Avisynth: Lastest revision to the MA script lowered CQ by 5 points
 

the encode isn't finished, but the same movie I did yesterday before the fully adaptive revision picked a CQ 5pts higher using the same prediction parameters. The movie is 90 minutes long and I"m only getting a CQ of 57 for 1 80minCD. I'll report back in 4 hours, but does this sound right?

I thought the CQ was supposed to go up slightly with the latest script? I use a prediction precision of .5%

i was coming to post like you when i see your news audi2honda!

my prediction low in 3 points after
adaptive in the TemporalSmoother filter
and ToK need more tries to find the values.
i don't compare the image quality,only CQs.

Jell post some tests using DCTFilter
and i try using his recomendations and got 1 more point in CQ,
see the post and recomendations:
http://www.kvcd.net/forum/viewtopic.php?t=4725

:!:

I just finished doing an encode with the latest changes and my final encoded video was 5% short from the predicted size. I'm redoing the prediction now with the old full prediction as my first one was done using fast prediction. I haven't tried going back to the previous change to MA, but perhaps I will and post what the results are here for predicted CQ

I experienced too, that the CQ value seems to be 3-5 points lower with the latest script! However, image quality is just superb in my eyes!! :D
It looks far better than my encode which was 3 CQ points higher! 8)
Backgrounds are more stable and Gibbs are almost gone! :D