I feel like that was super common. Apart from changing the volumes of entire channels (e.g. changing the level of Line In vs. digital sound), volume was a relatively “global” thing.

And I’m not sure if that was still the case in 1997, but most likely changing the volume of digital sound meant the CPU having to process the samples in realtime. Now on one hand, that’s probably dwarfed by what the CPU had to do for decompressing the video. On the other hand, if you’re already starved for CPU time…