If you think of old cassette tapes, the tape has a coating of ferric oxide powder (Iron oxide (FeO) is the red rust we commonly see. Ferric oxide (Fe2O3) is another oxide of iron.) and the sounds are recorded onto the tape using an electromagnet which magnetizes the oxide.
So… if you take an old house with the right materials (like rusted iron pipes), could you magnetize the object and record a sound into it? I’m thinking it would take the right frequency of sounds and the right conditions (arid? Magnetized dust particles?). And then on stormy nights … say a “magnetic storm” the conditions of the storm act like the playback side and the sounds are heard.
So why are they heard as moans and such … If you record something onto a tape, you have to play it back at the exact same rate (ie. X inch/second) and in the same order. If you play it too fast or too slow it will sound completely different. My feeling is that the conditions of recording vs. playback would be different and it may play the bits of material in a different order as it wouldn’t be running linear.
If I ever had the money/time I would love to do a study on this. See if the houses themselves have common materials, if the times the sound is heard has any common factors or conditions.
If you think about it, the concept of recording is nothing more than magnetizing and object. Who said the object had to only be linear tape?