[x264-devel] MB mode selection

[Andras Boho]

Hi,

If I understand well, there is an analysis before choosing the final MB
mode (I_16x16, I_8x8 etc). I work in CAVLC mode and I would need some help
with finding the part where the decision is made so I can read out the
content of the right coefficient table.
I have been examining the code but I can't rally find it.

I would need something like this: cavlc.c::block_residual_write_cavlc(
x264_t *h, int i_ctxBlockCat, dctcoef *l, int nC ) but this function is
called way more times than the sum of my MB-s and the dctcoef *l content
seems to be different from the real quantized values to me, so it is not
the one either.. anybody has any knowledge / experience with it?

Thank you for help.