theora dev - Mar 2011 - Examples of matrices before FDCT

Why do we need a function x86/mmxencfrag:oc_enc_frag_sub_128_mmx()?

^ 28 28 28 28 28 28 28 28
^ 28 28 28 28 28 28 28 28
^ 28 28 28 29 29 29 28 28
^ 28 29 28 29 29 29 28 28
^ 28 29 28 28 28 28 28 28
^ 29 28 29 28 28 28 28 28
^ 29 28 29 29 29 29 28 28
^ 29 29 29 29 29 29 29 29

After _sub_128_mmx:

^ -100 -100 -100 -100 -100 -100 -100 -100
^ -100 -100 -100 -100 -100 -100 -100 -100
^ -100 -100 -100 -99 -99 -99 -100 -100
^ -100 -99 -100 -99 -99 -99 -100 -100
^ -100 -99 -100 -100 -100 -100 -100 -100
^ -99 -100 -99 -100 -100 -100 -100 -100
^ -99 -100 -99 -99 -99 -99 -100 -100
^ -99 -99 -99 -99 -99 -99 -99 -99

After FDCT, x86/sse2fdct:oc_enc_fdct8x8_x86_64sse2(), it is:

*    -3187        3       -3        2        0       -1        1       -1
*       -5       -1        0        1        0       -1       -1       -1
*        4        0        3        0        1        2        2        2
*       -4        4        5       -1       -1        1        1        2
*        2        0        2        0       -1       -2       -3       -2
*        2        1       -4        3        2        0        2       -1
*        2        1        2        0        1        1        1        1
*       -1        0        0        1        0        0        0       -1

It is right?

So, we need 13 bits for storing DCT coefficients (for 8-bit pixels)?

See formulas - http://www.cs.cf.ac.uk/Dave/Multimedia/node231.html
F(i,j) = 64*255*2/8 = 4080 = 12 bits; + Sign = 13 bits (-4096...+4095)?