You'd basically have to write your own homegrown shazam using spectrogram software (waveforms only show amplitude). It should allow you to analyze the entire 8 hour track and build an output of note/frequencies over time, chunk them, fingerprint them and then do the same with your audio files and scan looking for matches. Which means you could find it faster by just listening to an hour or two a few days until you find it haha.
https://www.sonicvisualiser.org/ seems like it could help with the analysis.
A little more searching and there are some options like freezam (
https://github.com/Lizzi-Busy/freezam) that let you build your own database, but I'm not sure that would work well with a single 8 hour clip. Though, you could split it up into 5 minute chunks, process them into a reference database in freezam and then find what 5 minute chunk your audio fits with.