*
* Copyright (c) 2002 Michael Niedermayer <michaelni@gmx.at>
*
- * This file is part of Libav.
+ * This file is part of FFmpeg.
*
- * Libav is free software; you can redistribute it and/or
+ * FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
- * Libav is distributed in the hope that it will be useful,
+ * FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
- * License along with Libav; if not, write to the Free Software
+ * License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
const uint16_t *qmat, *bias;
LOCAL_ALIGNED_16(int16_t, temp_block, [64]);
- assert((7&(int)(&temp_block[0])) == 0); //did gcc align it correctly?
+ av_assert2((7&(int)(&temp_block[0])) == 0); //did gcc align it correctly?
//s->fdct (block);
RENAMEl(ff_fdct) (block); //cannot be anything else ...
if (s->mb_intra) {
int dummy;
- if (n < 4)
+ if (n < 4){
q = s->y_dc_scale;
- else
+ bias = s->q_intra_matrix16[qscale][1];
+ qmat = s->q_intra_matrix16[qscale][0];
+ }else{
q = s->c_dc_scale;
+ bias = s->q_chroma_intra_matrix16[qscale][1];
+ qmat = s->q_chroma_intra_matrix16[qscale][0];
+ }
/* note: block[0] is assumed to be positive */
if (!s->h263_aic) {
__asm__ volatile (
block[0]=0; //avoid fake overflow
// temp_block[0] = (block[0] + (q >> 1)) / q;
last_non_zero_p1 = 1;
- bias = s->q_intra_matrix16[qscale][1];
- qmat = s->q_intra_matrix16[qscale][0];
} else {
last_non_zero_p1 = 0;
bias = s->q_inter_matrix16[qscale][1];
" js 1b \n\t"
PMAX(MM"3", MM"0")
"movd "MM"3, %%"REG_a" \n\t"
- "movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1
+ "movzbl %%al, %%eax \n\t" // last_non_zero_p1
: "+a" (last_non_zero_p1)
: "r" (block+64), "r" (qmat), "r" (bias),
"r" (inv_zigzag_direct16 + 64), "r" (temp_block + 64)
" js 1b \n\t"
PMAX(MM"3", MM"0")
"movd "MM"3, %%"REG_a" \n\t"
- "movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1
+ "movzbl %%al, %%eax \n\t" // last_non_zero_p1
: "+a" (last_non_zero_p1)
: "r" (block+64), "r" (qmat+64), "r" (bias+64),
"r" (inv_zigzag_direct16 + 64), "r" (temp_block + 64)
"psubusw "MM"1, "MM"4 \n\t"
"packuswb "MM"4, "MM"4 \n\t"
#if COMPILE_TEMPLATE_SSE2
- "packuswb "MM"4, "MM"4 \n\t"
+ "packsswb "MM"4, "MM"4 \n\t"
#endif
"movd "MM"4, %0 \n\t" // *overflow
: "=g" (*overflow)
block[0x3E] = temp_block[0x3D]; block[0x27] = temp_block[0x36];
block[0x3D] = temp_block[0x2F]; block[0x2F] = temp_block[0x37];
block[0x37] = temp_block[0x3E]; block[0x3F] = temp_block[0x3F];
+ }else if(s->dsp.idct_permutation_type == FF_LIBMPEG2_IDCT_PERM){
+ if(last_non_zero_p1 <= 1) goto end;
+ block[0x04] = temp_block[0x01];
+ block[0x08] = temp_block[0x08]; block[0x10] = temp_block[0x10];
+ if(last_non_zero_p1 <= 4) goto end;
+ block[0x0C] = temp_block[0x09]; block[0x01] = temp_block[0x02];
+ block[0x05] = temp_block[0x03];
+ if(last_non_zero_p1 <= 7) goto end;
+ block[0x09] = temp_block[0x0A]; block[0x14] = temp_block[0x11];
+ block[0x18] = temp_block[0x18]; block[0x20] = temp_block[0x20];
+ if(last_non_zero_p1 <= 11) goto end;
+ block[0x1C] = temp_block[0x19];
+ block[0x11] = temp_block[0x12]; block[0x0D] = temp_block[0x0B];
+ block[0x02] = temp_block[0x04]; block[0x06] = temp_block[0x05];
+ if(last_non_zero_p1 <= 16) goto end;
+ block[0x0A] = temp_block[0x0C]; block[0x15] = temp_block[0x13];
+ block[0x19] = temp_block[0x1A]; block[0x24] = temp_block[0x21];
+ block[0x28] = temp_block[0x28]; block[0x30] = temp_block[0x30];
+ block[0x2C] = temp_block[0x29]; block[0x21] = temp_block[0x22];
+ if(last_non_zero_p1 <= 24) goto end;
+ block[0x1D] = temp_block[0x1B]; block[0x12] = temp_block[0x14];
+ block[0x0E] = temp_block[0x0D]; block[0x03] = temp_block[0x06];
+ block[0x07] = temp_block[0x07]; block[0x0B] = temp_block[0x0E];
+ block[0x16] = temp_block[0x15]; block[0x1A] = temp_block[0x1C];
+ if(last_non_zero_p1 <= 32) goto end;
+ block[0x25] = temp_block[0x23]; block[0x29] = temp_block[0x2A];
+ block[0x34] = temp_block[0x31]; block[0x38] = temp_block[0x38];
+ block[0x3C] = temp_block[0x39]; block[0x31] = temp_block[0x32];
+ block[0x2D] = temp_block[0x2B]; block[0x22] = temp_block[0x24];
+ if(last_non_zero_p1 <= 40) goto end;
+ block[0x1E] = temp_block[0x1D]; block[0x13] = temp_block[0x16];
+ block[0x0F] = temp_block[0x0F]; block[0x17] = temp_block[0x17];
+ block[0x1B] = temp_block[0x1E]; block[0x26] = temp_block[0x25];
+ block[0x2A] = temp_block[0x2C]; block[0x35] = temp_block[0x33];
+ if(last_non_zero_p1 <= 48) goto end;
+ block[0x39] = temp_block[0x3A]; block[0x3D] = temp_block[0x3B];
+ block[0x32] = temp_block[0x34]; block[0x2E] = temp_block[0x2D];
+ block[0x23] = temp_block[0x26]; block[0x1F] = temp_block[0x1F];
+ block[0x27] = temp_block[0x27]; block[0x2B] = temp_block[0x2E];
+ if(last_non_zero_p1 <= 56) goto end;
+ block[0x36] = temp_block[0x35]; block[0x3A] = temp_block[0x3C];
+ block[0x3E] = temp_block[0x3D]; block[0x33] = temp_block[0x36];
+ block[0x2F] = temp_block[0x2F]; block[0x37] = temp_block[0x37];
+ block[0x3B] = temp_block[0x3E]; block[0x3F] = temp_block[0x3F];
}else{
if(last_non_zero_p1 <= 1) goto end;
block[0x01] = temp_block[0x01];
projects / ffmpeg / blobdiff