For example. If at a bar you approached a girl and she said "ew your such a player I saw you talking to other girls". The surface frame is the context here. She's framing you as a player for talking to girls.

The underlying frame is a power frame. If you accept this with "oh sorry yeah maybe your right" you're accepting the power frame she is setting and signaling she is in control. Thus losing the frame collisions.

Is it accurate to say frame control has 2 parts to it?

Different example with different underlying frame

Girl-why do you go to the gym so much it seems like you're obsessed! (Surface frame is gym is bad. Framing you going lots as not good) (Underlying frame again is a power frame if you accept it)

Response-Reframe and set a prize frame "lol I love being fit it feels incredible. I actually only find girls that workout attractive. Do you exercise ?" You don't accept her surface level context and also try to set a prize frame if she falls into it