To address the issue of videos with captions being tiny, for the CSE site we were able to create a workaround by creating a "Video With Caption" view mode and enabling it for videos. This provided the figure object with a view-mode attribute, which we were able to attach CSS to (the default view options do not provide this class):

(The machine name of our view mode is "video-widget")

figure.media--view-mode-video-widget.caption,

.media--view-mode-video-widget.caption > figcaption

{

display: block;

}

This may be a useful approach to apply at the Folwell level. Let us know if you have any questions.