Steal space... Half-D1... guidelines ??

[hallway]

Previously I used DVD Shrink and so that I'd get optimal quality for the movie, I'd end up doing "Movie only" with Shrink and could usually get under 20% compression to fit on a DVD5. There's times I'd like to have the menus, extras, etc so I've bought DVD-RB. Now I need opinions on when to use the "Steal space from extras", "Half-D1", and so on settings. Problem is, it depends... Right ?? :rolleyes: I use CCE SP 2.7 for the encoder, by the way.

I think it was my first run with DVD-RB and I did 50% steal and checked both of the Half-D1 and (the other option). In that case, the extras were downright horrible. Unwatchable, in fact.

I guess it comes down to how big the main movie is, doesn't it ?? It's a tossup of how good you want the movie vs how so-so good you want the extras.

What do others use or do ?? That's probably easier to answer.

[arsmori]

There's no rule, but a lot of personal recipes. I say if your source is clean (i.e. digital masterized blockbuster) you can go as low as 3200Kbs avr even lower with low bitrate matrices without too much perceptual detail lost. I backed up Downfall at ~2800Kbs avr with acceptable results. If the source is less than perfect, a dirty B&W print or an old telecine for instance, you should stay above 3400Kbs avr or use filtering, the later should be an exception if you value fidelity.

Once you have determined your main title target bitrate then you can decide what to do the extras, that is to strip them all, keep some, use higher compression rate on them, or any combination of the three. You have to balance and it's time consuming, but you learn a little bit more each time and get faster.

To answer your question more directly, I think I've only compressed extras to +50% once, being 50% more than the main title it's usually too much to tolerate. However there not much perceptual difference between 33% and 25%, so I usually pick 33% instead if I need to go there, else it's 10%, more of psychological use of the steal space option really, the space gain being insignificant usually.

One neat trick I use sometime on criterion release is VobBlanker still (with sound) on static interviews. When used on VTST title, this is truly a amazing space saver and at least you don't loose the extra information which is all verbal in those cases, better a still image that change each cells than loosing the content all together. Well, for that kind of material anyway. The subs are even preserved, pretty cool.

[blutach]

I'm sure there's lots of opinions but Half D1 and 25% is about the lowest I'd go and then I'd make sure I used a matrix consistent with the avg bit rate.

Regards