Weird idea: Daisywheel printers

Dear All,

Your thoughts on something that I realised about 15 years ago (when Daisywheel printers were common) ... just only got around to telling anyone about it:

On a daisywheel printer (for those of you who haven't used them), there is a wheel with all the letters and symbols attached. The printer spins the wheel (in both directions? maybe it depends on model) and at exactly the right moment a hammer hits the back of the wheel, causing it to hit the paper (through the ribbon of course)

It's a similar principle to some electric typewriters. Although a lot used "Golf balls" instead.

If you can record the sound that the printer makes while printing, is it possible to reconstruct the message from the interval between the bangs (and possibly other sounds that the printer makes)

These printers are very noisy, so you can potentially record them from quite some distance.

If the wheel goes both ways, you can't necessarily determine which way it went, but by analysis of the probable content of the message you might.

Also, you don't know where the wheel started. However there are only a finite number of positions. So by searching for some repeated sequence, you may be able to spot patterns and hence determine where the wheel is.

ahh, are you trying to implement that this is a security issue. i mean sure, if you were sending secret messages against the government or something like that you dont want anyone to find out. but those printers really are out of fashion these days so i wouldnt worry about it.

Some other cool thing I just thought of reading this:

A way of sending encrypted messages with an "One Time Path"..

send the resepient a daisy wheel (special one with scrambled letters) via UPS or something..
send the resepient an e-mail containing a encrypted message..

if the resepient prints the message with the wheel that would decrpyt it...

Downside.. you need to send 1 wheel per email.. else it won't be a OTP and won't be safe..

I think the idea has merit. If you were able to record the amount of time between whacks of the wheel, you might be able to figure out what keys were struck when.
A thought occurs to me, though: you'd have to know what key was struck first, and how long it takes to get from every key to every other key.
Come to think of it, no. I think it's impossible. Too many variables.

/me Hits self for useless post.

PS - Jinx, I like your *new* avatar.

good point slarty, no one can be too careful with sensitive data... yes, decoding the sound of your printer is very plausible... esp. now with modern survalence technologies, etc... your common PI (private investigator) can listen in on a conversation behind close door (via the glass on the window, this technique converts the vibration of the glass back to an audible format) anywa- i don't want ot get too off the subject... but yes, i would consider it a risk if the data is of any real importance, although why not get new printers? or was this just a hyp. situation... good thought!

-take it easy :)

Slarty: Methinks you just got back from lunch at the Turf.... Or was it the Queens around the corner..... :D

I'm sure there were some printers that always returned the head to a datum point at the end of printing but then you'd need to get lucky in so far as the target needs to be using such a printer.... otherwise the start point is an issue. Now if you could get a very high quality recording you may be able to determine the sound certain letters make..... In that case you have your datum right there.

As you pointed out this is 15 years old...... Lets see if you can do it with a laser printer...... :D

it just occured to me. with "modern day" technology, you can literally have a camera sitting on the desk ect

but your not thinking good enough. Try putting a camera in the printer above the wheel. watching every letter printed.

I do recall that "back in the day" during the cold war, the US actually inserted cameras into copying machines, and would go "service" the copy machines to exchange the film. Don't really know where I was going with this...

I think that is a really interesting idea. As Slarty, pointed out, this idea occured to him 15 years ago so I think we're missing the point of his post by saying things like "nobody uses daisywheel printers anymore.".

Slarty's idea is cool assumming you do not have and will not have direct access to the printer or what is being printed from it. For example, things like the camera someone mentioned....well, you would have to first have access to install the camera. However, setting up a bug to record the sounds from the printer would be much easier, unobtrusive, etc. I'm not sure if this is what you were thinking, Slarty, but these are just the thougts that occured to me.

Regarding all the variables that would be involved in making such a daisywheel decoder....yes, lots. However, I'm not sure writing off the idea at face value just because of the amount of variables involved is such a smart thing because even though the daisywheel printer is not widely used anymore, the "idea" itself can be applied to many current technologies. Slarty's idea, in my opinion, seems to be a good "springboard" for modern audio decoders.

Interesting idea, would probably work too...
Wasn't there a similar idea put forward by some researcher for a similar attack on ssh; since ssh (like telnet) pushes each character/keystroke as urgent (ie, no buffering), it was speculated you might be able to statistically deduce what keystrokes (assuming a human typer with average typing skills) were being sent based on the delay between each char sent...

Ammo

Follow up: yep, found a refrence to that: http://www.securityfocus.com/news/241

Oh, if anyone wonders, it was determined that this threat was pretty much of theoritical/academic nature because implementation difficulties and "environmental" variables (like latency...):
http://www.ssh.com/company/newsroom/article/204/
http://www.cs.virginia.edu/~evans/cs...orts/team4.pdf