[XviD-devel] [In progress] devapi4 -- mpeg matrices

[Edouard Gomez]

Edouard Gomez (ed.gomez at free.fr) wrote:
> skal (skal at planet-d.net) wrote:
> > 	i confirm: i simply forgot to add this usual poor man's
> > 	RD-opt. MPEG4's intra bias should be:
> > 		bias = (3*quant*intra_matrix[i]+2)>>2
> > 	Nice fishing! :) Thanks
> 
> Ok  i'll  change  the  bias  computation  in  the  quantization  context
> initializer (equivalent to your SKL_QUANTIZER type)
> 
> Btw,  i had left  a serious  bug, not  copying the  quant matrix  in the
> context, so trellis was simply unoptimizing the coefficients.

Fixing that stupid bug, i got back to 38.8x dB :-) 

Then i changed the bias value and got 39.11 dB w/o trellis (which is
0.01 dB better than old code)

Does this code looks right to you skal ?

void
quant_mpeg_init_context_c(quant_context_t *q,
                          const uint8_t *matrix,
                          const int32_t intra)
{
    int qp;

    for (qp=1; qp<32; qp++) {
        int i;
        for(i=0; i<64; i++) {
            const int S = qp*matrix[i];
            const int shift = mpeg_shifts[qp-1];
            q[0][qp-1][0][i] = FIX(S<<1, shift, 1);  /* quant */
            q[0][qp-1][1][i] = intra ? ((3*S+2)>>2) : 0;
            q[1][qp-1][0][i] = S<<1;                 /* dequant */
            q[1][qp-1][1][i] = intra ?      0       : (S>>1);
        }
    }
}

-- 
Edouard Gomez