--- decoder.c 2002/10/10 12:16:00 1.37.2.3 +++ decoder.c 2002/11/16 23:38:15 1.43 @@ -3,17 +3,13 @@ * XVID MPEG-4 VIDEO CODEC * - Decoder main module - * - * This program is an implementation of a part of one or more MPEG-4 - * Video tools as specified in ISO/IEC 14496-2 standard. Those intending - * to use this software module in hardware or software products are - * advised that its use may infringe existing patents or copyrights, and - * any such use would be at such party's own risk. The original - * developer of this software module and his/her company, and subsequent - * editors and their companies, will have no liability for use of this - * software or modifications or derivatives thereof. + * Copyright(C) 2002 MinChen + * 2002 Peter Ross * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by + * This file is part of XviD, a free MPEG-4 video encoder/decoder + * + * XviD is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * @@ -26,36 +22,36 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA * - *************************************************************************/ - -/************************************************************************** + * Under section 8 of the GNU General Public License, the copyright + * holders of XVID explicitly forbid distribution in the following + * countries: + * + * - Japan + * - United States of America + * + * Linking XviD statically or dynamically with other modules is making a + * combined work based on XviD. Thus, the terms and conditions of the + * GNU General Public License cover the whole combination. * - * History: + * As a special exception, the copyright holders of XviD give you + * permission to link XviD with independent modules that communicate with + * XviD solely through the VFW1.1 and DShow interfaces, regardless of the + * license terms of these independent modules, and to copy and distribute + * the resulting combined work under terms of your choice, provided that + * every copy of the combined work is accompanied by a complete copy of + * the source code of XviD (the version of XviD used to produce the + * combined work), being distributed under the terms of the GNU General + * Public License plus this exception. An independent module is a module + * which is not derived from or based on XviD. * - * 15.07.2002 fix a bug in B-frame decode at DIRECT mode - * MinChen - * 10.07.2002 added BFRAMES_DEC_DEBUG support - * Fix a little bug for low_delay flage - * MinChen - * 28.06.2002 added basic resync support to iframe/pframe_decode() - * 22.06.2002 added primative N_VOP support - * #define BFRAMES_DEC now enables Minchen's bframe decoder - * 08.05.2002 add low_delay support for B_VOP decode - * MinChen - * 05.05.2002 fix some B-frame decode problem - * 02.05.2002 add B-frame decode support(have some problem); - * MinChen - * 22.04.2002 add some B-frame decode support; chenm001 - * 29.03.2002 interlacing fix - compensated block wasn't being used when - * reconstructing blocks, thus artifacts - * interlacing speedup - used transfers to re-interlace - * interlaced decoding should be as fast as progressive now - * 26.03.2002 interlacing support - moved transfers outside decode loop - * 26.12.2001 decoder_mbinter: dequant/idct moved within if(coded) block - * 22.12.2001 lock based interpolation - * 01.12.2001 inital version; (c)2001 peter ross + * Note that people who make modified versions of XviD are not obligated + * to grant this special exception for their modified versions; it is + * their choice whether to do so. The GNU General Public License gives + * permission to release a modified version without this exception; this + * exception also makes it possible to release a modified version which + * carries forward this exception. * - * $Id: decoder.c,v 1.37.2.3 2002/10/10 12:16:00 Isibaar Exp $ + * $Id: decoder.c,v 1.43 2002/11/16 23:38:15 edgomez Exp $ * *************************************************************************/ @@ -136,15 +132,6 @@ return XVID_ERR_MEMORY; } - if (image_create(&dec->refh, dec->edged_width, dec->edged_height)) { - image_destroy(&dec->cur, dec->edged_width, dec->edged_height); - image_destroy(&dec->refn[0], dec->edged_width, dec->edged_height); - image_destroy(&dec->refn[1], dec->edged_width, dec->edged_height); - image_destroy(&dec->refn[2], dec->edged_width, dec->edged_height); - xvid_free(dec); - return XVID_ERR_MEMORY; - } - dec->mbs = xvid_malloc(sizeof(MACROBLOCK) * dec->mb_width * dec->mb_height, CACHE_LINE); @@ -153,7 +140,6 @@ image_destroy(&dec->refn[0], dec->edged_width, dec->edged_height); image_destroy(&dec->refn[1], dec->edged_width, dec->edged_height); image_destroy(&dec->refn[2], dec->edged_width, dec->edged_height); - image_destroy(&dec->refh, dec->edged_width, dec->edged_height); xvid_free(dec); return XVID_ERR_MEMORY; } @@ -171,7 +157,6 @@ image_destroy(&dec->refn[0], dec->edged_width, dec->edged_height); image_destroy(&dec->refn[1], dec->edged_width, dec->edged_height); image_destroy(&dec->refn[2], dec->edged_width, dec->edged_height); - image_destroy(&dec->refh, dec->edged_width, dec->edged_height); xvid_free(dec); return XVID_ERR_MEMORY; } @@ -197,7 +182,6 @@ image_destroy(&dec->refn[0], dec->edged_width, dec->edged_height); image_destroy(&dec->refn[1], dec->edged_width, dec->edged_height); image_destroy(&dec->refn[2], dec->edged_width, dec->edged_height); - image_destroy(&dec->refh, dec->edged_width, dec->edged_height); image_destroy(&dec->cur, dec->edged_width, dec->edged_height); xvid_free(dec); @@ -280,10 +264,8 @@ start_timer(); if (cbp & (1 << (5 - i))) // coded { - int direction = dec->alternate_vertical_scan ? - 2 : pMB->acpred_directions[i]; - - get_intra_block(bs, &block[i * 64], direction, start_coeff); + get_intra_block(bs, &block[i * 64], pMB->acpred_directions[i], + start_coeff); } stop_coding_timer(); @@ -370,30 +352,6 @@ uv_dx = (uv_dx & 3) ? (uv_dx >> 1) | 1 : uv_dx / 2; uv_dy = (uv_dy & 3) ? (uv_dy >> 1) | 1 : uv_dy / 2; - - start_timer(); - if(dec->quarterpel) { - interpolate16x16_quarterpel(dec->cur.y, dec->refn[0].y, dec->refh.y, dec->refh.y + 64, - dec->refh.y + 128, 16*x_pos, 16*y_pos, - pMB->mvs[0].x, pMB->mvs[0].y, stride, rounding); - } - else { - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos, - pMB->mvs[0].x, pMB->mvs[0].y, stride, rounding); - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos, - pMB->mvs[1].x, pMB->mvs[1].y, stride, rounding); - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos + 8, - pMB->mvs[2].x, pMB->mvs[2].y, stride, rounding); - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos + 8, - pMB->mvs[3].x, pMB->mvs[3].y, stride, rounding); - } - - interpolate8x8_switch(dec->cur.u, dec->refn[0].u, 8 * x_pos, 8 * y_pos, - uv_dx, uv_dy, stride2, rounding); - interpolate8x8_switch(dec->cur.v, dec->refn[0].v, 8 * x_pos, 8 * y_pos, - uv_dx, uv_dy, stride2, rounding); - stop_comp_timer(); - } else { int sum; sum = pMB->mvs[0].x + pMB->mvs[1].x + pMB->mvs[2].x + pMB->mvs[3].x; @@ -413,49 +371,44 @@ } uv_dy = (sum == 0 ? 0 : SIGN(sum) * (roundtab[ABS(sum) % 16] + (ABS(sum) / 16) * 2)); + } - start_timer(); - if(dec->quarterpel) { - interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, dec->refh.y, dec->refh.y + 64, - dec->refh.y + 128, 16*x_pos, 16*y_pos, - pMB->mvs[0].x, pMB->mvs[0].y, stride, rounding); - interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, dec->refh.y, dec->refh.y + 64, - dec->refh.y + 128, 16*x_pos + 8, 16*y_pos, - pMB->mvs[1].x, pMB->mvs[1].y, stride, rounding); - interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, dec->refh.y, dec->refh.y + 64, - dec->refh.y + 128, 16*x_pos, 16*y_pos + 8, - pMB->mvs[2].x, pMB->mvs[2].y, stride, rounding); - interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, dec->refh.y, dec->refh.y + 64, - dec->refh.y + 128, 16*x_pos + 8, 16*y_pos + 8, - pMB->mvs[3].x, pMB->mvs[3].y, stride, rounding); - } - else { - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos, + start_timer(); + if(dec->quarterpel) { + DPRINTF(DPRINTF_DEBUG, "QUARTERPEL\n"); + interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos, pMB->mvs[0].x, pMB->mvs[0].y, stride, rounding); - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos, + interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos, pMB->mvs[1].x, pMB->mvs[1].y, stride, rounding); - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos + 8, + interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos + 8, pMB->mvs[2].x, pMB->mvs[2].y, stride, rounding); - interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos + 8, + interpolate8x8_quarterpel(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos + 8, pMB->mvs[3].x, pMB->mvs[3].y, stride, rounding); - } - - interpolate8x8_switch(dec->cur.u, dec->refn[0].u, 8 * x_pos, 8 * y_pos, - uv_dx, uv_dy, stride2, rounding); - interpolate8x8_switch(dec->cur.v, dec->refn[0].v, 8 * x_pos, 8 * y_pos, - uv_dx, uv_dy, stride2, rounding); - stop_comp_timer(); + } + else { + interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos, + pMB->mvs[0].x, pMB->mvs[0].y, stride, rounding); + interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos, + pMB->mvs[1].x, pMB->mvs[1].y, stride, rounding); + interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos, 16*y_pos + 8, + pMB->mvs[2].x, pMB->mvs[2].y, stride, rounding); + interpolate8x8_switch(dec->cur.y, dec->refn[0].y, 16*x_pos + 8, 16*y_pos + 8, + pMB->mvs[3].x, pMB->mvs[3].y, stride, rounding); } - for (i = 0; i < 6; i++) { - int direction = dec->alternate_vertical_scan ? 2 : 0; + interpolate8x8_switch(dec->cur.u, dec->refn[0].u, 8 * x_pos, 8 * y_pos, + uv_dx, uv_dy, stride2, rounding); + interpolate8x8_switch(dec->cur.v, dec->refn[0].v, 8 * x_pos, 8 * y_pos, + uv_dx, uv_dy, stride2, rounding); + stop_comp_timer(); + for (i = 0; i < 6; i++) { if (cbp & (1 << (5 - i))) // coded { memset(&block[i * 64], 0, 64 * sizeof(int16_t)); // clear start_timer(); - get_inter_block(bs, &block[i * 64], direction); + get_inter_block(bs, &block[i * 64]); stop_coding_timer(); start_timer(); @@ -552,7 +505,7 @@ if (dec->interlacing) { mb->field_dct = BitstreamGetBit(bs); - DEBUG1("deci: field_dct: ", mb->field_dct); + DPRINTF(DPRINTF_DEBUG, "deci: field_dct: %d", mb->field_dct); } decoder_mbintra(dec, mb, x, y, acpred_flag, cbp, bs, quant, @@ -698,18 +651,18 @@ if (dec->interlacing) { if (cbp || intra) { mb->field_dct = BitstreamGetBit(bs); - DEBUG1("decp: field_dct: ", mb->field_dct); + DPRINTF(DPRINTF_DEBUG, "decp: field_dct: %d", mb->field_dct); } if (mb->mode == MODE_INTER || mb->mode == MODE_INTER_Q) { mb->field_pred = BitstreamGetBit(bs); - DEBUG1("decp: field_pred: ", mb->field_pred); + DPRINTF(DPRINTF_DEBUG, "decp: field_pred: %d", mb->field_pred); if (mb->field_pred) { mb->field_for_top = BitstreamGetBit(bs); - DEBUG1("decp: field_for_top: ", mb->field_for_top); + DPRINTF(DPRINTF_DEBUG, "decp: field_for_top: %d", mb->field_for_top); mb->field_for_bot = BitstreamGetBit(bs); - DEBUG1("decp: field_for_bot: ", mb->field_for_bot); + DPRINTF(DPRINTF_DEBUG, "decp: field_for_bot: %d", mb->field_for_bot); } } } @@ -749,7 +702,8 @@ rounding); } else // not coded { - DEBUG2("P-frame MB at (X,Y)=",x,y); + DPRINTF(DPRINTF_DEBUG, "P-frame MB at (X,Y)=(%d,%d)", x, y); + mb->mode = MODE_NOT_CODED; mb->mvs[0].x = mb->mvs[1].x = mb->mvs[2].x = mb->mvs[3].x = 0; mb->mvs[0].y = mb->mvs[1].y = mb->mvs[2].y = mb->mvs[3].y = 0; @@ -919,14 +873,12 @@ stop_comp_timer(); for (i = 0; i < 6; i++) { - int direction = dec->alternate_vertical_scan ? 2 : 0; - if (cbp & (1 << (5 - i))) // coded { memset(&block[i * 64], 0, 64 * sizeof(int16_t)); // clear start_timer(); - get_inter_block(bs, &block[i * 64], direction); + get_inter_block(bs, &block[i * 64]); stop_coding_timer(); start_timer(); @@ -1072,47 +1024,27 @@ interpolate8x8_switch(dec->refn[2].v, backward.v, 8 * x_pos, 8 * y_pos, b_uv_dx, b_uv_dy, stride2, 0); - interpolate8x8_avg2(dec->cur.y + (16 * y_pos * stride) + 16 * x_pos, - dec->cur.y + (16 * y_pos * stride) + 16 * x_pos, - dec->refn[2].y + (16 * y_pos * stride) + 16 * x_pos, - stride, 0); - - interpolate8x8_avg2(dec->cur.y + (16 * y_pos * stride) + 16 * x_pos + 8, - dec->cur.y + (16 * y_pos * stride) + 16 * x_pos + 8, - dec->refn[2].y + (16 * y_pos * stride) + 16 * x_pos + 8, - stride, 0); - - interpolate8x8_avg2(dec->cur.y + (16 * (y_pos + 8) * stride) + 16 * x_pos, - dec->cur.y + (16 * (y_pos + 8) * stride) + 16 * x_pos, - dec->refn[2].y + (16 * (y_pos + 8) * stride) + 16 * x_pos, - stride, 0); - - interpolate8x8_avg2(dec->cur.y + (16 * (y_pos + 8) * stride) + 16 * x_pos + 8, - dec->cur.y + (16 * (y_pos + 8) * stride) + 16 * x_pos + 8, - dec->refn[2].y + (16 * (y_pos + 8) * stride) + 16 * x_pos + 8, - stride, 0); - - interpolate8x8_avg2(dec->cur.u + (8 * y_pos * stride) + 8 * x_pos, - dec->cur.u + (8 * y_pos * stride) + 8 * x_pos, - dec->refn[2].u + (8 * y_pos * stride) + 8 * x_pos, - stride2, 0); - - interpolate8x8_avg2(dec->cur.v + (8 * y_pos * stride) + 8 * x_pos, - dec->cur.v + (8 * y_pos * stride) + 8 * x_pos, - dec->refn[2].v + (8 * y_pos * stride) + 8 * x_pos, - stride2, 0); - + interpolate8x8_c(dec->cur.y, dec->refn[2].y, 16 * x_pos, 16 * y_pos, + stride); + interpolate8x8_c(dec->cur.y, dec->refn[2].y, 16 * x_pos + 8, 16 * y_pos, + stride); + interpolate8x8_c(dec->cur.y, dec->refn[2].y, 16 * x_pos, 16 * y_pos + 8, + stride); + interpolate8x8_c(dec->cur.y, dec->refn[2].y, 16 * x_pos + 8, + 16 * y_pos + 8, stride); + interpolate8x8_c(dec->cur.u, dec->refn[2].u, 8 * x_pos, 8 * y_pos, + stride2); + interpolate8x8_c(dec->cur.v, dec->refn[2].v, 8 * x_pos, 8 * y_pos, + stride2); stop_comp_timer(); for (i = 0; i < 6; i++) { - int direction = dec->alternate_vertical_scan ? 2 : 0; - if (cbp & (1 << (5 - i))) // coded { memset(&block[i * 64], 0, 64 * sizeof(int16_t)); // clear start_timer(); - get_inter_block(bs, &block[i * 64], direction); + get_inter_block(bs, &block[i * 64]); stop_coding_timer(); start_timer(); @@ -1343,7 +1275,7 @@ break; default: - DEBUG1("Not support B-frame mb_type =", mb->mb_type); + DPRINTF(DPRINTF_ERROR, "Not support B-frame mb_type = %d", mb->mb_type); } } // end of FOR @@ -1401,24 +1333,24 @@ decoder_pframe(dec, &bs, rounding, quant, fcode_forward, intra_dc_threshold); #ifdef BFRAMES_DEC - DEBUG1("P_VOP Time=", dec->time); + DPRINTF(DPRINTF_DEBUG, "P_VOP Time=%d", dec->time); #endif break; case I_VOP: decoder_iframe(dec, &bs, quant, intra_dc_threshold); #ifdef BFRAMES_DEC - DEBUG1("I_VOP Time=", dec->time); + DPRINTF(DPRINTF_DEBUG, "I_VOP Time=%d", dec->time); #endif break; case B_VOP: #ifdef BFRAMES_DEC if (dec->time_pp > dec->time_bp) { - DEBUG1("B_VOP Time=", dec->time); + DPRINTF(DPRINTF_DEBUG, "B_VOP Time=%d", dec->time); decoder_bframe(dec, &bs, quant, fcode_forward, fcode_backward); } else { - DEBUG("broken B-frame!"); + DPRINTF(DPRINTF_DEBUG, "Broken B_VOP"); } #else image_copy(&dec->cur, &dec->refn[0], dec->edged_width, dec->height); @@ -1436,7 +1368,7 @@ #ifdef BFRAMES_DEC_DEBUG if (frame->length != BitstreamPos(&bs) / 8){ - DEBUG2("InLen/UseLen",frame->length, BitstreamPos(&bs) / 8); + DPRINTF(DPRINTF_DEBUG, "InLen: %d / UseLen: %d", frame->length, BitstreamPos(&bs) / 8); } #endif frame->length = BitstreamPos(&bs) / 8;