mask and subtracting objects

 sure this is the screenshot i still see

 here is the screenshot I also still see the shadow of the candle even when subtracting the object. when i put a mask video over it which i intend to do i just dont want to see the refelction of the candle.

Content-Aware Fill needs enough similar content around the edges of the mask on adjacent frames to find pixels to fill the frame. 

Imagine someone walking down a beach. They are on the right side of the frame at the start of the video, and they walk to the left side of the frame a few seconds later. The shot is from above, so you see a little shoreline, the person, and the water. You draw a mask around the person, track them or manually keyframe the mask, so the person is always inside the mask. CAF looks at the hole and then looks forward and backward in time to see if there is a time when there is no person where the hole is now. If there is no person, CAF takes the pixels from the later or earlier frames and fills the hole. It's basically an AI-assisted time shift that fills the hole. 

Now imagine the shot is someone walking through times square on New Year's Eve. People are shoulder to shoulder and moving in different directions. You manage to mask out your person successfully, leaving a hole, but every frame is different. Nothing ahead or behind the current frame that closely resembles the scene if the masked actor was gone. CAF will fail on a shot like that.

Show us your shot, and we should be able to tell you if it will work. 

You have to show us screenshots. Content Aware Fill is not a magic tool and how well it works can vary from shot to shot. perhaps another approach might work better.

Mylenium