PDA

View Full Version : compressibility test totally out of bounds


theReal
6th February 2002, 01:27
I've been using Gknot for four or five movies now, and every time it worked just like in the guide and/or the help file:

load project, clip, resize, do a compression test, see if the result is somewhere between 60% and 75% (maybe adjust bitrate and/or resolution if it's not).

Now I have this movie (a German movie called "der Schuh des Manitu") that should be fine for a one-cd rip: 81 minutes, nice quality. Everything works fine until I do the compressibility test:
The result is around 4400% !!!
So, no matter what bitrate or resolution or format I chose, I will obviously never get this in the 60-75% range.

I tried 5%, 6%, 10% and 15% compressibility test, it's still the same.

The fact is, it must be plain wrong! The projected filesize in the nandub status window (while doing the compressibility test) shows values that should result in something like 60%.
The desired size is 700MB, 5% of that is 35MB - so the projected size in a 5% compressibility test is perfect at ~60MB, right? So why does it say 4400% in the end ???? I mean a little out of bounds, ok, I still believe it, but 4400% is just so far out it can't be right!


Does somebody have an idea why Nandub would write a stats file completely wrong while it shows plausible values in the encoding status window?

theReal
6th February 2002, 02:30
I just finished a whole first pass with nandub and loaded the stats file - I was very reliefed to see 42% instead of some weird 4000ish value...

Now I gotta do another first pass, because 42% is too low, but at least it works!

Still, why could it be that the compressibility test is so totally wrong? Any ideas?

rmatei
6th February 2002, 19:05
Are you sure you didn't load the comp-check stats as a regular stats? That would explain it... But then again you'd know better, right?

theReal
6th February 2002, 19:14
I finally encoded the movie, ripped the next one ('Trainspotting') and did exactly the same - this time the compressibility test again worked like it should...

I can't think of any possible reason for the weird values with that one movie - but now it's encoded and I hope the problem will never occur again :)