So - I'm putting this here because it's not specific to a program but was inspired by the question in the PREL13 board about cross dissolving text.
Bob suggested changing the opacity - and I suggested using a blank piece of video - but neither of these actually do the "right" thing for cross dissolving multiple tracks. This would seem to apply to all transitions that use variable transparency between the clips. A picture or three to illustrate:
Say we have the garish title as follows made of two tracks - one chequerboard pattern and one text track:
This is harder to see when you're actually dissolving so imagine that we are dissolving blank video (as opposed to fading to black which works differently). If we're most of the way through the fade then the programs will have reduced the opacity of both the pattern and the video and you'll get something like this:
Each track is being faded into it's next track and because that's increasing transparency you see the chequerboard showing through the text. Whilst technically correct what we probably wanted was to fade the title into the other piece of video and the fade would actually look something like this.
Now - obviously it's reasonably simple to achieve a fade to black that looks like the last example but a fade between two videos and titles seems harder. What I think is needed is some sort of "composite before transition" switch where a group of tracks or events would have the composite work done on them - then the transition applied rather than the other way round. Do any of the more "grown up" packages do this? Just for interests sake....
Final examples. Here's a cross dissolve between two titles. In this one you can see both check patterns in both pieces of text.
Whereas this one shows one check pattern in each piece of text.