Roto shape clipping issue with Silhouette Paint

Using Silhouette Paint in Nuke 12.1, I have tracked 4 Roto shapes using the Mocha Tracker option.

All tracks are stable. 3 of my patches and all of my blemish fixes are solid. The large patch at the bottom is the issue in that it seems to clip some sort of safe boundary at top and bottom of screen. This does not make sense to me as I have other patches passing the boundary of the top of the screen.

Does anyone know of a reason why I am getting this issue?

Can you confirm that you are running the newest version of Silhouette Paint which is 2020.5.5.

Confirmed, running the most current version downloaded a couple weeks ago.

Can you direct message me a download link with your Nuke script, the Silhouette project and 1 frame of the sequence where the problem is occurring and I’ll take a look?

I have attempted to attach the requested files a number of times, however, it appears as though it may be too large for this messaging service to accept as the Silhouette project is almost 85 MB.

Have you got a direct email I can attach the .zip file with the project scripts and requested frames?

I don’t think an 85 mb email attachment will work. Can you direct message me a download link from something like Google Drive or Dropbox? Thanks.

https://drive.google.com/drive/folders/18XuYf7-aaMXwC6uSI28Ej1TAUdOajzQ9?usp=sharing

This contains both script files (I’m not sure if the Silhouette project file uploaded correctly) and the requested frames, I have also included the entire source clip.

Frame range: 50-120
Reference frame: 105
Clipping frame: 112

@walker.jerrid I am not sure exactly what the problem was, but I deleted the Large Scratch group and did a Rebuild. I then re-cloned the area on one frame and repeated on all others and it solved the problem. Please make sure that the area that you are cloning from will be visible in all frames if you are painting one frame and replaying on other all frames with tracking. Otherwise, the area that you are cloning from can slide off the frame.

That doesn’t seem to make sense as the Large Scratch group was the only portion of the tracked areas that was within the frame the entire time, I based the frame range upon that criteria. Regardless, I will attempt the same fix you describe and see what results I get.