D2SRoBa Conditional Sizing Pass -- Why?

[brett]

I can understand the reason for a Conditional Sizing Pass in D2SRoBa when your target is undersized or oversized. That's obvious. But...

Why is there a "Q over" option for the Conditional Sizing Pass? If your movie is 100% on target, why do you want a sizing pass just because it has a high Q? Is this actually supposed to help quality? If so, in what way does it help? I mean, if it helps, I'll just leave it at 25 or lower since I usually encode at night.

Manolito's post here seems to imply that the "Q over" sizing pass is totally useless:
Does a sizing pass always improve quality? (http://forum.doom9.org/showthread.php?s=&postid=413101)

If this is true, why does the option exist?

[tylo]

Look from this post:
http://forum.doom9.org/showthread.php?s=&postid=411117#post411117

The CCE manual says something like : Q1 ~ 40 : Priority of quality over compressability. Above this, CCE will do compromizes on bitrate allocation. I guess this is what can be improved by doing more passes, but I haven't done any comparisons on this.

[bobwillis]

Hi,

My 'test' was done with Saving Private Ryan (region 1). The Q of this movie is over 70. If you look at chapter 9 (the scene in the church when Miller talks to Oppem), black blocking is obvious on Miller's face with a single OPV pass. The improvement with the additional VBR pass was pretty obvious to me.

I didn't do any more comparative tests, but this proved to me that enabling the sizing pass was a good idea for high Q material.

Regards,
Bob