Changing High % setting of alt-CC does *nothing* between encodes?

[unplugged]

I'm using 30-08-2002 Koepi's compile, this should be based on 18-8 binaries and almost the same as the latest (4-09).

The disarming thing :( that I'm encountering is that alt-CC behavior seems to not care about High % curve compression setting

I have an example of two 2nd-pass encodes that differs only by high % value, one with 250%, the other with 60%, the result is identical (!).
res: 640x272
Starting from XviD base settings, plus some my changes...:
Motion search precision: 6
Quant type: MPEG
Min I-frame interval: 1, Max: 1500
Min I-frame: 3, Max: 31
Min P-frame: 3, Max: 31
I-frame boost: 0%, below frame dist.: 10, I-frame bitrate reduction: 30%
AltCC enabled, Medium type, High (several tries) , Low 99%
Auto MRC strength 0% (no intervention by MRC)
Manual bonus bias: 100% (setting this to 0 or auto does change almost nothing to log...)
http://space.virgilio.it/hehflo@tin.it/AltCC-test.rar (200k, log files)

I have also experienced this problem before, it seems that higher part of curve "have no intention" to compress :), :(. It only reflects proportionally following the 1st-pass/2nd-pass ratio.

Please, tell me what I'm missing if so or give a try... it's very strange.

P.S.: at least WinRAR 2.90 to decompress.

[Didée]

Hey, unplugged,

where did you have your brains lately

If you enable auto MRC, setting Strength=0 *disables* any curve compression, giving you pure VBR according to framesizes.

You should already have known ...

[unplugged]

Originally posted by Didée
Hey, unplugged,

where did you have your brains lately

If you enable auto MRC, setting Strength=0 *disables* any curve compression, giving you pure VBR according to framesizes.
D'OH!

I have missed completely the relativity of this param, I wrongly believed that reference point was average frame size instead of linear-scaled frame size, so I have set it to 0 to disable MRQ (the opposite!)... sob! :(

Originally posted by Didée
You should already have known ...
Every time MRQ was discussed I never focused that reference point for MRQ factor was linear-scaled frame size...

Thanx