AFAIK, where the CC appears on your screen is pretty much determined by the transcriber/software at the source of the captioning. I agree with you. I have found far too many LAZY people leaving the caption right where they appear obscure titles of speakers and other text that was already a part of the finished production. Now, the bugs added by the TV channels by on-air TD's are a different story. However, I think some TV's give you a very narrow option of having the CC appear further from the raster by like a half-inch or so, but the rest is all OUT OF OUR CONTROL and determined by the individual transcribing. Live transcription is the WORST and sometimes useless as the transcriber is either too slow or can't hear well (often listening via POTS hundreds or thousands of miles away from the broadcast, and they don't know the names of the local communities a local newscast is refereeing to, so lots of WRONG names and spellings and outright just plain skipping a whole sentence or more because they just don't understand or are too slow.
For example, as a rule of standardization (while some institutions' conventions may be used by many as the standard reference, there is no real single standardization of CC) the CC should appear near the bottom. Then when a super title appears to identify a speaker, for example, the CC should be momentarily moved to the TOP so that people can read the speaker ID and not obscure the speaker's face. Then when the speaker title is no longer on screen, the CC should re-appear or continue at the bottom. This requires a few extra keystrokes, but we are talking about a "pre-produced" with CC material, not a live broadcast. If the CC appears and obscures something important like text identifying a speaker, that is the OPERATOR's HEADSPACE and LAZINESS, and, unfortunately, you can't change a bloody thing about it.
However, there is great control, in most devices, to set your font, size, color foreground, color background and even capitals. Dish boxes, thankfully, offers a wide range of options for appearance. As suggested, try the transparent or translucent (I use that and set it for others, and they prefer it) to help with text not being obscured. Also, anytime you press SELECT or otherwise cause the channel banner at the top to appear, it temporarily prevents CC from appearing until the channel banner disappears by pressing CANCEL or waiting a few seconds for it to disappear on its own.
I can't tell you how often I have to skip back, press SELECT to see the speaker title, press CANCEL to rid the channel banner faster, then often skip back to catch the entire CC string. Different houses provide the CC and different PEOPLE are the biggest factor on how well the CC's are done. I could always tell on the Match Game on GSN that there were two people who split the duty of transcribing to CC: One person CC's every remark, every half sentence interrupted, every small chatter off camera (often identifying who was speaking) with NOTHING left not transcribed, and that same person knew the voice of the actor or celebrity that Dawson or others were imitating (such as CC'ing "In the voice of Paul Lynd") The other person would often skip entire sentences, even several sentences, would never CC the chatter of the panel, even when it was loud and clear and they did display some ignorance of what the panel or contestants were speaking of. In short, a really crappy job.
Sorry, my point is that CC's are still far too dependent on a really good transcriber to be effective. Now, scripted TV shows and movies are usually done properly because they have the script in front of them. Oh, and they should properly synch the CC with the action on screen.
Good luck.
