Understanding the --crop option

[emerylin]

My understanding is that --crop should be used if I am going to be injecting the RPU into an encoded video with the black bars cropped out. But why doesn't the --crop option also accept the specific cropping dimensions used in the encode (i.e. --crop 0:0:128:128)?

[quietvoid]

Because the dimensions aren't necessary.
--crop sets the active area offsets to 0, it doesn't need anything else.

Now if your video still has letterbox bars, you would have to edit the RPU accordingly, but only if the resolution changed.