Somebody already mentioned the digital cassettes. Commodore had them as well as big mainframes with reel tape at one time. There were/are the phone modems. You can listen to those, but they are digital in nature transmitting digital information. Then there were keypunch cards and paper tape.
As for Diffmaker, it can do amazing things, and about as often fail and more often simply blow up. Now when it is amazing it takes two files that aren't digitally synched, adjusts for sample rate differences and then for sample rate drift. Then gives a nice deep null. In the course of doing that it will blow up nearly half the time. Heck I tried a file in it yesterday which was consecutive runs where nothing changed. It was a loopback so no synching issues involved. Manually there were nulls of about -106 db. Diffmaker said correlation null of 14 db. Nothing weird about the file.
@pkane is working on a similar bit of software. So maybe he'll get it finished and we can get reliable results more often than with Diffmaker.
When you get to deep nulling it takes so little to corrupt it. I'll sometimes embed two tones an octave apart. If you null them and the null residual is 6 db higher for the higher frequency tone you know the bulk of the residual is a time shift.
The reason I started work on DeltaWave was that AudioDiffMaker was so unpredictable. Sometimes it worked, sometimes it crashed, and sometimes the result seemed completely wrong.
While DeltaWave is far from commercial quality software, it's close enough to be usable by 'knowledgeable' enthusiasts, so I'll try to get a beta version of it out in the next couple of weeks. Just need to free up some time between work and family obligations to package it up