decoder.cpp

MPEG-4编解码的实现(包括MPEG4视音频编解码)

/*************************************************************************

This software module was originally developed by 

	Ming-Chieh Lee ([e-mail]), Microsoft Corporation
	Wei-ge Chen ([e-mail]), Microsoft Corporation
	Bruce Lin ([e-mail]), Microsoft Corporation
	Chuang Gu ([e-mail]), Microsoft Corporation
	(date: March, 1996)

and edited by

	Hiroyuki Katata ([e-mail]), Sharp Corporation
	Norio Ito ([e-mail]), Sharp Corporation
	Shuichi Watanabe ([e-mail]), Sharp Corporation
	(date: October, 1997)

and also edited by
	Dick van Smirren ([e-mail]), KPN Research
	Cor Quist ([e-mail]), KPN Research
	(date: July, 1998)

in the course of development of the MPEG-4 Video (ISO/IEC 14496-2). 
This software module is an implementation of a part of one or more MPEG-4 Video tools 
as specified by the MPEG-4 Video. 
ISO/IEC gives users of the MPEG-4 Video free license to this software module or modifications 
thereof for use in hardware or software products claiming conformance to the MPEG-4 Video. 
Those intending to use this software module in hardware or software products are advised that its use may infringe existing patents. 
The original developer of this software module and his/her company, 
the subsequent editors and their companies, 
and ISO/IEC have no liability for use of this software module or modifications thereof in an implementation. 
Copyright is not released for non MPEG-4 Video conforming products. 
Sharp retains full right to use the code for his/her own purpose, 
assign or donate the code to a third party and to inhibit third parties from using the code for non <MPEG standard> conforming products. 
This copyright notice must be included in all copies or derivative works. 

Copyright (c) 1997.

Module Name:

	tps_decoder.cpp

Abstract:

	caller for decoder for temporal scalability

Revision History:

*************************************************************************/

#include <stdio.h>
#include <stdlib.h>
#include <iostream.h>
#include <fstream.h>

#ifdef __PC_COMPILER_
#include <windows.h>
#include <mmsystem.h>
#endif // __PC_COMPILER_

#include "typeapi.h"
#include "mode.hpp"
#include "vopses.hpp"
#include "entropy/bitstrm.hpp"
#include "tps_enhcbuf.hpp"
#include "decoder/enhcbufdec.hpp"
#include "decoder/vopsedec.hpp"

#include "dataStruct.hpp"

#ifndef __GLOBAL_VAR_
#define __GLOBAL_VAR_
#endif

#include "global.hpp"

#ifdef __MFC_
#ifdef _DEBUG
#undef THIS_FILE
static char BASED_CODE THIS_FILE[] = __FILE__;
#endif
#define new DEBUG_NEW
#endif // __MFC_

Void dumpFrame (const CVOPU8YUVBA* pvopcQuant, FILE* pfYUV, FILE* pfSeg, AlphaUsage, CRct& rct, UInt nBits, Int iAlphaScale, Int DumpSkip);


#define MAX_TIME	99999			// for bitstreams stopped at BVOP
#define USAGE "Usage: \tDECODER bitstream_file [enhn_layer_bitstream_file] output_file width height\n\
or:    \tDECODER -vtc bitstream_file output_file target_spatial_layer target_snr_layer\n"

// Main Routine
int main(int argc, char* argv[])
{
	Bool bScalability = 0;
	Bool bWavelet = 0;

	Bool main_short_video_header; // Added by KPN for short headers [FDS]
	
	main_short_video_header = FALSE; 

	if(argc>1 && strcmp(argv[1],"-vtc")==0)
	{
		bWavelet = 1;
		if(argc != 6)
		{
			fprintf(stderr,USAGE);
			exit(1);
		}
	}
	else
	{
		if (argc < 5 || argc > 6) {
			fprintf (stderr,USAGE);
			exit(1);
		}
		else if (argc == 6)
			bScalability = 1;
	}


	FILE *pfBits;
	if ((pfBits = fopen (argv [bWavelet + 1], "rb")) == NULL )
		fatal_error("Bitstream File Not Found");
	fclose (pfBits);
	if (bScalability) {
		if ((pfBits = fopen (argv [2], "rb")) == NULL )
			fatal_error("Bitstream File for Enhancement Layer Not Found\n");
		fclose (pfBits);
	}


	if(bWavelet)
	{
	/////  WAVELET VTC: begin   /////////////////////////
		CVTCDecoder vtcdec;

		// start VTC decoding 
		// argv[1]: "-vtc"
		// argv[2]: bitstream file
		// argv[3]: decoded image file
		// argv[4]: target spatial layer
		// argv[5]: target SNR layer
		vtcdec.decode(argv[2], argv[3], atoi(argv[4]), atoi(argv[5]) );

		return 0;
	/////// WAVELET VTC: end   /////////////////////////
	}

	CRct rctDisplay       (0, 0, atoi (argv [3 + bScalability]), atoi (argv [4 + bScalability]));
/*Added */
	CRct rctDisplay_SSenh (0, 0, atoi (argv [3 + bScalability]), atoi (argv [4 + bScalability]));

	Bool bTemporalScalability = 0;

	if (bScalability){
		CVideoObjectDecoder* pvodec_tps[2];
		cout << "Checking scalability type...\n";
		pvodec_tps[BASE_LAYER] = new CVideoObjectDecoder (argv[1], rctDisplay.width, rctDisplay.height (), NULL, &main_short_video_header);
		pvodec_tps[ENHN_LAYER] = new CVideoObjectDecoder (argv[2], rctDisplay.width, rctDisplay.height (), NULL, &main_short_video_header);
		pvodec_tps[BASE_LAYER]->setClockRateScale( pvodec_tps[ENHN_LAYER] ); // added by Sharp (98/6/26)
		Time FirstBase = pvodec_tps[BASE_LAYER] -> senseTime ();
		Time FirstEnhn = pvodec_tps[ENHN_LAYER] -> senseTime ();

		if ( FirstBase != FirstEnhn ){
			bTemporalScalability = 1;
			cout << "Starting temporal scalability decoding...\n\n";
		}
		else{
			bTemporalScalability = 0;
			cout << "Starting spatial scalability decoding...\n\n";
		}
		
		delete pvodec_tps[BASE_LAYER];
		delete pvodec_tps[ENHN_LAYER];
	}

	if ( !bScalability || !bTemporalScalability ){
		CVideoObjectDecoder* pvodec [2];
		Bool bSpatialScalable = FALSE;
		//CRct rctDisplayBackup = rctDisplay; // for recovery from sprite decoding

		pvodec [BASE_LAYER] = new CVideoObjectDecoder (argv [1], rctDisplay.width, rctDisplay.height(), &bSpatialScalable, &main_short_video_header);
	
		// NBIT: get nBits information
		UInt nBits = pvodec [BASE_LAYER]->volmd ().nBits;

		/* Added*/
		if (bScalability){
			pvodec [ENHN_LAYER] = new CVideoObjectDecoder (argv [2], rctDisplay.width, rctDisplay.height(), &bSpatialScalable, &main_short_video_header);
			/*Added*/
			if(bSpatialScalable)
				pvodec[ENHN_LAYER]->set_enh_display_size(rctDisplay, &rctDisplay_SSenh);
		}
		if (pvodec [BASE_LAYER] -> fSptUsage () == 1) {
				// decode the initial sprite: //wchen: instead of calling decode ()
			pvodec [BASE_LAYER] -> decodeInitSprite ();
		}

		FILE* pfReconYUV [2];
		Char pchTmp [100];
		sprintf (pchTmp, "%s.yuv", argv [2 + bScalability]);
		pfReconYUV [BASE_LAYER] = fopen (pchTmp, "wb");
		fatal_error("cant open output yuv file",pfReconYUV [BASE_LAYER] != NULL);
		if (bScalability)	{
			sprintf (pchTmp, "%s_e.yuv", argv [2 + bScalability]);
			pfReconYUV [ENHN_LAYER]= fopen (pchTmp, "wb");
			fatal_error("cant open enhancement layer output yuv file",pfReconYUV [ENHN_LAYER] != NULL);
		}
		sprintf (pchTmp, "%s.seg", argv [2 + bScalability]);
		FILE* pfReconSeg = NULL;
		if (pvodec [BASE_LAYER]-> volmd ().fAUsage != RECTANGLE)	{
			pfReconSeg = fopen (pchTmp, "wb");
			fatal_error("cant open output seg file",pfReconSeg != NULL);
		}

		Int iEof = 1;
		Int nFrames = 0;
		const CVOPU8YUVBA* pvopcQuant;
#ifdef __PC_COMPILER_
		Int tickBegin = ::GetTickCount ();
#endif // __PC_COMPILER_
		Bool bCachedRefFrame = FALSE;
		Bool bCachedRefFrameCoded = FALSE;
		while (iEof != EOF)
		{
			if (main_short_video_header) // Added by KPN for short headers
			{
				fprintf(stderr,"Frame number: %d\n", nFrames);
				iEof = pvodec [BASE_LAYER] -> h263_decode ();
			}
			else
				iEof = pvodec [BASE_LAYER] -> decode ();  
			if (iEof != EOF)
				nFrames++;
			if(pvodec [BASE_LAYER] -> fSptUsage () == 1)
			{
				// sprite
				if(iEof != EOF)
				{
					pvopcQuant = pvodec [BASE_LAYER]->pvopcReconCurr();
					dumpFrame (pvopcQuant, pfReconYUV [BASE_LAYER], pfReconSeg, pvodec[BASE_LAYER]->volmd().fAUsage, rctDisplay,
						nBits, pvodec [BASE_LAYER]->vopmd().iVopConstantAlphaValue, 0);
				}
			}
			else
			{
				if(iEof == EOF)
				{
					// dump final cached frame if present
					if(bCachedRefFrame)
					{
						bCachedRefFrame = FALSE;
						if(bCachedRefFrameCoded)
						{
							pvopcQuant = pvodec [BASE_LAYER]->pvopcRefQLater(); // future reference
							dumpFrame (pvopcQuant, pfReconYUV [BASE_LAYER], pfReconSeg, pvodec[BASE_LAYER]->volmd().fAUsage, rctDisplay,
								nBits, pvodec [BASE_LAYER]->vopmd().iVopConstantAlphaValue, 0);
						}
						else // non coded
							dumpNonCodedFrame(pfReconYUV [BASE_LAYER], pfReconSeg, rctDisplay, nBits);
					}
				}
				else
				{
					// dump if bvop
					if(pvodec[BASE_LAYER]->vopmd().vopPredType == BVOP)
					{
						// BVOP
						if(iEof != FALSE)
						{
							pvopcQuant = pvodec [BASE_LAYER]->pvopcReconCurr(); // current vop
							dumpFrame (pvopcQuant, pfReconYUV [BASE_LAYER], pfReconSeg, pvodec[BASE_LAYER]->volmd().fAUsage, rctDisplay,
								nBits, pvodec [BASE_LAYER]->vopmd().iVopConstantAlphaValue, 0);
						}
						else // non coded BVOP
							dumpNonCodedFrame(pfReconYUV [BASE_LAYER], pfReconSeg, rctDisplay, nBits);
					}
					else
					{
						// not a BVOP, so dump any previous cached frame
						if(bCachedRefFrame)
						{
							bCachedRefFrame = FALSE;
							if(bCachedRefFrameCoded)
							{
								pvopcQuant = pvodec [BASE_LAYER]->pvopcRefQPrev(); // past reference
								dumpFrame (pvopcQuant, pfReconYUV [BASE_LAYER], pfReconSeg, pvodec[BASE_LAYER]->volmd().fAUsage, rctDisplay,
									nBits, pvodec [BASE_LAYER]->vopmd().iVopConstantAlphaValue, 0);
							}
							else // non coded
								dumpNonCodedFrame(pfReconYUV [BASE_LAYER], pfReconSeg, rctDisplay, nBits);
						}

						// cache current reference
						bCachedRefFrame = TRUE;
						bCachedRefFrameCoded = (iEof != FALSE);
					}
				}
			}

			if (bSpatialScalable == TRUE) {
				// if base layer is non coded, pvopcQuant=NULL, but enh layer should also be non coded.
				if(iEof == FALSE)
					pvopcQuant = NULL;
				else
					pvopcQuant = pvodec [BASE_LAYER]->pvopcReconCurr();
				iEof = pvodec [ENHN_LAYER] -> decode (pvopcQuant);
				if(iEof!=EOF)
					nFrames++; // include enhancement layer
				if(iEof==FALSE)
					dumpNonCodedFrame(pfReconYUV [ENHN_LAYER], NULL, rctDisplay_SSenh, nBits);
/*					dumpNonCodedFrame(pfReconYUV [ENHN_LAYER], NULL, rctDisplay * 2, nBits);*/

				if (iEof != EOF && iEof!=FALSE)
/*Added*/
					dumpFrame (pvodec [ENHN_LAYER]->pvopcReconCurr (), pfReconYUV [ENHN_LAYER], NULL, RECTANGLE, rctDisplay_SSenh,
						nBits, pvodec [BASE_LAYER]->vopmd().iVopConstantAlphaValue, 0); // still base layer
/*
					dumpFrame (pvodec [ENHN_LAYER]->pvopcReconCurr (), pfReconYUV [ENHN_LAYER], NULL, RECTANGLE, rctDisplay * 2,
						nBits, pvodec [BASE_LAYER]->vopmd().iVopConstantAlphaValue, 0); // still base layer
*/
			}
		}

#ifdef __PC_COMPILER_
		Int tickAfter = ::GetTickCount ();
		printf ("Total time: %d\n", tickAfter - tickBegin);
		Double dAverage = (Double) (tickAfter - tickBegin) / (Double) (nFrames);
		printf ("Total frames: %d\tAverage time: %.6lf\n", nFrames, dAverage);
		printf ("FPS %.6lf\n", 1000.0 / dAverage);
#endif // __PC_COMPILER_

		fclose (pfReconYUV [BASE_LAYER]);
		if (pvodec [BASE_LAYER]->volmd ().fAUsage != RECTANGLE)
			fclose (pfReconSeg);
		delete pvodec [BASE_LAYER];
		if (bScalability)	{
			fclose (pfReconYUV [ENHN_LAYER]);
			delete pvodec [ENHN_LAYER];
		}

	}
	else { // loop for temporal scalability
		CVideoObjectDecoder* pvodec[2];
		pvodec[BASE_LAYER] = new CVideoObjectDecoder (argv[1], rctDisplay.width, rctDisplay.height (), NULL, &main_short_video_header); // modified by Sharp (98/7/16)

		UInt nBits = pvodec [BASE_LAYER]->volmd ().nBits; // added by Sharp (98/11/11)
		if(bScalability){
			pvodec[ENHN_LAYER] = new CVideoObjectDecoder (argv[2], rctDisplay.width, rctDisplay.height (), NULL, &main_short_video_header); // modified by Sharp (98/7/16)
			// for back/forward shape
			pvodec[ENHN_LAYER]->rgpbfShape[0] = new CVideoObjectDecoder (rctDisplay.width, rctDisplay.height ());
			pvodec[ENHN_LAYER]->rgpbfShape[1] = new CVideoObjectDecoder (rctDisplay.width, rctDisplay.height ());
			pvodec[BASE_LAYER]->setClockRateScale( pvodec[ENHN_LAYER] );
			// copy pointers
			pvodec[ENHN_LAYER]->copyTobfShape ();
		}

		Char pchTmp [100];
		FILE* pfReconYUV [2]; // following two lines are swapped by Sharp (98/10/26)
#ifndef __OUT_ONE_FRAME_
		sprintf (pchTmp, "%s.yuv", argv [2 + bScalability]);
		pfReconYUV [BASE_LAYER] = fopen (pchTmp, "wb");
		fatal_error("cant open output yuv file",pfReconYUV [BASE_LAYER] != NULL);