MilesAhead, you likely already know this but I'll mention it anyways-- there is a HC .pdf in the HC miscellaneous sub-folder (though apparently not with the Oct. HC?) which provides info regarding parameters like bias
, adaptive quantization
From a logical point of view, I don't get why I should use a high bias value (say greater than 50) when performing a VBR encode.
the free version of DVD-RB uses a value of 20--seems more logical to me. I searched a while back and saw iirc where jdobbs (DVD-RB developer) recommended a bias value of 20.
FYI when I remarked earlier about my "seriously screwy results" with the September HC 024, I was referring to something I'd never seen before--on the 2nd pass HC was dropping bitrate like crazy (15-20% less than the 1st pass) rendering significant undersizing. I'd never seen more than a 1-2% variation between passes before, so something? was very clearly wrong. My guess
is that my experience was related to the new "deadzone quantization" parameter which is now adaptive with Oct. HC, but until I have time to run multiple trials that's all it is-- a guess.