decoder.c

mpeg4代码,比较具体

/**************************************************************************
 *
 *  XVID MPEG-4 VIDEO CODEC
 *  -  Decoder main module  -
 *
 *  Copyright(C) 2002 MinChen <[e-mail]>
 *               2002 Peter Ross <[e-mail]>
 *
 *  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.
 *
 *  This program 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 General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  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.
 *
 *  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.
 *
 *  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 1.46 2003/02/11 21:56:31 edgomez Exp $
 *
 *************************************************************************/

#include <stdlib.h>
#include <string.h>

#ifdef BFRAMES_DEC_DEBUG
	#define BFRAMES_DEC
#endif

#include "xvid.h"
#include "portab.h"

#include "decoder.h"
#include "bitstream/bitstream.h"
#include "bitstream/mbcoding.h"

#include "quant/quant_h263.h"
#include "quant/quant_mpeg4.h"
#include "dct/idct.h"
#include "dct/fdct.h"
#include "utils/mem_transfer.h"
#include "image/interpolate8x8.h"

#include "bitstream/mbcoding.h"
#include "prediction/mbprediction.h"
#include "utils/timer.h"
#include "utils/emms.h"

#include "image/image.h"
#include "image/colorspace.h"
#include "utils/mem_align.h"

int
decoder_create(XVID_DEC_PARAM * param)
{
	DECODER *dec;

	dec = xvid_malloc(sizeof(DECODER), CACHE_LINE);
	if (dec == NULL) {
		return XVID_ERR_MEMORY;
	}
	param->handle = dec;

	dec->width = param->width;
	dec->height = param->height;

	dec->mb_width = (dec->width + 15) / 16;
	dec->mb_height = (dec->height + 15) / 16;

	dec->edged_width = 16 * dec->mb_width + 2 * EDGE_SIZE;
	dec->edged_height = 16 * dec->mb_height + 2 * EDGE_SIZE;
	dec->low_delay = 0;

	if (image_create(&dec->cur, dec->edged_width, dec->edged_height)) {
		xvid_free(dec);
		return XVID_ERR_MEMORY;
	}

	if (image_create(&dec->refn[0], dec->edged_width, dec->edged_height)) {
		image_destroy(&dec->cur, dec->edged_width, dec->edged_height);
		xvid_free(dec);
		return XVID_ERR_MEMORY;
	}
	/* add by chenm001 <[e-mail]> */
	/* for support B-frame to reference last 2 frame */
	if (image_create(&dec->refn[1], 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);
		xvid_free(dec);
		return XVID_ERR_MEMORY;
	}
	if (image_create(&dec->refn[2], 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);
		xvid_free(dec);
		return XVID_ERR_MEMORY;
	}

	dec->mbs =
		xvid_malloc(sizeof(MACROBLOCK) * dec->mb_width * dec->mb_height,
					CACHE_LINE);
	if (dec->mbs == NULL) {
		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;
	}

	memset(dec->mbs, 0, sizeof(MACROBLOCK) * dec->mb_width * dec->mb_height);

	/* add by chenm001 <[e-mail]> */
	/* for skip MB flag */
	dec->last_mbs =
		xvid_malloc(sizeof(MACROBLOCK) * dec->mb_width * dec->mb_height,
					CACHE_LINE);
	if (dec->last_mbs == NULL) {
		xvid_free(dec->mbs);
		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;
	}

	memset(dec->last_mbs, 0, sizeof(MACROBLOCK) * dec->mb_width * dec->mb_height);

	init_timer();

	/* add by chenm001 <[e-mail]> */
	/* for support B-frame to save reference frame's time */
	dec->frames = -1;
	dec->time = dec->time_base = dec->last_time_base = 0;

	return XVID_ERR_OK;
}


int
decoder_destroy(DECODER * dec)
{
	xvid_free(dec->last_mbs);
	xvid_free(dec->mbs);
	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->cur, dec->edged_width, dec->edged_height);
	xvid_free(dec);

	write_timer();
	return XVID_ERR_OK;
}



static const int32_t dquant_table[4] = {
	-1, -2, 1, 2
};




/* decode an intra macroblock */

void
decoder_mbintra(DECODER * dec,
				MACROBLOCK * pMB,
				const uint32_t x_pos,
				const uint32_t y_pos,
				const uint32_t acpred_flag,
				const uint32_t cbp,
				Bitstream * bs,
				const uint32_t quant,
				const uint32_t intra_dc_threshold,
				const unsigned int bound)
{

	DECLARE_ALIGNED_MATRIX(block, 6, 64, int16_t, CACHE_LINE);
	DECLARE_ALIGNED_MATRIX(data, 6, 64, int16_t, CACHE_LINE);

	uint32_t stride = dec->edged_width;
	uint32_t stride2 = stride / 2;
	uint32_t next_block = stride * 8;
	uint32_t i;
	uint32_t iQuant = pMB->quant;
	uint8_t *pY_Cur, *pU_Cur, *pV_Cur;

	pY_Cur = dec->cur.y + (y_pos << 4) * stride + (x_pos << 4);
	pU_Cur = dec->cur.u + (y_pos << 3) * stride2 + (x_pos << 3);
	pV_Cur = dec->cur.v + (y_pos << 3) * stride2 + (x_pos << 3);

	memset(block, 0, 6 * 64 * sizeof(int16_t));	/* clear */

	for (i = 0; i < 6; i++) {
		uint32_t iDcScaler = get_dc_scaler(iQuant, i < 4);
		int16_t predictors[8];
		int start_coeff;

		start_timer();
		predict_acdc(dec->mbs, x_pos, y_pos, dec->mb_width, i, &block[i * 64],
					 iQuant, iDcScaler, predictors, bound);
		if (!acpred_flag) {
			pMB->acpred_directions[i] = 0;
		}
		stop_prediction_timer();

		if (quant < intra_dc_threshold) {
			int dc_size;
			int dc_dif;

			dc_size = i < 4 ? get_dc_size_lum(bs) : get_dc_size_chrom(bs);
			dc_dif = dc_size ? get_dc_dif(bs, dc_size) : 0;

			if (dc_size > 8) {
				BitstreamSkip(bs, 1);	/* marker */
			}

			block[i * 64 + 0] = dc_dif;
			start_coeff = 1;

			DPRINTF(DPRINTF_COEFF,"block[0] %i", dc_dif);
		} else {
			start_coeff = 0;
		}

		start_timer();
		if (cbp & (1 << (5 - i)))	/* coded */
		{
			get_intra_block(bs, &block[i * 64], pMB->acpred_directions[i],
							start_coeff);
		}
		stop_coding_timer();

		start_timer();
		add_acdc(pMB, i, &block[i * 64], iDcScaler, predictors);
		stop_prediction_timer();

		start_timer();
		if (dec->quant_type == 0) {
			dequant_intra(&data[i * 64], &block[i * 64], iQuant, iDcScaler);
		} else {
			dequant4_intra(&data[i * 64], &block[i * 64], iQuant, iDcScaler);
		}
		stop_iquant_timer();

		start_timer();
		idct(&data[i * 64]);
		stop_idct_timer();
	}

	if (dec->interlacing && pMB->field_dct) {
		next_block = stride;
		stride *= 2;
	}

	start_timer();
	transfer_16to8copy(pY_Cur, &data[0 * 64], stride);
	transfer_16to8copy(pY_Cur + 8, &data[1 * 64], stride);
	transfer_16to8copy(pY_Cur + next_block, &data[2 * 64], stride);
	transfer_16to8copy(pY_Cur + 8 + next_block, &data[3 * 64], stride);
	transfer_16to8copy(pU_Cur, &data[4 * 64], stride2);
	transfer_16to8copy(pV_Cur, &data[5 * 64], stride2);
	stop_transfer_timer();
}





#define SIGN(X) (((X)>0)?1:-1)
#define ABS(X) (((X)>0)?(X):-(X))
static const uint32_t roundtab[16] =
	{ 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2 };


/* decode an inter macroblock */

void
decoder_mbinter(DECODER * dec,
				const MACROBLOCK * pMB,
				const uint32_t x_pos,
				const uint32_t y_pos,
				const uint32_t acpred_flag,
				const uint32_t cbp,
				Bitstream * bs,
				const uint32_t quant,
				const uint32_t rounding)
{

	DECLARE_ALIGNED_MATRIX(block, 6, 64, int16_t, CACHE_LINE);
	DECLARE_ALIGNED_MATRIX(data, 6, 64, int16_t, CACHE_LINE);

	uint32_t stride = dec->edged_width;
	uint32_t stride2 = stride / 2;
	uint32_t next_block = stride * 8;
	uint32_t i;
	uint32_t iQuant = pMB->quant;
	uint8_t *pY_Cur, *pU_Cur, *pV_Cur;
	int uv_dx, uv_dy;

	pY_Cur = dec->cur.y + (y_pos << 4) * stride + (x_pos << 4);
	pU_Cur = dec->cur.u + (y_pos << 3) * stride2 + (x_pos << 3);
	pV_Cur = dec->cur.v + (y_pos << 3) * stride2 + (x_pos << 3);

	if (pMB->mode == MODE_INTER || pMB->mode == MODE_INTER_Q) {
		uv_dx = pMB->mvs[0].x;
		uv_dy = pMB->mvs[0].y;

		uv_dx = (uv_dx & 3) ? (uv_dx >> 1) | 1 : uv_dx / 2;
		uv_dy = (uv_dy & 3) ? (uv_dy >> 1) | 1 : uv_dy / 2;
	} else {
		int sum;
		sum = pMB->mvs[0].x + pMB->mvs[1].x + pMB->mvs[2].x + pMB->mvs[3].x;

		uv_dx = (sum == 0 ? 0 : SIGN(sum) * (roundtab[ABS(sum) % 16] + (ABS(sum) / 16) * 2));

		sum = pMB->mvs[0].y + pMB->mvs[1].y + pMB->mvs[2].y + pMB->mvs[3].y;

		uv_dy = (sum == 0 ? 0 : SIGN(sum) * (roundtab[ABS(sum) % 16] + (ABS(sum) / 16) * 2));
	}

	start_timer();
	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();

	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]);
			stop_coding_timer();

			start_timer();
			if (dec->quant_type == 0) {
				dequant_inter(&data[i * 64], &block[i * 64], iQuant);
			} else {
				dequant4_inter(&data[i * 64], &block[i * 64], iQuant);
			}
			stop_iquant_timer();

			start_timer();
			idct(&data[i * 64]);
			stop_idct_timer();
		}
	}

	if (dec->interlacing && pMB->field_dct) {
		next_block = stride;
		stride *= 2;
	}

	start_timer();
	if (cbp & 32)
		transfer_16to8add(pY_Cur, &data[0 * 64], stride);
	if (cbp & 16)
		transfer_16to8add(pY_Cur + 8, &data[1 * 64], stride);
	if (cbp & 8)
		transfer_16to8add(pY_Cur + next_block, &data[2 * 64], stride);