I've film tested a 400mm f5.6 Nikkor with and without a Nikon made filter and found no measurable difference in resolution using Air Force charts (as used by "Modern Photography" magazine for its lens tests) and focusing with extreme magnification to get repeatable results at a level of care next to impossible in real world shooting. The only filter I tested was a UV. And there certainly could have been a slight contrast reduction but I was not testing for that.
I briefly had a Nikkor 400/3.5 with a front UV that was bonded in place and that lens was as sharp as other f3.5's w/o front filters.
Also in the past I owned a Century Tele Athenar 650mm f6.3 (later models were f6.8) and found no sharpness difference using a Hoya UV compared to the unfiltered lens, testing with the above charts. But when I tried a (non Nikon) yellow filter (for B&W use) the image went to "mush" so much so that the image was unfocusable. It was a very sensitive lens to filter issues. Current (sharper) Canons and Nikkors would likely be more so.
These were tests of the optics alone, so I used electronic flash as my light source in a dark room (or in the case of the 650mm outside at night). That eliminated mirror or shutter bounce issues in the results.
But the issue HERE for use with Red and Scarlet is not really UV filters but ND's (and others) and I have no experience with them in terms of image degradation at extreme tele magnifications yet.