Background: While tab capture is active, if a page goes fullscreen, an alternate browser UI is provided: The browser window does not go fullscreen, and the web page becomes surrounded by gray matte, and the web page continues to be rendered within the tab. This allows the user to switch to other tabs/apps on their desktop while capture is remoting content to another device.

Problem 1: For a long time, the web page will see that it's size is something much smaller what a fullscreened page would normally be. For example, because it is still rendered within a tab, the page would detect its size at something like 600x400 to 900x600. Thus, most video streaming sites will choose to download the standard definition content and not the high-definition content.

Problem 2: In some cases, a page will make size assumptions when rendering fullscreened media controls; and so these controls will become obnoxiously-sized on the both the local screen and the remote screen (where the captured video is played out).

Problem 3: Because tab capture is often capturing at a high-def resolution, such as 1080p; having the web page render at a smaller size means we must execute expensive scaling operations on the GPU. In addition, because the web page sizes are very arbitrary (based on the browser window's size), there is quality degradation in most of these scaling operations (because the scale factor is almost never an integer multiple).

Solution: When a page enters fullscreen mode, it's rendering size should be set to the native tab video capture resolution. However, it will then need to be scaled to fit on the local screen in the browser tab content area. Prototypes show this delivers a number of benefits: 1) Pages download high-def streams and can be tab-captured without scaling, improving quality and performance; 2) the controls are still interactable locally, and are almost always a reasonable size for practical use; 3) the down-scaling of the tab content for display in the local UI is cheaper; 4) the down-scaled tab content view with the gray matte surrounding is a much better indication/reminder to the user that the page content is being captured and remoted.