On sob, 2006-02-04 at 21:51 +0100, Herman Robak wrote: > On Sat, 04 Feb 2006 21:18:23 +0100, <[EMAIL PROTECTED]> wrote: > > > On Sat, 4 Feb 2006, Nicolas wrote: > >> It worked. I would have preferred to be able to use the 1 and 4 keys of > >> the compositor, since that would have avoid me going from the main > >> window to the main compositor on each frame. But as I said, that worked, > > > > The 1 and 4 keys should still be usable... you just have to pay close > > attention to what direction you're moving, and understand that the cursor > > is *between* frames, not *on* a frame. > > Is there a justification for this somewhat odd behaviour? > Is it more logical? More usable (I think not!) Easier to program?
I've found the source of the mask problem - actually the mask is always rendered absolutely correctly... what is displayed wrong is the overlay... The reason is that in cwindowgui.C in function do_mask() there is no recollection of playback direction and position is taken from mwindow->edl->local_session->get_selectionstart(1) and PLAY_FORWARD is used for calculating the mask overlay position - which is naturally wrong whan playing backwards. The question is how to get data about how the frame we are overlaying was rendered. What we should do is probably store the position and direction of the image that is put into overlay buffer and then take positioning and direction information from that data when creating an overlay for it - pretty logical. Thus _not_ using get_selectionstart() when for positioning information at all. Quite some research is needed to do this properly... anyone? bye andraz _______________________________________________ Cinelerra mailing list [email protected] https://init.linpro.no/mailman/skolelinux.no/listinfo/cinelerra
