I was once told that given the right diagnostic tools Scientists would be able to analyse the plaster in a room’s ceiling and ‘play back’ every sound to have been in its earshot, by scrutinising the minute rearrangement of its matter wreaked by the sound waves over the ages.
Is there even a speck of possibility in this?
Sounds like a whole load of crap. Also mythbusters couldn’t do it 
If this were true, it would bring a whole new meaning to the phrase “The walls have ears”.
Well, you’d think that at the very least, you’d need to know the precise molecular configuration of the plaster when it was installed. Plus, there’s nothing to say that there aren’t multiple ways to get to the same state.
Ancient Pottery Recorded Audio
http://www.museumofhoaxes.com/hoax/weblog/comments/3992/
That said, there’s a stretch of interstate 80 to the west of Grand Island NE where they’ve done such an irregular job of grooving the pavement that the road actually seems to play tunes on your tires. Had they modulated the grooves better, I think they could have produced an intelligible voice.
This was done once, to warm drivers to slow down. It freaked people out so much it was discontinued.
There’s a road in Lancaster, California with grooves engineered to play Rossini’s “William Tell Overture”. From what I can tell from the videos, it isn’t a very accurate rendition. But it is definitely a melody.
I think in principle this would be possible, but in practice would require much more measurement and analysis than will be practical for a long time.
A simpler but similar thing is done by analyzing the temperature profile of deep polar ice. You can imagine that the temperature at the surface is constantly driving heat transfer in or out at all depths, although much less so the deeper you go. Still, there isn’t necessarily any depth at which the influence is zero. It is much easier to start with a temperature history at the top and figure out what the profile in depth would be now - it seems to me this is a one dimensional example of the Direchlet problem, or at least one with time added in, maybe better to say an application of the Poisson equation. It is much harder to start with the profile and deduce the history, an example of a deconvolution problem. A century or so ago Kelvin recognized that this was possible but that one could describe unexciting looking profiles that could not possibly have resulted from any real history, which is a nice example of some kind of imaginary mathematics.
I don’t expect to hear anything interesting about this in the next thirty years or so, as it will still be intractably difficult. I don’t expect to hear anything interesting about it in the years after that, either, though for different reasons.
Not sure if this is the same road, but there’s a car commercial (might have stopped playing by now) that was based on this exact premise, though I think the cuts in the road are supposed to simulate the car brand’s theme