Thread is 12 years old and I have it bookmarked and still constantly refer to it.
I'm still not sure I understand it, I'll fiddle around and see what I can figure out. Thanks again.
Parallel compression isn't all that complicated when it comes down to it — it's mostly a way to create a compression curve (more compressed for loud bits, less so for quiet bits, making the quiet bits louder without making the whole thing sound unnatural) without mucking around too much with numbers.
If the crowd noise is louder than whatever else is going on, dynamic compression is a great way to lower its relative volume; as kindms suggested above, just eyeball what dB level the actual music reaches, and compress the hell out of anything that exceeds that in the sections with crowd noise. If you have loud music plus loud crowd it's trickier, and starts requiring things like Izotope filters.