roba - not really one pass vbr???

[aristoteles1000]

In the last 3 of 4 movies roba always needs two passes in order toget a result of 99-101% on the target. By the last movie roba has a result of 95,2% on the target. So roba must start cce twice.
I use roba 3.60, which was in the bundle of DVD2S; is the new roba 3.77 better and can I install it over the old version?
What can I do else to get better results to have really One Pass VBR?

Thanks in advance.


PS:
What is to do, in order to get two dvds on one by DVD2s; each dvd encode with a “cd size” of 2.232 and then use a authoring proggie???? Is thar right?

[jsoto]

D2SRoBa accuracy:
There are some tips to get a better accuracy (like use a 2 % in the sample encodes, mainly in short movies). But there is always room for an estimation error.

If oversized, you can use a transcoder to reduce the size. And if undersized, well man, I think a very low quality improvement will be achieved with 2-5% of size. But, if your goal is the quality (may be important if you are going to put two DVD resolution movies in a DVD5), the best choice is to do 2 passes (with D2SRoBa). The final encode will have the best possible quality and the size will be accurate

Two movies in a DVD5:
The main problem for to two movies in a DVD5 is how to split the capacity of the DVD. IMHO, the goal should be to get the same Q for both movies, but I think it can not be done in an easy way.
A first approximation could be split the capacity taking into account the length, but how to take into account the compressibility?

May be the best option is to put both movies together (its not so easy, cell ids can be repeated in the VOBs) and launch D2SRoBa. See what Q it gets and:
- Encode one movie at fixed Q. See the final size
- Encode the other movie with D2SRoBa with the rest of the capacity. The calculated Q should be similar, (that's the goal).

BTW, I've never tried this procedure, but theoretically, it should work.

jsoto
NOTE: To "merge two movies" in an uniq DVD, you can use AddChapter+VobBlanker.