Large disturbances (losing a generator, for example) create electromechanical transients that propagate throughout the grid at relatively slow speeds (1000 miles/s).
My research group studies these things - you can learn more here: http://en.wikipedia.org/wiki/FNET
We also have some people working on the ENF authentication technique.

Re:Uniformity across the country (0)

Anonymous Coward | about a year ago | (#42260455)

Oops, I didn't phrase that well...what I was trying to say was that there really isn't much difference under steady-steady conditions.

And when the Romulan Senator dies... (0)

Anonymous Coward | about a year ago | (#42260283)

You see, when the Tal Shiar finishes examining the wreckage of Vreenak's shuttle, they'll find the burnt remnants of a Cardassian optolythic data rod which somehow miraculously survived the explosion. After painstaking forensic examination, they'll discover that the rod contains a recording of a high-level Dominion meeting, at which the invasion of Romulus was being planned. Any imperfections in the forgery will appear to be a result of the explosion. So - with a seemingly legitimate rod in one hand, and a dead senator in the other, I ask you, what conclusion would you draw?

On a completely unrelated note... (1)

cuncator (906265) | about a year ago | (#42260319)

Am I the only one who saw the "electrichum" tag and thought "electri-chum"? Think electrified chum would work pretty well for fishing. The scent would bring the fish in and the electricity zaps them/dolphins/penguins/scuba divers.

Re:On a completely unrelated note... (1)

Sulphur (1548251) | about a year ago | (#42261137)

Am I the only one who saw the "electrichum" tag and thought "electri-chum"? Think electrified chum would work pretty well for fishing. The scent would bring the fish in and the electricity zaps them/dolphins/penguins/scuba divers.

It also attracts electrified sharks.even if their lasers are off.

Upload them (1)

Hentes (2461350) | about a year ago | (#42260397)

Faking A/V gets easier and easier. The only way I see to be able to use CCTV records as proof is to upload them to a trusted organisation at real time. The feed should be encrypted, but everytime a court wants to use it as proof they can ask for the recording from the organisation and the key from the owner, and can be sure that the result has not been tampered with.

Re:Upload them (2)

KiloByte (825081) | about a year ago | (#42260581)

Or upload just hashes of segments of the recording, preferably signed and timestamped by a public server. The set of possible attacks is the same: a pre-recorded fake, a "suddenly broken" camera, and "oops, my disk broke and there are no backups".

Re:Upload them (1)

Hentes (2461350) | about a year ago | (#42261629)

You can't protect against the owners not recording themselves. After all, if they really don't want to appear on tape they won't put up cameras in the first place. Realistically, the footage can't be used against its owner.
Prerecorded fakes are a bigger problem, but they are also much harder to do for a number of reasons. First, they require lots of preparation, so they wouldn't work in cases where the crime wasn't, or wasn't properly planned. Second, there are a huge number of unpredictable events, which you can't fake. For example, if a camera is between two other cameras, both of which recorded a person or car passing by, its absence in the footage of the middle one will be a sign of forgery.

Example (1)

HeckRuler (1369601) | about a year ago | (#42260411)

A suspected terrorist has been taped planning a deadly attack...

Holy shit, that's amazing and extremely newsworthy, especially on Slashdot where the general consensus that the CCTV pantopicon society is a ineffectual waste of money and an over-reaching of power by asshole cops.

Oh... It's one of those hypothetical scenarios...
Yeah.....

Interesting seeing the Skepticism (2)

mykepredko (40154) | about a year ago | (#42260497)

Interesting seeing the different ways people come up with "spoofing" this evidence - I don't know if this is a function of the people on this site or if we've seen too many movies with criminal geniuses.

I'm sure that the video source could be tampered with to "prove" something but I don't know if this makes sense in the real world.

Wouldn't the video (with hum) evidence be provided by the police as proof that the accused committed the crime? In that case, rather than the prosecution proving that evidence provided was false the prosecution would have to just provide the chain of custody for the recordings and if there was any concern wouldn't it be up to the defense to verify the validity of the recordings?

I would think that with the hum information, this could be done fairly cheaply and quickly which would be a good thing for all concerned.

myke

Re:Interesting seeing the Skepticism (3, Insightful)

plover (150551) | about a year ago | (#42261871)

One real-world use would be tying the frequency fluctuations to anchor a recording in time. Remember when Osama bin Laden would send out his tapes via courier, and the CIA would analyze them and say things like "we can tell he's still alive as of last week because he mentioned the Yemen bombing." If his tapes matched a certain frequency pattern, they might have been able to say "we know he recorded this on December 12th, because the hum matches the power grid of Pakistan's Islamabad power generating plant." That also would have proven he was in Pakistan, and might even be why the CIA insisted he was.

With continual monitoring of the world's electrical grid (that doesn't seem like an impossible task for the CIA to do), they can discover in which section of the grid such a signal originated. I was watching some of the University of Tennessee Knoxville's video footage on YouTube [youtube.com] generated by the FNET [wikipedia.org], which shows impacts to the grid rippling across the country over time. Disturbances are surprisingly visible, and the way they spread via propagation delays would act like an echolocator to pinpoint someone who sent the recording. Not only could they tie it to a specific plant, but they could tie it to a subsection of the grid with a known delay.

And they could potentially discover this location without having all the recording instruments in place in advance. If they knew the signatures of the Islamabad generator and the Lahore generator, and they knew the recordings contained signatures delayed by 2 msec from Islamabad and 5 msec from Lahore, they could start quietly plugging in line monitors around Gujrat to map out where such a grid section might be.

Of course, if bin Laden's lighting was from a private generator, which is not uncommon in that area of the world, then they wouldn't know such info. But if he was using commercial electrical power, that might have located him to within a region.

Protect us from what, exactly? (0)

Anonymous Coward | about a year ago | (#42260547)

If the problem is detecting whether the facility owner "A" is offering a fake video recording, when "A" owns the facility and the recording apparatus, this isn't going to help. "A" can just stage a fake scene at the facility with an actor, and it will have all the correct audio. It would be much cheaper than hiring a video-editing studio, and fewer people to keep it all a secret.

Do we really need to worry this much about facility owners bringing fake evidence and false charges against innocent people? Chilling.

unbalances? (0)

Anonymous Coward | about a year ago | (#42260693)

The summary uses the work unbalances. Is this technical jargon in the EE field? Should be imbalances if not.

Frequency with Zero crossings? (1, Interesting)

gr8_phk (621180) | about a year ago | (#42260805)

So for the database they measure the line frequency by looking for zero crossings in the voltage waveform. They average this over many cycles, which sounds like a good idea but.... If you take 500 cycles over a minute in Europe, this averaging is still equivalent to taking the time between the first and last zero crossing. Or actually that would be the same as taking the average of the periods. Since the frequency is inversely proportional to the period, using the first and last zero crossings would be more accurate. Noise on the signal (or in the sampling) could shift a zero crossing, which would lengthen one period and shorten the next. That would have no effect on the average period (except the first or last in the batch) but would cause the average for the frequencies to be higher than the actual. This effect may be apparent in the data - the database has a consistent shift upward in frequency compared to the recording which we analyzed using FFT.

Three problems. (1)

iive (721743) | about a year ago | (#42261273)

1. Usually digital sound recoding is lossy. This means that a lot of "information" is discarded. Good audio codec would try to eliminate the noise before going on to simplify the "relevant" audible information. You will need to have raw PCM recoding for proper analysis.
This method makes a lot more sense with analog recoding.

2. The method rely only on variation of the 50Hz main frequency used for the power grid... it is definitely not as precise.
First, low frequencies cause low induction. That's why ac/dc adapters (PC PSU too) usually upscale frequency to 50'000Hz before using small transformer.
Second, the grid frequency deviations are changing slowly because of the inertia of the dynamos that generate the electricity. High power consumption tries to slow down the rotation and automatic feedback compensates with increasing power (aka more steam in the turbine).
This method may be good enough to pin-point the possible time of the day when the conversation happened, but it would not be good enough to say if few seconds here and there have been removed.
Even the article says "if you look at it over time, you can see minute fluctuations."

Actually the method may detect the point of a cut, if the 50Hz main frequency suddenly changes phase. However this could easily be avoided if the one doing the edit does the cuts in lengths of 1/50second (or just a whole seconds).

3. Since the forensic scientist logs the noise, he is in the excellent position to manipulate the recoding so it appears from any time period he desires.
Also... to disprove his analysis one needs access to the same noise logs, but preferably done by somebody else.

As others have pointed out, you can take previous recordings, clear the noise from them, edit them, then add current hum from the power grid. Then present the result as evidence that this conversation just happened.

When it comes to police recoding suspects... there are much better digital methods to ensure that the recoding is not manipulated by a non-expert.

Electrical hum? pah! (0)

Anonymous Coward | about a year ago | (#42261771)

I prefer to use an ominous hum to deal with snac^H^H^H^H bad guys.

Unfortunately... (0)

Anonymous Coward | about a year ago | (#42261787)

This technique is used by police to prove their recordings are genuine and yet the police, or at least the forensics department, obviously have access to the background noise database. So we're back to square one for proving if a recording is genuine.

It's a nice idea, but... (1)

GrumpySteen (1250194) | about a year ago | (#42261867)

That hum isn't just line voltage. Motors of all sorts, florescent lights, switching power supplies, transformers... the list of things that can generate electrical noise at frequencies close to line frequency is huge. There's no way to reliably separate the noise components when the frequencies are close, so you could never be certain which frequency was the line frequency at any given moment. Without being able to differentiate between the noise sources, you'd never be able to match a recording to the researcher's database.

Load More Comments
Slashdot Account

Need an Account?

Forgot your password?

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>
Sign up for Slashdot Newsletters
Create a Slashdot Account

Loading...