Hi! This just my guess, I have no idea - maybe fpu just returns nan for something/0, and not a divide by zero exception ? > error. I only don't know why it appeared first when I migrated to > integer calculations in tvi_v4l.c, where possible. Denes