point_cloud_xyz input constraints

edit

Dear all, I could really need some help as I spent many hours on the point cloud, and don't know anymore where to search further. Thus I want to gain some general understanding. What are the constraints for the input topics:

• Do the frame_ids of all incoming topics need to match? (in my case, they are all the optical frame of the rgb camera)
• Does the time sync need to be exact?
• Does the depth image need to have any special format? (I use the convert_metric nodelet and it has now encoding "32FC1", so I guess I should have canonical representation )
• ... Anything else I miss here?

The status in my case is, that the point cloud topic appears (for example in rostopic list or in rviz), but there is no data published on it. In rviz, it has a "green status".

I would apperciate any ideas as I don't have any left ;-)