Here's another approach:

Assuming your tracks are audio (and if there are midi tracks, they are bounced to audio...recommended).

1. Duplicate your tracks (meaning copy just in case you can't undo or undo should fail) and mute the tracks you don't want to modify.
2. SELECT (using the mouse) region on the first track you want the fade-out on (e.g., the last 6 bars), then
3. SELECT the remaining tracks (this assumes you are either selecting the duplicate or original tracks, but not both)
4. From the Edit Menu, SELECT (Audio Effects, Gain Change, Fade-Out (from the pull-down menu), under the Variable Gain options to the right, SELECT Fade Out, or adjustable fade and set the Start Gain and End Gain percents as desired.). The Fade Out (non-adjustable) should give you a linear slope. You can also save the settings as a named preset and later use or modifiy to suit
5. CLICK Process (you can undo if you want to try different settings...but, I still recommend doing this on duplicate tracks ust in case).

NOTE: You can replace step 4 by RIGHT-CLICKING the selected track(s), scrolling down to Audio Effects, and then at the bottom you'll see Fade In and Fade Out (note the subtle division line just above these two options in this view, and that these options don't appear in the Edit/Audio Effects menu). Then SELECT the options you want as in step 4.

The benefit of using the Audio Effects method, is that you avoid inaccuracies introduced with human error interacting with the mouse/track ball.

What would be nice is to have a preset collection of curves (linear, forward slope, aft slope, etc., with the ability to see the line of the curve the ability to add nodes to the curve and pull/bend the curve as desired. I may put this in the wish list for Realband.

The reason I suggest bouncing midi tracks is that it reduces the processing overhead (which may not be a factor on many systems, but certainly reduces the complexity, since midi deals in velocity, while audio deals in gain, and the two fade Outs/Fade Ins must be handled differently.

Hope this is of use.

Richard