How mains hum is helping forensic scientists to fight crime

[jucole]

http://www.bbc.co.uk/news/science-environment-20629671

[ejeffrey]

It sounds like a great forensic technique for identifying a random recording, but not one that is terribly hard to fake if you know your forgery will be subjected to such analysis. 

[tom66]

Very clever. Someone mentioned this at work today. But they got it the wrong way around with mains frequency. It goes up under greater supply than demand, down under greater demand than supply.

It would be hard to fake unless you had records of all of the mains frequency at the alleged time.

[G7PSK]

The frequency will go down on first meeting greater demand and when the demand cannot be met by increasing the field current the speed is wound up. I have seen the frequency vary between 49 and 51 HZ.

[aluck]

It would be hard to fake unless you had records of all of the mains frequency at the alleged time.

and you will manage to remove original frequencies...

I asked the guy I know from forensics lab. He said that it's a well-known and a widely-used method.

[ejeffrey]

It would be hard to fake unless you had records of all of the mains frequency at the alleged time.

Right, but that isn't exactly secret information, nor is it particularly hard to find after the fact.  Just find some other recording that takes place at the time you want to impersonate.

and you will manage to remove original frequencies...

That is a potentially harder problem.  Like most forms of digital figerprinting, it is hard to do so without leaving any traces, but it all depends on how sophisticated the analysis is compared to how good the forgery is.  For sure a simple automated technique that uses a 2 Hz bandpass filter and a level crossing detector would be easy to fool.

As one tool in a suite of forgery detection tools I imagine it is useful, but it needs to be used carefully.

[SeanB]

How would it cope with highly compressed audio, which generally does a high pass of low frequencies to reduce bitrate.

[aluck]

It is not a universal method, nor that is the only one used. Also that method wouldn't work outside; 5-7 meters from a building and the 50/60 Hz signal is gone.

[Ed.Kloonk]

Time to dig up Nixon and find out what happened to the missing 18 minutes.