Why is D2S prediction so inconsistent?
 

I have great success with D2S and its prediction method one time and completely out of wack the very next. All my settings stay the same and one comes out great and the next will be 200mb too big. What could be wrong?

Look at the CQ values tested and you will see where is the problem.

In my last test there were the values :

70 (starting value) -> Too low
83.76 -> too high
69.5 -> too low (Why does D2S test 29.5 when 70 when already
too low :?:
83.36 -> too high
69.5 -> too low (69.5 was already tested in step #3 :!: :!: :!:)
....

At the end the encoding starts with 69.5 (:evil) and the file was 560 MB long :-(

So what do I do to remedy the problem? I'm starting to wonder if I should just switch back to 2 pass?

Not at all ! It's a lot longer and a lot worse (in result).

The best thing to do is to stop D2S when it is in the video encoding phase, then use CQmatic to find the correct CQ, relaunch D2S, do a recover, set the CQ find by CQMaticv and let D2S continue it's job.

Do I need to make any adjustment to CQmatic due to creating chapters with D2S?

No. Chapters do not take any place on the disc.

It works pretty good for me
 

In the 20 or so movies I've converted it has done very well with the built in file prediction. Gets within 20 megs or so on average. I tell it to go to 1 800 meg cd and choose 1.01x's size and it gets to 810-820 pretty consistently (with overburn). And the dvd2kvcd works really well for me too. I've been luck. All NTSC stuff, maybe that has something to do with it.