I thought "--qlp-coeff-precision-search" was only suppose to search for a greater compression possibility.
I encoded a file using "flac -8 --verify(since all -v seems to do is show version info) ---exhaustive-model-search" and I got a file that was 11,572k.
That is down from 11,706k when using only "flac -8 --verify".
So why is it when I use "flac -8 --verify --exhaustive-model-search --qlp-coeff-precision-search" I end up with a file that is 11,601k?
Why does "--qlp-coeff-precision-search" make the file bigger and how should I know when to use this feature (that is suppose to make the files smaller from what I read) considering what just happened?

Are there any possible instances where "--exhaustive-model-search" could cause files to be larger than without this command?

something is wierd since '-8' include '--exhaustive-model-search'.

the encoder has to do some estimations in a few places. the small size difference you are seeing is a result of that. using anything beyond -8 is usually not worthwhile.

Josh

So what is the point of "--qlp-coeff-precision-search" if you sometimes might get larger files?
It would seem a more useful feature if the "--qlp-coeff-precision-search" output file size was auto compared to the non "--qlp-coeff-precision-search" output file size to see which one was really smaller.
That would mean that 2 estimates would have to be done but its what you would have to manually anyway in order to find out which would really be the smallest file.

it tends to work better on higher sample resolutions other than 16bps.

Josh