As Mike indicated, the cause for the differences in sound probably is where in the signalchain you are picking the recorded sound from.
Normally (at least for me) you want to record the signal “au naturel”. For this you pick off right at the input -> the recording sounds exactly the same as what is coming into the desk in a live situation. This also has the benefit that if you play back those recordings any processing you’re applying better represents what happens live.
In summation: what probably happened is that you recorded the ladies without processing and the rest with processing. Find the buttons Mike alluded to in the pic and you’re set.
Good luck !
Giga