home *** CD-ROM | disk | FTP | other *** search
- /**********************************************************************
- Copyright (c) 1991 MPEG/audio software simulation group, All Rights Reserved
- musicin.c
- **********************************************************************/
- /**********************************************************************
- * MPEG/audio coding/decoding software, work in progress *
- * NOT for public distribution until verified and approved by the *
- * MPEG/audio committee. For further information, please contact *
- * Davis Pan, 508-493-2241, e-mail: pan@3d.enet.dec.com *
- * *
- * VERSION 4.0 *
- * changes made since last update: *
- * date programmers comment *
- * 3/01/91 Douglas Wong, start of version 1.1 records *
- * Davis Pan *
- * 3/06/91 Douglas Wong, rename: setup.h to endef.h *
- * removed extraneous variables *
- * 3/21/91 J.Georges Fritsch introduction of the bit-stream *
- * package. This package allows you *
- * to generate the bit-stream in a *
- * binary or ascii format *
- * 3/31/91 Bill Aspromonte replaced the read of the SB matrix *
- * by an "code generated" one *
- * 5/10/91 W. Joseph Carter Ported to Macintosh and Unix. *
- * Incorporated Jean-Georges Fritsch's *
- * "bitstream.c" package. *
- * Modified to strictly adhere to *
- * encoded bitstream specs, including *
- * "Berlin changes". *
- * Modified user interface dialog & code *
- * to accept any input & output *
- * filenames desired. Also added *
- * de-emphasis prompt and final bail-out *
- * opportunity before encoding. *
- * Added AIFF PCM sound file reading *
- * capability. *
- * Modified PCM sound file handling to *
- * process all incoming samples and fill *
- * out last encoded frame with zeros *
- * (silence) if needed. *
- * Located and fixed numerous software *
- * bugs and table data errors. *
- * 27jun91 dpwe (Aware Inc) Used new frame_params struct. *
- * Clear all automatic arrays. *
- * Changed some variable names, *
- * simplified some code. *
- * Track number of bits actually sent. *
- * Fixed padding slot, stereo bitrate *
- * Added joint-stereo : scales L+R. *
- * 6/12/91 Earle Jennings added fix for MS_DOS in obtain_param *
- * 6/13/91 Earle Jennings added stack length adjustment before *
- * main for MS_DOS *
- * 7/10/91 Earle Jennings conversion of all float to FLOAT *
- * port to MsDos from MacIntosh completed*
- * 8/ 8/91 Jens Spille Change for MS-C6.00 *
- * 8/22/91 Jens Spille new obtain_parameters() *
- *10/ 1/91 S.I. Sudharsanan, Ported to IBM AIX platform. *
- * Don H. Lee, *
- * Peter W. Farrett *
- *10/ 3/91 Don H. Lee implemented CRC-16 error protection *
- * newly introduced functions are *
- * I_CRC_calc, II_CRC_calc and encode_CRC*
- * Additions and revisions are marked *
- * with "dhl" for clarity *
- *11/11/91 Katherine Wang Documentation of code. *
- * (variables in documentation are *
- * surround by the # symbol, and an '*'*
- * denotes layer I or II versions) *
- * 2/11/92 W. Joseph Carter Ported new code to Macintosh. Most *
- * important fixes involved changing *
- * 16-bit ints to long or unsigned in *
- * bit alloc routines for quant of 65535 *
- * and passing proper function args. *
- * Removed "Other Joint Stereo" option *
- * and made bitrate be total channel *
- * bitrate, irrespective of the mode. *
- * Fixed many small bugs & reorganized. *
- * 2/25/92 Masahiro Iwadare made code cleaner and more consistent *
- * 8/07/92 Mike Coleman make exit() codes return error status *
- * made slight changes for portability *
- *19 aug 92 Soren H. Nielsen Changed MS-DOS file name extensions. *
- * 8/25/92 Shaun Astarabadi Replaced rint() function with explicit*
- * rounding for portability with MSDOS. *
- * 9/22/92 jddevine@aware.com Fixed _scale_factor_calc() calls. *
- *10/19/92 Masahiro Iwadare added info->mode and info->mode_ext *
- * updates for AIFF format files *
- * 3/10/93 Kevin Peterson In parse_args, only set non default *
- * bit rate if specified in arg list. *
- * Use return value from aiff_read_hdrs *
- * to fseek to start of sound data *
- * 7/26/93 Davis Pan fixed bug in printing info->mode_ext *
- * value for joint stereo condition *
- * 8/27/93 Seymour Shlien, Fixes in Unix and MSDOS ports, *
- * Daniel Lauzon, and *
- * Bill Truerniet *
- **********************************************************************/
-
- #include "common.h"
- #include "encoder.h"
-
- /* Global variable definitions for "musicin.c" */
-
- FILE *musicin;
- Bit_stream_struc bs;
- char *programName;
-
- /* Implementations */
-
-
- /************************************************************************
- *
- * parse_args
- *
- * PURPOSE: Sets encoding parameters to the specifications of the
- * command line. Default settings are used for parameters
- * not specified in the command line.
- *
- * SEMANTICS: The command line is parsed according to the following
- * syntax:
- *
- * -l is followed by the layer number
- * -m is followed by the mode
- * -p is followed by the psychoacoustic model number
- * -s is followed by the sampling rate
- * -b is followed by the total bitrate, irrespective of the mode
- * -d is followed by the emphasis flag
- * -c is followed by the copyright/no_copyright flag
- * -o is followed by the original/not_original flag
- * -e is followed by the error_protection on/off flag
- *
- * If the input file is in AIFF format, the sampling frequency is read
- * from the AIFF header.
- *
- * The input and output filenames are read into #inpath# and #outpath#.
- *
- ************************************************************************/
-
- void
- parse_args(argc, argv, fr_ps, psy, num_samples, inPath, outPath)
- int argc;
- char **argv;
- frame_params *fr_ps;
- int *psy;
- unsigned long *num_samples;
- char inPath[MAX_NAME_SIZE];
- char outPath[MAX_NAME_SIZE];
- {
- FLOAT srate;
- int brate;
- layer *info = fr_ps->header;
- int err = 0, i = 0;
- long samplerate;
- long soundPosition;
-
- /* preset defaults */
- inPath[0] = '\0'; outPath[0] = '\0';
- info->lay = DFLT_LAY;
- switch(DFLT_MOD) {
- case 's': info->mode = MPG_MD_STEREO; info->mode_ext = 0; break;
- case 'd': info->mode = MPG_MD_DUAL_CHANNEL; info->mode_ext=0; break;
- case 'j': info->mode = MPG_MD_JOINT_STEREO; break;
- case 'm': info->mode = MPG_MD_MONO; info->mode_ext = 0; break;
- default:
- fprintf(stderr, "%s: Bad mode dflt %c\n", programName, DFLT_MOD);
- abort();
- }
- *psy = DFLT_PSY;
- if((info->sampling_frequency = SmpFrqIndex((long)(1000*DFLT_SFQ))) < 0) {
- fprintf(stderr, "%s: bad sfrq default %.2f\n", programName, DFLT_SFQ);
- abort();
- }
- if((info->bitrate_index = BitrateIndex(info->lay, DFLT_BRT)) < 0) {
- fprintf(stderr, "%s: bad default bitrate %u\n", programName, DFLT_BRT);
- abort();
- }
- switch(DFLT_EMP) {
- case 'n': info->emphasis = 0; break;
- case '5': info->emphasis = 1; break;
- case 'c': info->emphasis = 3; break;
- default:
- fprintf(stderr, "%s: Bad emph dflt %c\n", programName, DFLT_EMP);
- abort();
- }
- info->copyright = 0; info->original = 0; info->error_protection = FALSE;
-
- /* process args */
- while(++i<argc && err == 0) {
- char c, *token, *arg, *nextArg;
- int argUsed;
-
- token = argv[i];
- if(*token++ == '-') {
- if(i+1 < argc) nextArg = argv[i+1];
- else nextArg = "";
- argUsed = 0;
- while(c = *token++) {
- if(*token /* NumericQ(token) */) arg = token;
- else arg = nextArg;
- switch(c) {
- case 'l': info->lay = atoi(arg); argUsed = 1;
- if(info->lay<1 || info->lay>2) {
- fprintf(stderr,"%s: -l layer must be 1 or 2, not %s\n",
- programName, arg);
- err = 1;
- }
- break;
- case 'm': argUsed = 1;
- if (*arg == 's')
- { info->mode = MPG_MD_STEREO; info->mode_ext = 0; }
- else if (*arg == 'd')
- { info->mode = MPG_MD_DUAL_CHANNEL; info->mode_ext=0; }
- else if (*arg == 'j')
- { info->mode = MPG_MD_JOINT_STEREO; }
- else if (*arg == 'm')
- { info->mode = MPG_MD_MONO; info->mode_ext = 0; }
- else {
- fprintf(stderr,"%s: -m mode must be s/d/j/m not %s\n",
- programName, arg);
- err = 1;
- }
- break;
- case 'p': *psy = atoi(arg); argUsed = 1;
- if(*psy<1 || *psy>2) {
- fprintf(stderr,"%s: -p model must be 1 or 2, not %s\n",
- programName, arg);
- err = 1;
- }
- break;
-
- case 's':
- argUsed = 1;
- srate = atof( arg );
- /* samplerate = rint( 1000.0 * srate ); $A */
- samplerate = (long) (( 1000.0 * srate ) + 0.5);
- if( (info->sampling_frequency = SmpFrqIndex((long) samplerate)) < 0 )
- err = 1;
- break;
-
- case 'b':
- argUsed = 1;
- brate = atoi(arg);
- if( (info->bitrate_index = BitrateIndex(info->lay, brate)) < 0)
- err=1;
- break;
- case 'd': argUsed = 1;
- if (*arg == 'n') info->emphasis = 0;
- else if (*arg == '5') info->emphasis = 1;
- else if (*arg == 'c') info->emphasis = 3;
- else {
- fprintf(stderr,"%s: -d emp must be n/5/c not %s\n",
- programName, arg);
- err = 1;
- }
- break;
- case 'c': info->copyright = 1; break;
- case 'o': info->original = 1; break;
- case 'e': info->error_protection = TRUE; break;
- default: fprintf(stderr,"%s: unrec option %c\n",
- programName, c);
- err = 1; break;
- }
- if(argUsed) {
- if(arg == token) token = ""; /* no more from token */
- else ++i; /* skip arg we used */
- arg = ""; argUsed = 0;
- }
- }
- }
- else {
- if(inPath[0] == '\0') strcpy(inPath, argv[i]);
- else if(outPath[0] == '\0') strcpy(outPath, argv[i]);
- else {
- fprintf(stderr,"%s: excess arg %s\n", programName, argv[i]);
- err = 1;
- }
- }
- }
-
- if(err || inPath[0] == '\0') usage(); /* never returns */
-
- if(outPath[0] == '\0') {
- strcpy(outPath, inPath);
- strcat(outPath, DFLT_EXT);
- }
-
- if ((musicin = fopen(inPath, "rb")) == NULL) {
- printf("Could not find \"%s\".\n", inPath);
- exit(1);
- }
-
- open_bit_stream_w(&bs, outPath, BUFFER_SIZE);
-
- if (fseek(musicin, 0, SEEK_SET) != 0) {
- printf("Could not seek to PCM sound data in \"%s\".\n", inPath);
- exit(1);
- }
-
- /* Declare sound file to have "infinite" number of samples. */
- *num_samples = MAX_U_32_NUM;
-
- }
-
-
-
- /************************************************************************
- *
- * main
- *
- * PURPOSE: MPEG I Encoder supporting layers 1 and 2, and
- * psychoacoustic models 1 (MUSICAM) and 2 (AT&T)
- *
- * SEMANTICS: One overlapping frame of audio of up to 2 channels are
- * processed at a time in the following order:
- * (associated routines are in parentheses)
- *
- * 1. Filter sliding window of data to get 32 subband
- * samples per channel.
- * (window_subband,filter_subband)
- *
- * 2. If joint stereo mode, combine left and right channels
- * for subbands above #jsbound#.
- * (*_combine_LR)
- *
- * 3. Calculate scalefactors for the frame, and if layer 2,
- * also calculate scalefactor select information.
- * (*_scale_factor_calc)
- *
- * 4. Calculate psychoacoustic masking levels using selected
- * psychoacoustic model.
- * (*_Psycho_One, psycho_anal)
- *
- * 5. Perform iterative bit allocation for subbands with low
- * mask_to_noise ratios using masking levels from step 4.
- * (*_main_bit_allocation)
- *
- * 6. If error protection flag is active, add redundancy for
- * error protection.
- * (*_CRC_calc)
- *
- * 7. Pack bit allocation, scalefactors, and scalefactor select
- * information (layer 2) onto bitstream.
- * (*_encode_bit_alloc,*_encode_scale,II_transmission_pattern)
- *
- * 8. Quantize subbands and pack them into bitstream
- * (*_subband_quantization, *_sample_encoding)
- *
- ************************************************************************/
-
- main(argc, argv)
- int argc;
- char **argv;
- {
- typedef double SBS[2][3][SCALE_BLOCK][SBLIMIT];
- SBS FAR *sb_sample;
- typedef double JSBS[3][SCALE_BLOCK][SBLIMIT];
- JSBS FAR *j_sample;
- typedef double IN[2][HAN_SIZE];
- IN FAR *win_que;
- typedef unsigned int SUB[2][3][SCALE_BLOCK][SBLIMIT];
- SUB FAR *subband;
-
- frame_params fr_ps;
- layer info;
- char original_file_name[MAX_NAME_SIZE];
- char encoded_file_name[MAX_NAME_SIZE];
- short FAR **win_buf;
- static short FAR buffer[2][1152];
- static unsigned int bit_alloc[2][SBLIMIT], scfsi[2][SBLIMIT];
- static unsigned int scalar[2][3][SBLIMIT], j_scale[3][SBLIMIT];
- static double FAR ltmin[2][SBLIMIT], lgmin[2][SBLIMIT], max_sc[2][SBLIMIT];
- FLOAT snr32[32];
- short sam[2][1056];
- int whole_SpF, extra_slot = 0;
- double avg_slots_per_frame, frac_SpF, slot_lag;
- int model, stereo, error_protection;
- static unsigned int crc;
- int i, j, k, adb;
- unsigned long bitsPerSlot, samplesPerFrame, frameNum = 0;
- unsigned long frameBits, sentBits = 0;
- unsigned long num_samples;
-
- /* Most large variables are declared dynamically to ensure
- compatibility with smaller machines */
-
- sb_sample = (SBS FAR *) mem_alloc(sizeof(SBS), "sb_sample");
- j_sample = (JSBS FAR *) mem_alloc(sizeof(JSBS), "j_sample");
- win_que = (IN FAR *) mem_alloc(sizeof(IN), "Win_que");
- subband = (SUB FAR *) mem_alloc(sizeof(SUB),"subband");
- win_buf = (short FAR **) mem_alloc(sizeof(short *)*2, "win_buf");
-
- /* clear buffers */
- memset((char *) buffer, 0, sizeof(buffer));
- memset((char *) bit_alloc, 0, sizeof(bit_alloc));
- memset((char *) scalar, 0, sizeof(scalar));
- memset((char *) j_scale, 0, sizeof(j_scale));
- memset((char *) scfsi, 0, sizeof(scfsi));
- memset((char *) ltmin, 0, sizeof(ltmin));
- memset((char *) lgmin, 0, sizeof(lgmin));
- memset((char *) max_sc, 0, sizeof(max_sc));
- memset((char *) snr32, 0, sizeof(snr32));
- memset((char *) sam, 0, sizeof(sam));
-
- fr_ps.header = &info;
- fr_ps.tab_num = -1; /* no table loaded */
- fr_ps.alloc = NULL;
- info.version = MPEG_AUDIO_ID;
-
- programName = argv[0];
- parse_args(argc, argv, &fr_ps, &model, &num_samples,
- original_file_name, encoded_file_name);
-
- hdr_to_frps(&fr_ps);
- stereo = fr_ps.stereo;
- error_protection = info.error_protection;
-
- if (info.lay == 1) { bitsPerSlot = 32; samplesPerFrame = 384; }
- else { bitsPerSlot = 8; samplesPerFrame = 1152; }
- /* Figure average number of 'slots' per frame. */
- /* Bitrate means TOTAL for both channels, not per side. */
- avg_slots_per_frame = ((double)samplesPerFrame /
- s_freq[info.sampling_frequency]) *
- ((double)bitrate[info.lay-1][info.bitrate_index] /
- (double)bitsPerSlot);
- whole_SpF = (int) avg_slots_per_frame;
- frac_SpF = avg_slots_per_frame - (double)whole_SpF;
- slot_lag = -frac_SpF;
-
- if (frac_SpF == 0) info.padding = 0;
-
- while (get_audio(musicin, buffer, num_samples, stereo, info.lay) > 0) {
-
- fprintf(stderr, "{%4lu}", frameNum++); fflush(stderr);
- win_buf[0] = &buffer[0][0];
- win_buf[1] = &buffer[1][0];
- if (frac_SpF != 0) {
- if (slot_lag > (frac_SpF-1.0) ) {
- slot_lag -= frac_SpF;
- extra_slot = 0;
- info.padding = 0;
- /* printf("No padding for this frame\n"); */
- }
- else {
- extra_slot = 1;
- info.padding = 1;
- slot_lag += (1-frac_SpF);
- /* printf("Padding for this frame\n"); */
- }
- }
- adb = (whole_SpF+extra_slot) * bitsPerSlot;
-
- switch (info.lay) {
-
- /***************************** Layer I **********************************/
-
- case 1 :
- for (j=0;j<SCALE_BLOCK;j++)
- for (k=0;k<stereo;k++) {
- window_subband(&win_buf[k], &(*win_que)[k][0], k);
- filter_subband(&(*win_que)[k][0], &(*sb_sample)[k][0][j][0]);
- }
-
- I_scale_factor_calc(*sb_sample, scalar, stereo);
- if(fr_ps.actual_mode == MPG_MD_JOINT_STEREO) {
- I_combine_LR(*sb_sample, *j_sample);
- I_scale_factor_calc(j_sample, &j_scale, 1);
- }
-
- put_scale(scalar, &fr_ps, max_sc);
-
- if (model == 1) I_Psycho_One(buffer, max_sc, ltmin, &fr_ps);
- else {
- for (k=0;k<stereo;k++) {
- psycho_anal(&buffer[k][0],&sam[k][0], k, info.lay, snr32,
- (FLOAT)s_freq[info.sampling_frequency]*1000);
- for (i=0;i<SBLIMIT;i++) ltmin[k][i] = (double) snr32[i];
- }
- }
-
- I_main_bit_allocation(ltmin, bit_alloc, &adb, &fr_ps);
-
- if (error_protection) I_CRC_calc(&fr_ps, bit_alloc, &crc);
-
- encode_info(&fr_ps, &bs);
-
- if (error_protection) encode_CRC(crc, &bs);
-
- I_encode_bit_alloc(bit_alloc, &fr_ps, &bs);
- I_encode_scale(scalar, bit_alloc, &fr_ps, &bs);
- I_subband_quantization(scalar, *sb_sample, j_scale, *j_sample,
- bit_alloc, *subband, &fr_ps);
- I_sample_encoding(*subband, bit_alloc, &fr_ps, &bs);
- for (i=0;i<adb;i++) put1bit(&bs, 0);
- break;
-
- /***************************** Layer 2 **********************************/
-
- case 2 :
- for (i=0;i<3;i++) for (j=0;j<SCALE_BLOCK;j++)
- for (k=0;k<stereo;k++) {
- window_subband(&win_buf[k], &(*win_que)[k][0], k);
- filter_subband(&(*win_que)[k][0], &(*sb_sample)[k][i][j][0]);
- }
-
- II_scale_factor_calc(*sb_sample, scalar, stereo, fr_ps.sblimit);
- pick_scale(scalar, &fr_ps, max_sc);
- if(fr_ps.actual_mode == MPG_MD_JOINT_STEREO) {
- II_combine_LR(*sb_sample, *j_sample, fr_ps.sblimit);
- II_scale_factor_calc(j_sample, &j_scale, 1, fr_ps.sblimit);
- } /* this way we calculate more mono than we need */
- /* but it is cheap */
-
- if (model == 1) II_Psycho_One(buffer, max_sc, ltmin, &fr_ps);
- else {
- for (k=0;k<stereo;k++) {
- psycho_anal(&buffer[k][0],&sam[k][0], k,
- info.lay, snr32,
- (FLOAT)s_freq[info.sampling_frequency]*1000);
- for (i=0;i<SBLIMIT;i++) ltmin[k][i] = (double) snr32[i];
- }
- }
-
- II_transmission_pattern(scalar, scfsi, &fr_ps);
- II_main_bit_allocation(ltmin, scfsi, bit_alloc, &adb, &fr_ps);
-
- if (error_protection)
- II_CRC_calc(&fr_ps, bit_alloc, scfsi, &crc);
-
- encode_info(&fr_ps, &bs);
-
- if (error_protection) encode_CRC(crc, &bs);
-
- II_encode_bit_alloc(bit_alloc, &fr_ps, &bs);
- II_encode_scale(bit_alloc, scfsi, scalar, &fr_ps, &bs);
- II_subband_quantization(scalar, *sb_sample, j_scale,
- *j_sample, bit_alloc, *subband, &fr_ps);
- II_sample_encoding(*subband, bit_alloc, &fr_ps, &bs);
- for (i=0;i<adb;i++) put1bit(&bs, 0);
- break;
-
- /***************************** Layer 3 **********************************/
-
- case 3 : break;
-
- }
-
- frameBits = sstell(&bs) - sentBits;
- if(frameBits%bitsPerSlot) /* a program failure */
- fprintf(stderr,"Sent %ld bits = %ld slots plus %ld\n",
- frameBits, frameBits/bitsPerSlot,
- frameBits%bitsPerSlot);
- sentBits += frameBits;
-
- }
-
- close_bit_stream_w(&bs);
-
- if (fclose(musicin) != 0){
- printf("Could not close \"%s\".\n", original_file_name);
- exit(2);
- }
-
- exit(0);
- }
-
- /************************************************************************
- *
- * usage
- *
- * PURPOSE: Writes command line syntax to the file specified by #stderr#
- *
- ************************************************************************/
-
- void usage() /* print syntax & exit */
- {
- fprintf(stderr, "MPEG-1 Audio encoder 0.01\n\n");
- fprintf(stderr, "Ported by Henrik Bjerregaard Pedersen, 14-May-1995\n\n");
- fprintf(stderr, "Usage: Encode [switches] <infile> <outfile>\n");
- fprintf(stderr, "Switches:\n");
- fprintf(stderr, " -m <mode> mode (s for stereo, m for mono)\n");
- fprintf(stderr, " -s <freq> input sample rate (32, 44.1 or 48)\n");
- fprintf(stderr, " -b <kbps> output bit rate in kbps\n");
- exit(1);
- }
-
-