Not sure if that transfers to compressed music in general. If even the non-compressed guinea pigs had hearing loss, that suggest the level of 102dB, though below the legal safety level, is harmful to their ears. Total harm done if you listen to music that's too loud is proportional to <loudness>*<time>. Since compression makes quiet parts louder, that amount would increase, explaining the increased hearing loss for guinea pigs that listen to the compressed amount. But if the overall level is below the harmful threshold, does that also happen?