summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rwxr-xr-xMakefile33
-rw-r--r--dteutils.cpp47
-rw-r--r--lzss.cpp361
-rw-r--r--lzss.h3
-rw-r--r--psxdev/Makefile10
-rw-r--r--psxdev/bs.c348
-rw-r--r--psxdev/bs.h91
-rw-r--r--psxdev/common.h47
-rw-r--r--psxdev/idctfst.c287
-rw-r--r--psxdev/jfdctint.c291
-rw-r--r--psxdev/table.h102
-rw-r--r--psxdev/vlc.c602
-rw-r--r--str-util.cpp178
-rw-r--r--yazedc.cpp159
14 files changed, 1799 insertions, 760 deletions
diff --git a/Makefile b/Makefile
index 7a41281..6f57c0e 100755
--- a/Makefile
+++ b/Makefile
@@ -3,30 +3,33 @@
CPPFLAGS=-Wall -g -O3 -mcpu=i686 -pedantic -pedantic-errors -Werror
CXX=g++
-TARGET = lzss dlzss yazedc cd-tool dte-tool str-tool
+SUBDIRS = psxdev
+TARGET = lzss dlzss cd-tool str-tool
-all: ${TARGET}
+all: subdirs ${TARGET}
-lzss: lzss.cpp lzss.h generic.cpp generic.h fileutils.h fileutils.cpp
- ${CXX} ${CPPFLAGS} ${LDFLAGS} -DLZSS_MAIN lzss.cpp generic.cpp fileutils.cpp -o lzss
+subdirs:
+ for d in ${SUBDIRS} ; do (cd $$d ; make all) ; done
+
+lzss: lzss.o lzss.h generic.o generic.h fileutils.h fileutils.o lzss-main.o
+ ${CXX} ${LDFLAGS} -DLZSS_MAIN lzss.o generic.o fileutils.o lzss-main.o -o lzss
dlzss: lzss
ln -fs lzss dlzss
-yazedc: yazedc.cpp crctables crctable.out
- ${CXX} ${CPPFLAGS} ${LDFLAGS} yazedc.cpp -DMAIN -o yazedc
-
-cd-tool: cd-tool.cpp cdutils.cpp cdutils.h fileutils.cpp fileutils.h generic.cpp generic.h yazedc.cpp yazedc.h
- ${CXX} ${CPPFLAGS} ${LDFLAGAS} cd-tool.cpp cdutils.cpp fileutils.cpp yazedc.cpp generic.cpp -o cd-tool
+yazedc: yazedc.o crctables crctable.out yazedc-main.o
+ ${CXX} ${CPPFLAGS} ${LDFLAGS} yazedc.o yazedc-main.o -DMAIN -o yazedc
-dte-tool-asm: dteutils.cpp generic.h generic.cpp fileutils.cpp fileutils.h dte-asm.S
- ${CXX} ${CPPFLAGS} ${LDFLAGS} dteutils.cpp generic.cpp fileutils.cpp dte-asm.S -o dte-tool-asm -DUSEASM -DDTEMAIN
+cd-tool: cd-tool.o cdutils.o cdutils.h fileutils.o fileutils.h generic.o generic.h yazedc.o yazedc.h
+ ${CXX} ${oFLAGS} ${LDFLAGAS} cd-tool.o cdutils.o fileutils.o yazedc.o generic.o -o cd-tool
-dte-tool: dteutils.cpp generic.h generic.cpp fileutils.cpp fileutils.h
- ${CXX} ${CPPFLAGS} ${LDFLAGS} dteutils.cpp generic.cpp fileutils.cpp -o dte-tool -DDTEMAIN
+dte-tool: dteutils.o generic.h generic.o fileutils.o fileutils.h dtemain.o
+ ${CXX} ${oFLAGS} ${LDFLAGS} dteutils.o generic.o fileutils.o dtemain.o -o dte-tool
-str-tool: str-util.cpp generic.h generic.cpp fileutils.cpp fileutils.h cdutils.cpp cdutils.h yazedc.cpp yazedc.h
- ${CXX} ${CPPFLAGS} ${LDFLAGS} str-util.cpp generic.cpp fileutils.cpp cdutils.cpp yazedc.cpp -o str-tool -DSTR_MAIN
+str-tool: str-util.o generic.h generic.o fileutils.o fileutils.h cdutils.o cdutils.h yazedc.o yazedc.h
+ ${CXX} ${oFLAGS} ${LDFLAGS} str-util.o generic.o fileutils.o cdutils.o yazedc.o psxdev/bs.o psxdev/idctfst.o psxdev/jfdctint.o psxdev/vlc.o -o str-tool
clean:
rm -f *.o ${TARGET} compil.c
+ for d in ${SUBDIRS} ; do (cd $$d ; make clean) ; done
+
diff --git a/dteutils.cpp b/dteutils.cpp
index 7c5057d..4ae8668 100644
--- a/dteutils.cpp
+++ b/dteutils.cpp
@@ -340,50 +340,3 @@ void read_thingy_file(FILE * f) {
dte_text[ptr] = dte_text[ptr + 1] = dte_text[ptr + 2] = dte_text[ptr + 3] = 0;
dte_text_size = ptr;
}
-
-#ifdef DTEMAIN
-
-int main(int argc, char ** argv) {
- FILE * f, * t;
- long old_size;
- int i;
-
- verbosity = M_INFO;
-
- printm(M_STATUS, "Reading thingy table\n");
- t = fopen(argv[2], "r");
- read_thingy(t);
- fclose(t);
-
- f = fopen(argv[1], "r");
- dte_text_size = filesize(f);
- dte_text = (char *) calloc(dte_text_size + 4, 1);
- printm(M_STATUS, "Reading file, size = %li\n", dte_text_size);
- read_thingy_file(f);
- fclose(f);
-
- printm(M_STATUS, "True size = %li\n", old_size = dte_text_size);
-
- printm(M_STATUS, "Compressing file.\n");
- dte_compress();
-
- printm(M_STATUS, "Rereading file.\n");
- f = fopen(argv[1], "r");
- dte_text_size = filesize(f);
- dte_text = (char *) calloc(dte_text_size + 4, 1);
- read_thingy_file(f);
- fclose(f);
-
- printm(M_STATUS, "True size = %li, real gain = %li\n", dte_text_size, old_size - dte_text_size);
-
- printm(M_INFO, "DTE Usage:\n");
- for (i = 0; i < 256; i++) {
- printm(M_INFO, "Entry %i ('%s') used at %i\n", i, things[i], dte_usage[i]);
- }
-
- printm(M_INFO, "Number of couples: %i\n", tnb_dte);
-
- free(dte_text);
-}
-
-#endif
diff --git a/lzss.cpp b/lzss.cpp
index 3e50095..4f9e424 100644
--- a/lzss.cpp
+++ b/lzss.cpp
@@ -436,364 +436,3 @@ void lzss_comp(FILE * f_source, FILE * f_cible, long * delta) {
free(c);
free(r);
}
-
-#ifdef LZSS_MAIN
-
-char * fn1, * fn2, * pname;
-
-int lga = 0;
-int compress = 1;
-
-struct option long_options[] = {
- {"1iscomp", 1, &lga, 1 },
- {"overlap", 1, &lga, 2 },
- {"16bits", 1, &lga, 11 },
- {"negative", 1, &lga, 12 },
- {"ptrb", 1, &lga, 13 },
- {"filling", 1, &lga, 14 },
- {"inverse", 1, &lga, 23 },
- {"onejump", 1, &lga, 24 },
- {"window", 1, &lga, 25 },
- {"lmask1", 1, &lga, 3 },
- {"lshft1", 1, &lga, 4 },
- {"lmask2", 1, &lga, 5 },
- {"lshft2", 1, &lga, 6 },
- {"jmask1", 1, &lga, 7 },
- {"jshft1", 1, &lga, 8 },
- {"jmask2", 1, &lga, 9 },
- {"jshft2", 1, &lga, 10 },
- {"fmask1", 1, &lga, 15 },
- {"fshft1", 1, &lga, 16 },
- {"fmask2", 1, &lga, 17 },
- {"fshft2", 1, &lga, 18 },
- {"vmask1", 1, &lga, 19 },
- {"vshft1", 1, &lga, 20 },
- {"vmask2", 1, &lga, 21 },
- {"vshft2", 1, &lga, 22 },
- {"help", 0, NULL, 'h'},
- {"scheme", 1, NULL, 's'},
- {"length", 1, NULL, 'l'},
- {"version", 0, NULL, 'V'},
- {"verbose", 0, NULL, 'v'},
- {"show", 0, NULL, 'S'},
- {"dump", 0, NULL, 'D'},
- {"compress", 0, NULL, 'c'},
- {"decompress", 0, NULL, 'd'},
- {"blocks", 0, NULL, 'b'},
- {0, 0, NULL, 0 }
-};
-
-void showhelp(void) {
- printm(M_BARE,
-"Usages:\n"
-" %s <--help|-h|-H|-?> Show this help and exit.\n"
-" %s <--dump> Show the built-in schemes and exit.\n"
-" %s <--version|-V> Show the version and copyrights and exit.\n"
-" %s [-c|-d] [-s <scheme>] [-l <length>] [-S] [-v] <infile> <outfile>\n"
-"\n"
-"-c --compress Compress <infile> to <outfile>\n"
-"-d --decompress Decompress <infile> to <outfile>\n"
-"-s <scheme> --scheme=<scheme> Loads the built-in scheme number <scheme>\n"
-"-l <length> --length=<length> Specify the true length for decompression,\n"
-" or enable the padding behavior for compression.\n"
-"-S --show Show the actual scheme before processing.\n"
-"-v --verbose Display a *LOT* of informations.\n"
-"-b --blocks Switch to blocks decompression behaviour\n"
-"\n"
-"Additionnaly you have the scheme manipulation options:\n"
-"--1iscomp --overlap --negative --16bits --ptrb\n"
-"--filling --inverse --onejump --window\n"
-"--lmask1 --lshft1 --lmask2 --lshft2 --jmask1 --jshft1 --jmask2 --jshft2\n"
-"--vmask1 --vshft1 --vmask2 --vshft2 --fmask1 --fshft1 --fmask2 --fshft2\n"
-"\n"
-"If you don't know what they are, forget them.\n"
-"\n"
-"Default behavior is to %scompress.\n"
-"\n", pname, pname, pname, pname, compress ? "" : "de");
-}
-
-void showscheme(void) {
- printm(M_BARE,
-"Actual scheme:\n"
-"--1iscomp %i\n"
-"--overlap %i\n"
-"--negative %i\n"
-"--16bits %i\n"
-"--ptrb %i\n"
-"--filling %i\n"
-"--inverse %i\n"
-"--onejump %i\n"
-"--window %i\n"
-"--lmask1 0x%02x\n"
-"--lshft1 %i\n"
-"--lmask2 0x%02x\n"
-"--lshft2 %i\n"
-"--jmask1 0x%02x\n"
-"--jshft1 %i\n"
-"--jmask2 0x%02x\n"
-"--jshft2 %i\n"
-"--fmask1 0x%02x\n"
-"--fshft1 %i\n"
-"--fmask2 0x%02x\n"
-"--fshft2 %i\n"
-"--vmask1 0x%02x\n"
-"--vshft1 %i\n"
-"--vmask2 0x%02x\n"
-"--vshft2 %i\n"
-"\n", scheme.one_is_compressed, scheme.overlap_trick, scheme.negative_trick, scheme.sixteen_bits,
-scheme.ptrb, scheme.filling, scheme.bitmap_inversed, scheme.one_jump, scheme.window_start,
-scheme.l_mask_1, scheme.l_shft_1, scheme.l_mask_2, scheme.l_shft_2,
-scheme.j_mask_1, scheme.j_shft_1, scheme.j_mask_2, scheme.j_shft_2,
-scheme.f_mask_1, scheme.f_shft_1, scheme.f_mask_2, scheme.f_shft_2,
-scheme.v_mask_1, scheme.v_shft_1, scheme.v_mask_2, scheme.v_shft_2);
-}
-
-void dump(void) {
- int i;
-
- printm(M_BARE, "Built-in schemes:\n");
- for (i = 0; schemes[i].name; i++) {
- printm(M_BARE, "%2i - %s\n", i, schemes[i].name);
- }
-}
-
-int main(int argc, char ** argv) {
- long length = -1;
- FILE * f1, * f2;
- int p, show = 0;
- int c, s, t;
-
- pname = strdup(argv[0]);
- p = strlen(pname) - 5;
-
- verbosity = M_STATUS;
-
- if (!strcasecmp(pname + p, "dlzss")) {
- compress = 0;
- }
-
- printm(M_BARE, /*
-LZSS_MAIN " compressor/decompressor version " LZSS_VERSION ",\n" */
-"Copyright (C) 2002 Nicolas \"Pixel\" Noble\n"
-"This software comes with ABSOLUTELY NO WARRANTY; see COPYING for details\n"
-"Thanks to Czar Dragon, for his little 'lzss' schemes FAQ.\n"
-"Special thanks to Yazoo, who taught me PSX hacking.\n"
-"\n");
-
- while ((c = getopt_long(argc, argv, "Hhs:l:vVScdb", long_options, NULL)) != EOF) {
- switch (c) {
- case 0:
- switch (lga) {
- case 1:
- t = atoi(optarg);
- if ((t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for boolean: %s\n", optarg);
- } else {
- scheme.one_is_compressed = t;
- }
- break;
- case 2:
- t = atoi(optarg);
- if ((t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for boolean: %s\n", optarg);
- } else {
- scheme.overlap_trick = t;
- }
- break;
- case 3:
- sscanf(optarg, "%i", &scheme.l_mask_1);
- break;
- case 4:
- scheme.l_shft_1 = atoi(optarg);
- break;
- case 5:
- sscanf(optarg, "%i", &scheme.l_mask_2);
- break;
- case 6:
- scheme.l_shft_2 = atoi(optarg);
- break;
- case 7:
- sscanf(optarg, "%i", &scheme.j_mask_1);
- break;
- case 8:
- scheme.j_shft_1 = atoi(optarg);
- break;
- case 9:
- sscanf(optarg, "%i", &scheme.j_mask_2);
- break;
- case 10:
- scheme.j_mask_2 = atoi(optarg);
- break;
- case 11:
- t = atoi(optarg);
- if ((t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for boolean: %s\n", optarg);
- } else {
- scheme.sixteen_bits = t;
- }
- break;
- case 12:
- t = atoi(optarg);
- if ((t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for boolean: %s\n", optarg);
- } else {
- scheme.negative_trick = t;
- }
- break;
- case 13:
- t = atoi(optarg);
- if ((t != 2) && (t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for ter: %s\n", optarg);
- } else {
- scheme.ptrb = t;
- }
- break;
- case 14:
- t = atoi(optarg);
- if ((t != 2) && (t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for ter: %s\n", optarg);
- } else {
- scheme.filling = t;
- }
- break;
- case 15:
- sscanf(optarg, "%i", &scheme.f_mask_1);
- break;
- case 16:
- scheme.f_shft_1 = atoi(optarg);
- break;
- case 17:
- sscanf(optarg, "%i", &scheme.f_mask_2);
- break;
- case 18:
- scheme.f_shft_2 = atoi(optarg);
- break;
- case 19:
- sscanf(optarg, "%i", &scheme.v_mask_1);
- break;
- case 20:
- scheme.v_shft_1 = atoi(optarg);
- break;
- case 21:
- sscanf(optarg, "%i", &scheme.v_mask_2);
- break;
- case 22:
- scheme.v_mask_2 = atoi(optarg);
- break;
- case 23:
- t = atoi(optarg);
- if ((t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for boolean: %s\n", optarg);
- } else {
- scheme.bitmap_inversed = t;
- }
- break;
- case 24:
- t = atoi(optarg);
- if ((t != 1) && (t != 0)) {
- printm(M_ERROR, "Invalid value for boolean: %s\n", optarg);
- } else {
- scheme.one_jump = t;
- }
- break;
- case 25:
- t = sscanf(optarg, "%i", &scheme.window_start);
- break;
- default:
- showhelp();
- printm(M_ERROR, "Unknow option.\n");
- exit(-1);
- }
- break;
- case '?':
- case 'H':
- case 'h':
- showhelp();
- exit(0);
- case 's':
- s = atoi(optarg);
- scheme = schemes[s];
- break;
- case 'l':
- length = atoi(optarg);
- break;
- case 'v':
- verbosity = M_INFO;
- break;
- case 'V':
- exit(0);
- case 'S':
- show = 1;
- break;
- case 'D':
- dump();
- exit(0);
- break;
- case 'd':
- compress = 0;
- break;
- case 'c':
- compress = 1;
- break;
- case 'b':
- blockb = 1;
- break;
- default:
- showhelp();
- printm(M_ERROR, "Unknow option.\n");
- exit(-1);
- }
- }
-
- if (show) showscheme();
-
- if (optind != (argc - 2)) {
- if (optind > (argc - 2)) {
- printm(M_ERROR, "Not enough filenames\n");
- exit(-1);
- } else {
- printm(M_ERROR, "Too much arguments\n");
- exit(-1);
- }
- }
-
- fn1 = argv[optind++];
- fn2 = argv[optind++];
-
- if (!(f1 = fopen(fn1, "r"))) {
- printm(M_ERROR, "Error opening file %s for reading.\n", fn1);
- }
-
- if (!(f2 = fopen(fn2, "w"))) {
- printm(M_ERROR, "Error opening file %s for writing.\n", fn2);
- }
-
- if (compress) {
- printm(M_STATUS, "Compressing `%s' to `%s'...\n", fn1, fn2);
- } else {
- printm(M_STATUS, "Decompressing `%s' to `%s'...\n", fn1, fn2);
- }
-
- if (compress) {
- if (length == -1) {
- lzss_comp(f1, f2);
- } else {
- lzss_comp(f1, f2, &length);
- }
- } else {
- length = lzss_decomp(f1, f2, length);
- }
-
- printm(M_STATUS, "Done, filesize changed from %i to %i.\n", filesize(f1), filesize(f2));
- if (!bitmap_count)
- bitmap_count = 8;
- if (compress)
- printm(M_STATUS, "Compressed %i = 0x%08x blocs, containing %i = 0x%08x chunks.\n", blk + 1, blk + 1, blk * 8 + bitmap_count, blk * 8 + bitmap_count);
-
- fclose(f1);
- fclose(f2);
-
- exit(0);
-}
-
-#endif
diff --git a/lzss.h b/lzss.h
index c0d05b8..a4fc71f 100644
--- a/lzss.h
+++ b/lzss.h
@@ -52,7 +52,8 @@ enum {
extern scheme_t scheme, schemes[];
-extern int tolerate;
+extern int tolerate, blockb;
+extern long blk, bitmap_count;
unsigned long lzss_decomp(FILE * f_source, FILE * f_cible, long true_length = -1);
void lzss_comp(FILE * f_source, FILE * f_cible, long * delta = NULL);
diff --git a/psxdev/Makefile b/psxdev/Makefile
new file mode 100644
index 0000000..3db2a0f
--- /dev/null
+++ b/psxdev/Makefile
@@ -0,0 +1,10 @@
+TARGETS=bs.o idctfst.o jfdctint.o vlc.o
+
+CC=gcc
+CPPFLAGS=-march=i386 -O3
+
+all: $(TARGETS)
+
+clean:
+ rm -f *.o
+
diff --git a/psxdev/bs.c b/psxdev/bs.c
new file mode 100644
index 0000000..5d7e185
--- /dev/null
+++ b/psxdev/bs.c
@@ -0,0 +1,348 @@
+/*
+ (c)2000 by BERO bero@geocities.co.jp
+
+ under GPL
+
+ some changes by dbalster@psxdev.de
+
+ - all globals now in a context (to use it as shlib)
+ - removed debugging printfs
+*/
+
+typedef struct {
+/* bit i/o */
+ unsigned int bitbuf;
+ int bitcount,bs_size,totalbit;
+ unsigned short *bsbuf;
+/* huffman */
+ int last_dc[3];
+ int _type;
+ int rlsize;
+ unsigned char *iqtab;
+} bs_context_t;
+
+#include "bs.h"
+#include "common.h"
+
+static const char *copyright = N_("Copyright (C) 2000 by Daniel Balster <dbalster@psxdev.de>");
+
+enum {B,G,R};
+typedef int BLOCK;
+
+#define DCTSIZE2 64
+#define RGB2Y(r,g,b) ( 0.299*(r) + 0.587*(g) + 0.114*(b) )
+#define RGB2Cb(r,g,b) ( -0.16874*(r) - 0.33126*(g) +0.5*(b) )
+#define RGB2Cr(r,g,b) ( 0.5*(r) - 0.41869*(g) - 0.08131*(b) )
+
+/*
+16x16 RGB -> 8x8 Cb,Cr,Y0,Y1,Y2,Y3
+
+[Y0][Y1] [Cb] [Cr]
+[Y2][Y3]
+*/
+#define Cb 0
+#define Cr DCTSIZE2
+
+static void rgb2yuv (unsigned char image[][3], BLOCK *blk)
+{
+ int x,y,i;
+ int tmpblk[16*16][3],(*yuv)[3];
+ BLOCK *yblk;
+
+ yuv=tmpblk;
+ for(i=0;i<16*16;i++) {
+ yuv[0][0] = RGB2Y (image[0][R],image[0][G],image[0][B])-128;
+ yuv[0][1] = RGB2Cb(image[0][R],image[0][G],image[0][B]);
+ yuv[0][2] = RGB2Cr(image[0][R],image[0][G],image[0][B]);
+ yuv++; image++;
+ }
+
+ yuv = tmpblk;
+ yblk = blk+DCTSIZE2*2;
+ for(y=0;y<16;y+=2,blk+=4,yblk+=8,yuv+=8+16) {
+ if (y==8) yblk+=DCTSIZE2;
+ for(x=0;x<4;x++,blk++,yblk+=2,yuv+=2) {
+ blk[Cb] = (yuv[0][1]+yuv[1][1]+yuv[16][1]+yuv[17][1])/4;
+ blk[Cr] = (yuv[0][2]+yuv[1][2]+yuv[16][2]+yuv[17][2])/4;
+ yblk[0] = yuv[ 0][0];
+ yblk[1] = yuv[ 1][0];
+ yblk[8] = yuv[16][0];
+ yblk[9] = yuv[17][0];
+
+ blk[4+Cb] = (yuv[8+0][1]+yuv[8+1][1]+yuv[8+16][1]+yuv[8+17][1])/4;
+ blk[4+Cr] = (yuv[8+0][2]+yuv[8+1][2]+yuv[8+16][2]+yuv[8+17][2])/4;
+ yblk[DCTSIZE2+0] = yuv[8+ 0][0];
+ yblk[DCTSIZE2+1] = yuv[8+ 1][0];
+ yblk[DCTSIZE2+8] = yuv[8+16][0];
+ yblk[DCTSIZE2+9] = yuv[8+17][0];
+ }
+ }
+}
+
+#undef Cb
+#undef Cr
+
+/* bit i/o */
+#define BITBUFSIZE 16
+#define WriteWord(x) ctxt->bsbuf[ctxt->bs_size++]=(x)
+
+static void putbits_init (bs_context_t *ctxt)
+{
+ ctxt->bitbuf = 0;
+ ctxt->bitcount = BITBUFSIZE;
+ ctxt->bs_size = 0;
+ ctxt->totalbit = 0;
+}
+
+static void putbits_flush (bs_context_t *ctxt)
+{
+ WriteWord(ctxt->bitbuf);
+}
+
+static void putbits (bs_context_t *ctxt, unsigned int x, unsigned int n)
+{
+ ctxt->totalbit+=n;
+
+ if (n<ctxt->bitcount) {
+ ctxt->bitcount-=n;
+ ctxt->bitbuf |= x << ctxt->bitcount;
+ } else {
+ n-=ctxt->bitcount;
+ WriteWord(ctxt->bitbuf | (x>>n) );
+ if (n<BITBUFSIZE) {
+ ctxt->bitcount = BITBUFSIZE-n;
+ } else {
+ WriteWord( x>>(n-BITBUFSIZE) );
+ ctxt->bitcount = BITBUFSIZE*2-n;
+ }
+ ctxt->bitbuf = x << ctxt->bitcount;
+ }
+}
+
+typedef struct {
+ unsigned char code,nbits;
+} huff_t;
+
+const static huff_t dc_y_table[] = {
+ {4,3},{0,2},{1,2},{5,3},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8}
+};
+
+const static huff_t dc_c_table[] = {
+ {0,2},{1,2},{2,2},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8},{510,9}
+};
+
+#include "table.h"
+
+static void encode_init (bs_context_t *ctxt, void *outbuf, int type, int q_scale)
+{
+ ctxt->_type = type;
+ ctxt->last_dc[0] = 0;
+ ctxt->last_dc[1] = 0;
+ ctxt->last_dc[2] = 0;
+ ctxt->rlsize = 0;
+ putbits_init(ctxt);
+
+ ctxt->bsbuf = outbuf;
+ ctxt->bsbuf[1] = 0x3800;
+ ctxt->bsbuf[2] = q_scale;
+ ctxt->bsbuf[3] = type;
+ ctxt->bs_size+=4;
+}
+
+static void encode_finish (bs_context_t *ctxt)
+{
+ putbits_flush(ctxt);
+ ctxt->bsbuf[0] = (((ctxt->rlsize+1)/2)+31)&~31;
+}
+
+static void encode_dc (bs_context_t *ctxt, int n, int level)
+{
+ if (ctxt->_type==2) {
+ putbits(ctxt,level&0x3ff,10);
+ } else {
+ const huff_t *table;
+ int prev,cnt;
+
+ level = level/4;
+ if (n<2) {
+ table = dc_c_table;
+ prev = ctxt->last_dc[n];
+ ctxt->last_dc[n] = level;
+ } else {
+ table = dc_y_table;
+ prev = ctxt->last_dc[2];
+ ctxt->last_dc[2] = level;
+ }
+ level -= prev;
+ if (level==0) cnt=0;
+ else {
+ int alevel = level;
+ if (alevel<0) alevel=-alevel;
+ for(cnt=8;(alevel>>cnt)==0;cnt--);
+ cnt++;
+ if (level<0) level--;
+ }
+ putbits(ctxt,table[cnt].code,table[cnt].nbits);
+ if (cnt) putbits(ctxt,level&((1<<cnt)-1),cnt);
+ }
+ ctxt->rlsize++;
+}
+
+static void encode_ac (bs_context_t *ctxt, int run, int level)
+{
+ int abslevel,sign;
+ if (level>0) {
+ abslevel = level;
+ sign = 0;
+ } else {
+ abslevel = -level;
+ sign = 1;
+ }
+ if (run<=31 && abslevel<=maxlevel[run]) {
+ putbits(ctxt,huff_table[run][abslevel-1].code+sign,huff_table[run][abslevel-1].nbits);
+ } else {
+ /* ESCAPE */
+ putbits(ctxt,1,6);
+ putbits(ctxt,(run<<10)+(level&0x3ff),16);
+ }
+ ctxt->rlsize++;
+}
+
+static void encode_eob (bs_context_t *ctxt)
+{
+ putbits(ctxt, 2,2);
+ ctxt->rlsize++;
+}
+
+extern DCT(BLOCK *blk);
+
+unsigned char zscan[DCTSIZE2] = {
+ 0 ,1 ,8 ,16,9 ,2 ,3 ,10,
+ 17,24,32,25,18,11,4 ,5 ,
+ 12,19,26,33,40,48,41,34,
+ 27,20,13,6 ,7 ,14,21,28,
+ 35,42,49,56,57,50,43,36,
+ 29,22,15,23,30,37,44,51,
+ 58,59,52,45,38,31,39,46,
+ 53,60,61,54,47,55,62,63
+};
+
+static unsigned char xxx_iqtab[DCTSIZE2] = {
+ 2,16,19,22,26,27,29,34,
+ 16,16,22,24,27,29,34,37,
+ 19,22,26,27,29,34,34,38,
+ 22,22,26,27,29,34,37,40,
+ 22,26,27,29,32,35,40,48,
+ 26,27,29,32,35,40,48,58,
+ 26,27,29,34,38,46,56,69,
+ 27,29,35,38,46,56,69,83
+};
+
+const unsigned char *bs_iqtab (void) { return xxx_iqtab; }
+
+static void blk2huff (bs_context_t *ctxt,BLOCK *blk,int q_scale)
+{
+ int i,k,run,level;
+ for(i=0;i<6;i++) {
+ DCT(blk);
+ for(k=0;k<DCTSIZE2;k++) blk[k]>>=3;
+ level = blk[0]/ctxt->iqtab[0];
+ encode_dc(ctxt,i,level);
+ run = 0;
+ for(k=1;k<64;k++) {
+ level = blk[zscan[k]]*8/(ctxt->iqtab[zscan[k]]*q_scale);
+ if (level==0) {
+ run++;
+ } else {
+ encode_ac(ctxt,run,level);
+ run=0;
+ }
+ }
+ encode_eob(ctxt);
+ blk+=DCTSIZE2;
+ }
+}
+
+u_char bs_roundtbl[256*3];
+
+void bs_init (void)
+{
+ int i;
+ for(i=0;i<256;i++) {
+ bs_roundtbl [i]=0;
+ bs_roundtbl [i+256]=i;
+ bs_roundtbl [i+512]=255;
+ }
+}
+
+int bs_encode (bs_header_t *outbuf,bs_input_image_t *img,int type,int q_scale,
+ unsigned char *myiqtab)
+{
+ unsigned char image[16][16][3];
+ BLOCK blk[6][DCTSIZE2];
+ bs_context_t *ctxt = malloc(sizeof(bs_context_t));
+
+ int x,y,xw,yw,rl;
+
+ ctxt->iqtab = myiqtab ? myiqtab : bs_iqtab();
+
+ encode_init (ctxt,outbuf,type,q_scale);
+
+ for(x=0;x<img->width;x+=16) {
+ xw = img->width-x; if (xw>16) xw = 16;
+ for(y=0;y<img->height;y+=16) {
+ char *p0 = img->top + x*(img->bit)/8 + y*img->nextline;
+ int i,j=0;
+ yw = img->height-y; if (yw>16) yw = 16;
+
+ /* get 16x16 image */
+
+ for(i=0;i<yw;i++) {
+ unsigned char *p = p0;
+ p0+=img->nextline;
+ switch(img->bit) {
+ case 16:
+ for(j=0;j<xw;j++) {
+ int c = *(unsigned short*)p;
+ image[i][j][B] = ((c>>10)&31)*8;
+ image[i][j][G] = ((c>>5)&31)*8;
+ image[i][j][R] = ((c&31))*8;
+ p+=2;
+ }
+ break;
+ case 24:
+ for(j=0;j<xw;j++) {
+ image[i][j][R] = p[R];
+ image[i][j][G] = p[G];
+ image[i][j][B] = p[B];
+ p+=3;
+ }
+ break;
+ }
+ for(;j<16;j++) {
+ image[i][j][R] = image[i][xw-1][R];
+ image[i][j][G] = image[i][xw-1][G];
+ image[i][j][B] = image[i][xw-1][B];
+ }
+ }
+
+ for(;i<16;i++) {
+ for(j=0;j<16;j++) {
+ image[i][j][R] = image[yw-1][j][R];
+ image[i][j][G] = image[yw-1][j][G];
+ image[i][j][B] = image[yw-1][j][B];
+ }
+ }
+
+ rgb2yuv((unsigned char*)image[0],blk[0]);
+ blk2huff(ctxt,blk[0],q_scale);
+ }
+ }
+
+ encode_finish(ctxt);
+
+ rl = (ctxt->bs_size * 2);
+ free (ctxt);
+
+ return rl;
+}
diff --git a/psxdev/bs.h b/psxdev/bs.h
new file mode 100644
index 0000000..d8d9a6c
--- /dev/null
+++ b/psxdev/bs.h
@@ -0,0 +1,91 @@
+/* $Id: bs.h,v 1.1 2002-06-21 23:45:51 Pixel Exp $ */
+
+/*
+ libbs - library for the bitstream image format
+
+ Copyright (C) 1999, 2000 by these people, who contributed to this project
+
+ bero@geocities.co.jp
+ Daniel Balster <dbalster@psxdev.de>
+
+ 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
+ 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., 675 Mass Ave, Cambridge, MA 02139, USA.
+*/
+
+/*
+ DCT code is based on Independent JPEG Group's sotfware
+*/
+
+#ifndef __LIB_BS_H
+#define __LIB_BS_H
+
+#define _GNU_SOURCE
+
+#include <sys/types.h>
+#include <stdarg.h>
+
+typedef struct {
+ int width,height;
+ int bit;
+ int nextline;
+ unsigned char *top,*lpbits;
+} bs_input_image_t;
+
+#define BS_MAGIC 0x3800
+#define BS_TYPE 2
+
+typedef struct {
+ u_short length;
+ u_short magic;
+ u_short q_scale;
+ u_short type;
+} bs_header_t;
+
+/* prototypes */
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+void bs_init (void);
+
+int bs_encode ( // returns BS image size in bytes
+ bs_header_t *outbuf, // output BS image
+ bs_input_image_t *img, // input image descriptor
+ int type, // image type (use BS_TYPE)
+ int q_scale, // Q scaling factor (1=best,>= lower quality)
+ unsigned char *myiqtab // provide own iqtab (NULL == default)
+ );
+
+void bs_decode_rgb24 (
+ unsigned char *outbuf, // output RGB bytes (width*height*3)
+ bs_header_t *img, // input BS image
+ int width, int height, // dimension of BS image
+ unsigned char *myiqtab
+ );
+
+void bs_decode_rgb15 (
+ unsigned short *outbuf, // output RGB bytes (width*height*2)
+ bs_header_t *img, // input BS image
+ int width, int height, // dimension of BS image
+ unsigned char *myiqtab
+ );
+
+const unsigned char *bs_iqtab (void);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* __LIB_BS_H */
diff --git a/psxdev/common.h b/psxdev/common.h
new file mode 100644
index 0000000..ee6e3a1
--- /dev/null
+++ b/psxdev/common.h
@@ -0,0 +1,47 @@
+/* $Id: common.h,v 1.1 2002-06-21 23:45:51 Pixel Exp $ */
+
+/*
+ common stuff
+
+ Copyright (C) 1997, 1998, 1999, 2000 by these people, who contributed to this project
+
+ Daniel Balster <dbalster@psxdev.de>
+
+ 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
+ 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., 675 Mass Ave, Cambridge, MA 02139, USA.
+*/
+
+#ifndef __COMMON_H
+#define __COMMON_H
+
+#define _GNU_SOURCE
+#define _USE_GNU
+
+#include <sys/types.h>
+
+#if ENABLE_NLS
+#if HAVE_LOCALE_H
+#include <locale.h>
+#endif
+#if HAVE_LIBINTL_H
+#include <libintl.h>
+#endif
+#define _(string) gettext(string)
+#define N_(string) (string)
+#else
+#define _(string) (string)
+#define N_(string) (string)
+#endif
+
+#endif
diff --git a/psxdev/idctfst.c b/psxdev/idctfst.c
new file mode 100644
index 0000000..22faae6
--- /dev/null
+++ b/psxdev/idctfst.c
@@ -0,0 +1,287 @@
+/*
+ * jidctfst.c
+ *
+ * Copyright (C) 1994-1996, Thomas G. Lane.
+ * This file is part of the Independent JPEG Group's software.
+ * For conditions of distribution and use, see the accompanying README file.
+ *
+ * This file contains a fast, not so accurate integer implementation of the
+ * inverse DCT (Discrete Cosine Transform). In the IJG code, this routine
+ * must also perform dequantization of the input coefficients.
+ *
+ * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT
+ * on each row (or vice versa, but it's more convenient to emit a row at
+ * a time). Direct algorithms are also available, but they are much more
+ * complex and seem not to be any faster when reduced to code.
+ *
+ * This implementation is based on Arai, Agui, and Nakajima's algorithm for
+ * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in
+ * Japanese, but the algorithm is described in the Pennebaker & Mitchell
+ * JPEG textbook (see REFERENCES section in file README). The following code
+ * is based directly on figure 4-8 in P&M.
+ * While an 8-point DCT cannot be done in less than 11 multiplies, it is
+ * possible to arrange the computation so that many of the multiplies are
+ * simple scalings of the final outputs. These multiplies can then be
+ * folded into the multiplications or divisions by the JPEG quantization
+ * table entries. The AA&N method leaves only 5 multiplies and 29 adds
+ * to be done in the DCT itself.
+ * The primary disadvantage of this method is that with fixed-point math,
+ * accuracy is lost due to imprecise representation of the scaled
+ * quantization values. The smaller the quantization table entry, the less
+ * precise the scaled value, so this implementation does worse with high-
+ * quality-setting files than with low-quality ones.
+ */
+
+/*
+ * This module is specialized to the case DCTSIZE = 8.
+ */
+
+/* Scaling decisions are generally the same as in the LL&M algorithm;
+ * see jidctint.c for more details. However, we choose to descale
+ * (right shift) multiplication products as soon as they are formed,
+ * rather than carrying additional fractional bits into subsequent additions.
+ * This compromises accuracy slightly, but it lets us save a few shifts.
+ * More importantly, 16-bit arithmetic is then adequate (for 8-bit samples)
+ * everywhere except in the multiplications proper; this saves a good deal
+ * of work on 16-bit-int machines.
+ *
+ * The dequantized coefficients are not integers because the AA&N scaling
+ * factors have been incorporated. We represent them scaled up by PASS1_BITS,
+ * so that the first and second IDCT rounds have the same input scaling.
+ * For 8-bit JSAMPLEs, we choose IFAST_SCALE_BITS = PASS1_BITS so as to
+ * avoid a descaling shift; this compromises accuracy rather drastically
+ * for small quantization table entries, but it saves a lot of shifts.
+ * For 12-bit JSAMPLEs, there's no hope of using 16x16 multiplies anyway,
+ * so we use a much larger scaling factor to preserve accuracy.
+ *
+ * A final compromise is to represent the multiplicative constants to only
+ * 8 fractional bits, rather than 13. This saves some shifting work on some
+ * machines, and may also reduce the cost of multiplication (since there
+ * are fewer one-bits in the constants).
+ */
+
+#define BITS_IN_JSAMPLE 8
+
+#if BITS_IN_JSAMPLE == 8
+#define CONST_BITS 8
+#define PASS1_BITS 2
+#else
+#define CONST_BITS 8
+#define PASS1_BITS 1 /* lose a little precision to avoid overflow */
+#endif
+
+/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus
+ * causing a lot of useless floating-point operations at run time.
+ * To get around this we use the following pre-calculated constants.
+ * If you change CONST_BITS you may want to add appropriate values.
+ * (With a reasonable C compiler, you can just rely on the FIX() macro...)
+ */
+
+#if CONST_BITS == 8
+#define FIX_1_082392200 (277) /* FIX(1.082392200) */
+#define FIX_1_414213562 (362) /* FIX(1.414213562) */
+#define FIX_1_847759065 (473) /* FIX(1.847759065) */
+#define FIX_2_613125930 (669) /* FIX(2.613125930) */
+#else
+#define FIX_1_082392200 FIX(1.082392200)
+#define FIX_1_414213562 FIX(1.414213562)
+#define FIX_1_847759065 FIX(1.847759065)
+#define FIX_2_613125930 FIX(2.613125930)
+#endif
+
+
+/* We can gain a little more speed, with a further compromise in accuracy,
+ * by omitting the addition in a descaling shift. This yields an incorrectly
+ * rounded result half the time...
+ */
+
+
+/* Multiply a DCTELEM variable by an INT32 constant, and immediately
+ * descale to yield a DCTELEM result.
+ */
+
+#define MULTIPLY(var,const) (DESCALE((var) * (const), CONST_BITS))
+
+
+/* Dequantize a coefficient by multiplying it by the multiplier-table
+ * entry; produce a DCTELEM result. For 8-bit data a 16x16->16
+ * multiplication will do. For 12-bit data, the multiplier table is
+ * declared INT32, so a 32-bit multiply will be used.
+ */
+
+#if BITS_IN_JSAMPLE == 8
+#define DEQUANTIZE(coef,quantval) (coef)
+#else
+#define DEQUANTIZE(coef,quantval) \
+ DESCALE((coef), IFAST_SCALE_BITS-PASS1_BITS)
+#endif
+
+
+/* Like DESCALE, but applies to a DCTELEM and produces an int.
+ * We assume that int right shift is unsigned if INT32 right shift is.
+ */
+
+#define DESCALE(x,n) ((x)>>(n))
+#define RANGE(n) (n)
+#define BLOCK int
+
+/*
+ * Perform dequantization and inverse DCT on one block of coefficients.
+ */
+#define DCTSIZE 8
+#define DCTSIZE2 64
+
+static void IDCT1(BLOCK *block)
+{
+ int val = RANGE(DESCALE(block[0], PASS1_BITS+3));
+ int i;
+ for(i=0;i<DCTSIZE2;i++) block[i]=val;
+}
+
+void IDCT(BLOCK *block,int k)
+{
+ int tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
+ int z5, z10, z11, z12, z13;
+ BLOCK *ptr;
+ int i;
+
+ /* Pass 1: process columns from input, store into work array. */
+ switch(k){
+ case 1:IDCT1(block); return;
+ }
+
+ ptr = block;
+ for (i = 0; i< DCTSIZE; i++,ptr++) {
+ /* Due to quantization, we will usually find that many of the input
+ * coefficients are zero, especially the AC terms. We can exploit this
+ * by short-circuiting the IDCT calculation for any column in which all
+ * the AC terms are zero. In that case each output is equal to the
+ * DC coefficient (with scale factor as needed).
+ * With typical images and quantization tables, half or more of the
+ * column DCT calculations can be simplified this way.
+ */
+
+ if ((ptr[DCTSIZE*1] | ptr[DCTSIZE*2] | ptr[DCTSIZE*3] |
+ ptr[DCTSIZE*4] | ptr[DCTSIZE*5] | ptr[DCTSIZE*6] |
+ ptr[DCTSIZE*7]) == 0) {
+ /* AC terms all zero */
+ ptr[DCTSIZE*0] =
+ ptr[DCTSIZE*1] =
+ ptr[DCTSIZE*2] =
+ ptr[DCTSIZE*3] =
+ ptr[DCTSIZE*4] =
+ ptr[DCTSIZE*5] =
+ ptr[DCTSIZE*6] =
+ ptr[DCTSIZE*7] =
+ ptr[DCTSIZE*0];
+
+ continue;
+ }
+
+ /* Even part */
+
+ z10 = ptr[DCTSIZE*0] + ptr[DCTSIZE*4]; /* phase 3 */
+ z11 = ptr[DCTSIZE*0] - ptr[DCTSIZE*4];
+ z13 = ptr[DCTSIZE*2] + ptr[DCTSIZE*6]; /* phases 5-3 */
+ z12 = MULTIPLY(ptr[DCTSIZE*2] - ptr[DCTSIZE*6], FIX_1_414213562) - z13; /* 2*c4 */
+
+ tmp0 = z10 + z13; /* phase 2 */
+ tmp3 = z10 - z13;
+ tmp1 = z11 + z12;
+ tmp2 = z11 - z12;
+
+ /* Odd part */
+
+ z13 = ptr[DCTSIZE*3] + ptr[DCTSIZE*5]; /* phase 6 */
+ z10 = ptr[DCTSIZE*3] - ptr[DCTSIZE*5];
+ z11 = ptr[DCTSIZE*1] + ptr[DCTSIZE*7];
+ z12 = ptr[DCTSIZE*1] - ptr[DCTSIZE*7];
+
+ z5 = MULTIPLY(z12 - z10, FIX_1_847759065);
+ tmp7 = z11 + z13; /* phase 5 */
+ tmp6 = MULTIPLY(z10, FIX_2_613125930) + z5 - tmp7; /* phase 2 */
+ tmp5 = MULTIPLY(z11 - z13, FIX_1_414213562) - tmp6;
+ tmp4 = MULTIPLY(z12, FIX_1_082392200) - z5 + tmp5;
+
+ ptr[DCTSIZE*0] = (tmp0 + tmp7);
+ ptr[DCTSIZE*7] = (tmp0 - tmp7);
+ ptr[DCTSIZE*1] = (tmp1 + tmp6);
+ ptr[DCTSIZE*6] = (tmp1 - tmp6);
+ ptr[DCTSIZE*2] = (tmp2 + tmp5);
+ ptr[DCTSIZE*5] = (tmp2 - tmp5);
+ ptr[DCTSIZE*4] = (tmp3 + tmp4);
+ ptr[DCTSIZE*3] = (tmp3 - tmp4);
+
+ }
+
+ /* Pass 2: process rows from work array, store into output array. */
+ /* Note that we must descale the results by a factor of 8 == 2**3, */
+ /* and also undo the PASS1_BITS scaling. */
+
+ ptr = block;
+ for (i = 0; i < DCTSIZE; i++ ,ptr+=DCTSIZE) {
+ /* Rows of zeroes can be exploited in the same way as we did with columns.
+ * However, the column calculation has created many nonzero AC terms, so
+ * the simplification applies less often (typically 5% to 10% of the time).
+ * On machines with very fast multiplication, it's possible that the
+ * test takes more time than it's worth. In that case this section
+ * may be commented out.
+ */
+
+#ifndef NO_ZERO_ROW_TEST
+ if ((ptr[1] | ptr[2] | ptr[3] | ptr[4] | ptr[5] | ptr[6] |
+ ptr[7]) == 0) {
+ /* AC terms all zero */
+ ptr[0] =
+ ptr[1] =
+ ptr[2] =
+ ptr[3] =
+ ptr[4] =
+ ptr[5] =
+ ptr[6] =
+ ptr[7] =
+ RANGE(DESCALE(ptr[0], PASS1_BITS+3));;
+
+ continue;
+ }
+#endif
+
+ /* Even part */
+
+ z10 = ptr[0] + ptr[4];
+ z11 = ptr[0] - ptr[4];
+ z13 = ptr[2] + ptr[6];
+ z12 = MULTIPLY(ptr[2] - ptr[6], FIX_1_414213562) - z13;
+
+ tmp0 = z10 + z13;
+ tmp3 = z10 - z13;
+ tmp1 = z11 + z12;
+ tmp2 = z11 - z12;
+
+ /* Odd part */
+
+ z13 = ptr[3] + ptr[5];
+ z10 = ptr[3] - ptr[5];
+ z11 = ptr[1] + ptr[7];
+ z12 = ptr[1] - ptr[7];
+
+ z5 = MULTIPLY(z12 - z10, FIX_1_847759065);
+ tmp7 = z11 + z13; /* phase 5 */
+ tmp6 = MULTIPLY(z10, FIX_2_613125930) + z5 - tmp7; /* phase 2 */
+ tmp5 = MULTIPLY(z11 - z13, FIX_1_414213562) - tmp6;
+ tmp4 = MULTIPLY(z12, FIX_1_082392200) - z5 + tmp5;
+
+ /* Final output stage: scale down by a factor of 8 and range-limit */
+
+ ptr[0] = RANGE(DESCALE(tmp0 + tmp7, PASS1_BITS+3));;
+ ptr[7] = RANGE(DESCALE(tmp0 - tmp7, PASS1_BITS+3));;
+ ptr[1] = RANGE(DESCALE(tmp1 + tmp6, PASS1_BITS+3));;
+ ptr[6] = RANGE(DESCALE(tmp1 - tmp6, PASS1_BITS+3));;
+ ptr[2] = RANGE(DESCALE(tmp2 + tmp5, PASS1_BITS+3));;
+ ptr[5] = RANGE(DESCALE(tmp2 - tmp5, PASS1_BITS+3));;
+ ptr[4] = RANGE(DESCALE(tmp3 + tmp4, PASS1_BITS+3));;
+ ptr[3] = RANGE(DESCALE(tmp3 - tmp4, PASS1_BITS+3));;
+
+ }
+}
+
diff --git a/psxdev/jfdctint.c b/psxdev/jfdctint.c
new file mode 100644
index 0000000..58bc2e4
--- /dev/null
+++ b/psxdev/jfdctint.c
@@ -0,0 +1,291 @@
+/*
+ * jfdctint.c
+ *
+ * Copyright (C) 1991-1994, Thomas G. Lane.
+ * This file is part of the Independent JPEG Group's software.
+ * For conditions of distribution and use, see the accompanying README file.
+ *
+ * This file contains a slow-but-accurate integer implementation of the
+ * forward DCT (Discrete Cosine Transform).
+ *
+ * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT
+ * on each column. Direct algorithms are also available, but they are
+ * much more complex and seem not to be any faster when reduced to code.
+ *
+ * This implementation is based on an algorithm described in
+ * C. Loeffler, A. Ligtenberg and G. Moschytz, "Practical Fast 1-D DCT
+ * Algorithms with 11 Multiplications", Proc. Int'l. Conf. on Acoustics,
+ * Speech, and Signal Processing 1989 (ICASSP '89), pp. 988-991.
+ * The primary algorithm described there uses 11 multiplies and 29 adds.
+ * We use their alternate method with 12 multiplies and 32 adds.
+ * The advantage of this method is that no data path contains more than one
+ * multiplication; this allows a very simple and accurate implementation in
+ * scaled fixed-point arithmetic, with a minimal number of shifts.
+ */
+
+#define DCT_ISLOW_SUPPORTED
+#define DCTSIZE 8
+#define DCTELEM int
+#define INT32 int
+#define DESCALE(x,n) RIGHT_SHIFT((x) + (1 << ((n)-1)), n)
+#define RIGHT_SHIFT(x,n) ((x)>>(n))
+#define GLOBAL
+#define jpeg_fdct_islow DCT
+#define SHIFT_TEMPS
+//#define BITS_IN_JSAMPLE 8
+//#define MULTIPLY16C16(var,const) ((var) * (const))
+
+
+#ifdef DCT_ISLOW_SUPPORTED
+
+
+/*
+ * This module is specialized to the case DCTSIZE = 8.
+ */
+
+#if DCTSIZE != 8
+ Sorry, this code only copes with 8x8 DCTs. /* deliberate syntax err */
+#endif
+
+
+/*
+ * The poop on this scaling stuff is as follows:
+ *
+ * Each 1-D DCT step produces outputs which are a factor of sqrt(N)
+ * larger than the true DCT outputs. The final outputs are therefore
+ * a factor of N larger than desired; since N=8 this can be cured by
+ * a simple right shift at the end of the algorithm. The advantage of
+ * this arrangement is that we save two multiplications per 1-D DCT,
+ * because the y0 and y4 outputs need not be divided by sqrt(N).
+ * In the IJG code, this factor of 8 is removed by the quantization step
+ * (in jcdctmgr.c), NOT in this module.
+ *
+ * We have to do addition and subtraction of the integer inputs, which
+ * is no problem, and multiplication by fractional constants, which is
+ * a problem to do in integer arithmetic. We multiply all the constants
+ * by CONST_SCALE and convert them to integer constants (thus retaining
+ * CONST_BITS bits of precision in the constants). After doing a
+ * multiplication we have to divide the product by CONST_SCALE, with proper
+ * rounding, to produce the correct output. This division can be done
+ * cheaply as a right shift of CONST_BITS bits. We postpone shifting
+ * as long as possible so that partial sums can be added together with
+ * full fractional precision.
+ *
+ * The outputs of the first pass are scaled up by PASS1_BITS bits so that
+ * they are represented to better-than-integral precision. These outputs
+ * require BITS_IN_JSAMPLE + PASS1_BITS + 3 bits; this fits in a 16-bit word
+ * with the recommended scaling. (For 12-bit sample data, the intermediate
+ * array is INT32 anyway.)
+ *
+ * To avoid overflow of the 32-bit intermediate results in pass 2, we must
+ * have BITS_IN_JSAMPLE + CONST_BITS + PASS1_BITS <= 26. Error analysis
+ * shows that the values given below are the most effective.
+ */
+
+#if BITS_IN_JSAMPLE == 8
+#define CONST_BITS 13
+#define PASS1_BITS 2
+#else
+#define CONST_BITS 13
+#define PASS1_BITS 1 /* lose a little precision to avoid overflow */
+#endif
+
+/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus
+ * causing a lot of useless floating-point operations at run time.
+ * To get around this we use the following pre-calculated constants.
+ * If you change CONST_BITS you may want to add appropriate values.
+ * (With a reasonable C compiler, you can just rely on the FIX() macro...)
+ */
+
+#if CONST_BITS == 13
+#define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */
+#define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */
+#define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */
+#define FIX_0_765366865 ((INT32) 6270) /* FIX(0.765366865) */
+#define FIX_0_899976223 ((INT32) 7373) /* FIX(0.899976223) */
+#define FIX_1_175875602 ((INT32) 9633) /* FIX(1.175875602) */
+#define FIX_1_501321110 ((INT32) 12299) /* FIX(1.501321110) */
+#define FIX_1_847759065 ((INT32) 15137) /* FIX(1.847759065) */
+#define FIX_1_961570560 ((INT32) 16069) /* FIX(1.961570560) */
+#define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */
+#define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */
+#define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */
+#else
+#define FIX_0_298631336 FIX(0.298631336)
+#define FIX_0_390180644 FIX(0.390180644)
+#define FIX_0_541196100 FIX(0.541196100)
+#define FIX_0_765366865 FIX(0.765366865)
+#define FIX_0_899976223 FIX(0.899976223)
+#define FIX_1_175875602 FIX(1.175875602)
+#define FIX_1_501321110 FIX(1.501321110)
+#define FIX_1_847759065 FIX(1.847759065)
+#define FIX_1_961570560 FIX(1.961570560)
+#define FIX_2_053119869 FIX(2.053119869)
+#define FIX_2_562915447 FIX(2.562915447)
+#define FIX_3_072711026 FIX(3.072711026)
+#endif
+
+
+/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result.
+ * For 8-bit samples with the recommended scaling, all the variable
+ * and constant values involved are no more than 16 bits wide, so a
+ * 16x16->32 bit multiply can be used instead of a full 32x32 multiply.
+ * For 12-bit samples, a full 32-bit multiplication will be needed.
+ */
+
+#if BITS_IN_JSAMPLE == 8
+#define MULTIPLY(var,const) MULTIPLY16C16(var,const)
+#else
+#define MULTIPLY(var,const) ((var) * (const))
+#endif
+
+
+/*
+ * Perform the forward DCT on one block of samples.
+ */
+
+GLOBAL void
+jpeg_fdct_islow (DCTELEM * data)
+{
+ INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
+ INT32 tmp10, tmp11, tmp12, tmp13;
+ INT32 z1, z2, z3, z4, z5;
+ DCTELEM *dataptr;
+ int ctr;
+ SHIFT_TEMPS
+
+ /* Pass 1: process rows. */
+ /* Note results are scaled up by sqrt(8) compared to a true DCT; */
+ /* furthermore, we scale the results by 2**PASS1_BITS. */
+
+ dataptr = data;
+ for (ctr = DCTSIZE-1; ctr >= 0; ctr--) {
+ tmp0 = dataptr[0] + dataptr[7];
+ tmp7 = dataptr[0] - dataptr[7];
+ tmp1 = dataptr[1] + dataptr[6];
+ tmp6 = dataptr[1] - dataptr[6];
+ tmp2 = dataptr[2] + dataptr[5];
+ tmp5 = dataptr[2] - dataptr[5];
+ tmp3 = dataptr[3] + dataptr[4];
+ tmp4 = dataptr[3] - dataptr[4];
+
+ /* Even part per LL&M figure 1 --- note that published figure is faulty;
+ * rotator "sqrt(2)*c1" should be "sqrt(2)*c6".
+ */
+
+ tmp10 = tmp0 + tmp3;
+ tmp13 = tmp0 - tmp3;
+ tmp11 = tmp1 + tmp2;
+ tmp12 = tmp1 - tmp2;
+
+ dataptr[0] = (DCTELEM) ((tmp10 + tmp11) << PASS1_BITS);
+ dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << PASS1_BITS);
+
+ z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100);
+ dataptr[2] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp13, FIX_0_765366865),
+ CONST_BITS-PASS1_BITS);
+ dataptr[6] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp12, - FIX_1_847759065),
+ CONST_BITS-PASS1_BITS);
+
+ /* Odd part per figure 8 --- note paper omits factor of sqrt(2).
+ * cK represents cos(K*pi/16).
+ * i0..i3 in the paper are tmp4..tmp7 here.
+ */
+
+ z1 = tmp4 + tmp7;
+ z2 = tmp5 + tmp6;
+ z3 = tmp4 + tmp6;
+ z4 = tmp5 + tmp7;
+ z5 = MULTIPLY(z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */
+
+ tmp4 = MULTIPLY(tmp4, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */
+ tmp5 = MULTIPLY(tmp5, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */
+ tmp6 = MULTIPLY(tmp6, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */
+ tmp7 = MULTIPLY(tmp7, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */
+ z1 = MULTIPLY(z1, - FIX_0_899976223); /* sqrt(2) * (c7-c3) */
+ z2 = MULTIPLY(z2, - FIX_2_562915447); /* sqrt(2) * (-c1-c3) */
+ z3 = MULTIPLY(z3, - FIX_1_961570560); /* sqrt(2) * (-c3-c5) */
+ z4 = MULTIPLY(z4, - FIX_0_390180644); /* sqrt(2) * (c5-c3) */
+
+ z3 += z5;
+ z4 += z5;
+
+ dataptr[7] = (DCTELEM) DESCALE(tmp4 + z1 + z3, CONST_BITS-PASS1_BITS);
+ dataptr[5] = (DCTELEM) DESCALE(tmp5 + z2 + z4, CONST_BITS-PASS1_BITS);
+ dataptr[3] = (DCTELEM) DESCALE(tmp6 + z2 + z3, CONST_BITS-PASS1_BITS);
+ dataptr[1] = (DCTELEM) DESCALE(tmp7 + z1 + z4, CONST_BITS-PASS1_BITS);
+
+ dataptr += DCTSIZE; /* advance pointer to next row */
+ }
+
+ /* Pass 2: process columns.
+ * We remove the PASS1_BITS scaling, but leave the results scaled up
+ * by an overall factor of 8.
+ */
+
+ dataptr = data;
+ for (ctr = DCTSIZE-1; ctr >= 0; ctr--) {
+ tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7];
+ tmp7 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7];
+ tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6];
+ tmp6 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6];
+ tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5];
+ tmp5 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5];
+ tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4];
+ tmp4 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4];
+
+ /* Even part per LL&M figure 1 --- note that published figure is faulty;
+ * rotator "sqrt(2)*c1" should be "sqrt(2)*c6".
+ */
+
+ tmp10 = tmp0 + tmp3;
+ tmp13 = tmp0 - tmp3;
+ tmp11 = tmp1 + tmp2;
+ tmp12 = tmp1 - tmp2;
+
+ dataptr[DCTSIZE*0] = (DCTELEM) DESCALE(tmp10 + tmp11, PASS1_BITS);
+ dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp10 - tmp11, PASS1_BITS);
+
+ z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100);
+ dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp13, FIX_0_765366865),
+ CONST_BITS+PASS1_BITS);
+ dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp12, - FIX_1_847759065),
+ CONST_BITS+PASS1_BITS);
+
+ /* Odd part per figure 8 --- note paper omits factor of sqrt(2).
+ * cK represents cos(K*pi/16).
+ * i0..i3 in the paper are tmp4..tmp7 here.
+ */
+
+ z1 = tmp4 + tmp7;
+ z2 = tmp5 + tmp6;
+ z3 = tmp4 + tmp6;
+ z4 = tmp5 + tmp7;
+ z5 = MULTIPLY(z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */
+
+ tmp4 = MULTIPLY(tmp4, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */
+ tmp5 = MULTIPLY(tmp5, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */
+ tmp6 = MULTIPLY(tmp6, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */
+ tmp7 = MULTIPLY(tmp7, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */
+ z1 = MULTIPLY(z1, - FIX_0_899976223); /* sqrt(2) * (c7-c3) */
+ z2 = MULTIPLY(z2, - FIX_2_562915447); /* sqrt(2) * (-c1-c3) */
+ z3 = MULTIPLY(z3, - FIX_1_961570560); /* sqrt(2) * (-c3-c5) */
+ z4 = MULTIPLY(z4, - FIX_0_390180644); /* sqrt(2) * (c5-c3) */
+
+ z3 += z5;
+ z4 += z5;
+
+ dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp4 + z1 + z3,
+ CONST_BITS+PASS1_BITS);
+ dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp5 + z2 + z4,
+ CONST_BITS+PASS1_BITS);
+ dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp6 + z2 + z3,
+ CONST_BITS+PASS1_BITS);
+ dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp7 + z1 + z4,
+ CONST_BITS+PASS1_BITS);
+
+ dataptr++; /* advance pointer to next column */
+ }
+}
+
+#endif /* DCT_ISLOW_SUPPORTED */
diff --git a/psxdev/table.h b/psxdev/table.h
new file mode 100644
index 0000000..3e50b18
--- /dev/null
+++ b/psxdev/table.h
@@ -0,0 +1,102 @@
+const static huff_t table0[]={
+ {6,3},{8,5},{10,6},{12,8},{76,9},{66,9},{20,11},{58,13},{48,13},{38,13},{32,13},{52,14},{50,14},{48,14},{46,14},{62,15},{60,15},{58,15},{56,15},{54,15},{52,15},{50,15},{48,15},{46,15},{44,15},{42,15},{40,15},{38,15},{36,15},{34,15},{32,15},{48,16},{46,16},{44,16},{42,16},{40,16},{38,16},{36,16},{34,16},{32,16},
+};
+const static huff_t table1[]={
+ {6,4},{12,7},{74,9},{24,11},{54,13},{44,14},{42,14},{62,16},{60,16},{58,16},{56,16},{54,16},{52,16},{50,16},{38,17},{36,17},{34,17},{32,17},
+};
+const static huff_t table2[]={
+ {10,5},{8,8},{22,11},{40,13},{40,14},
+};
+const static huff_t table3[]={
+ {14,6},{72,9},{56,13},{38,14},
+};
+const static huff_t table4[]={
+ {12,6},{30,11},{36,13},
+};
+const static huff_t table5[]={
+ {14,7},{18,11},{36,14},
+};
+const static huff_t table6[]={
+ {10,7},{60,13},{40,17},
+};
+const static huff_t table7[]={
+ {8,7},{42,13},
+};
+const static huff_t table8[]={
+ {14,8},{34,13},
+};
+const static huff_t table9[]={
+ {10,8},{34,14},
+};
+const static huff_t table10[]={
+ {78,9},{32,14},
+};
+const static huff_t table11[]={
+ {70,9},{52,17},
+};
+const static huff_t table12[]={
+ {68,9},{50,17},
+};
+const static huff_t table13[]={
+ {64,9},{48,17},
+};
+const static huff_t table14[]={
+ {28,11},{46,17},
+};
+const static huff_t table15[]={
+ {26,11},{44,17},
+};
+const static huff_t table16[]={
+ {16,11},{42,17},
+};
+const static huff_t table17[]={
+ {62,13},
+};
+const static huff_t table18[]={
+ {52,13},
+};
+const static huff_t table19[]={
+ {50,13},
+};
+const static huff_t table20[]={
+ {46,13},
+};
+const static huff_t table21[]={
+ {44,13},
+};
+const static huff_t table22[]={
+ {62,14},
+};
+const static huff_t table23[]={
+ {60,14},
+};
+const static huff_t table24[]={
+ {58,14},
+};
+const static huff_t table25[]={
+ {56,14},
+};
+const static huff_t table26[]={
+ {54,14},
+};
+const static huff_t table27[]={
+ {62,17},
+};
+const static huff_t table28[]={
+ {60,17},
+};
+const static huff_t table29[]={
+ {58,17},
+};
+const static huff_t table30[]={
+ {56,17},
+};
+const static huff_t table31[]={
+ {54,17},
+};
+const static huff_t *huff_table[]={
+ table0,table1,table2,table3,table4,table5,table6,table7,table8,table9,table10,table11,table12,table13,table14,table15,table16,table17,table18,table19,table20,table21,table22,table23,table24,table25,table26,table27,table28,table29,table30,table31,
+};
+const static int maxlevel[]={
+ 40,18,5,4,3,3,3,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
+};
diff --git a/psxdev/vlc.c b/psxdev/vlc.c
new file mode 100644
index 0000000..aa7e610
--- /dev/null
+++ b/psxdev/vlc.c
@@ -0,0 +1,602 @@
+#include "bs.h"
+#include <sys/types.h>
+
+#define SOFT
+
+#define CODE1(a,b,c) (((a)<<10)|((b)&0x3ff)|((c)<<16))
+/* run, level, bit */
+#define CODE(a,b,c) CODE1(a,b,c+1),CODE1(a,-b,c+1)
+#define CODE0(a,b,c) CODE1(a,b,c),CODE1(a,b,c)
+#define CODE2(a,b,c) CODE1(a,b,c+1),CODE1(a,b,c+1)
+#define RUNOF(a) ((a)>>10)
+#define VALOF(a) ((short)((a)<<6)>>6)
+#define BITOF(a) ((a)>>16)
+#define EOB 0xfe00
+#define ESCAPE_CODE CODE1(63,0,6)
+#define EOB_CODE CODE1(63,512,2)
+
+/*
+ DC code
+ Y U,V
+0 100 00 0
+1 00x 01x -1,1
+2 01xx 10xx -3,-2,2,3
+3 101xxx 110xxx -7..-4,4..7
+4 110xxxx 1110 -15..-8,8..15
+5 1110xxxxx 11110 -31..-16,16..31
+6 11110xxxxxx 111110 -63..-32,32..63
+7 111110 1111110 -127..-64,64..127
+8 1111110 11111110 -255..-128,128..255
+ 7+8 8+8
+*/
+
+/*
+ This table based on MPEG2DEC by MPEG Software Simulation Group
+*/
+
+/* Table B-14, DCT coefficients table zero,
+* codes 0100 ... 1xxx (used for all other coefficients)
+*/
+static const u_long VLCtabnext[12*2] = {
+ CODE(0,2,4), CODE(2,1,4), CODE2(1,1,3), CODE2(1,-1,3),
+ CODE0(63,512,2), CODE0(63,512,2), CODE0(63,512,2), CODE0(63,512,2), /*EOB*/
+ CODE2(0,1,2), CODE2(0,1,2), CODE2(0,-1,2), CODE2(0,-1,2)
+};
+
+/* Table B-14, DCT coefficients table zero,
+* codes 000001xx ... 00111xxx
+*/
+static const u_long VLCtab0[60*2] = {
+ CODE0(63,0,6), CODE0(63,0,6),CODE0(63,0,6), CODE0(63,0,6), /* ESCAPE */
+ CODE2(2,2,7), CODE2(2,-2,7), CODE2(9,1,7), CODE2(9,-1,7),
+ CODE2(0,4,7), CODE2(0,-4,7), CODE2(8,1,7), CODE2(8,-1,7),
+ CODE2(7,1,6), CODE2(7,1,6), CODE2(7,-1,6), CODE2(7,-1,6),
+ CODE2(6,1,6), CODE2(6,1,6), CODE2(6,-1,6), CODE2(6,-1,6),
+ CODE2(1,2,6), CODE2(1,2,6), CODE2(1,-2,6), CODE2(1,-2,6),
+ CODE2(5,1,6), CODE2(5,1,6), CODE2(5,-1,6), CODE2(5,-1,6),
+ CODE(13,1,8), CODE(0,6,8), CODE(12,1,8), CODE(11,1,8),
+ CODE(3,2,8), CODE(1,3,8), CODE(0,5,8), CODE(10,1,8),
+ CODE2(0,3,5), CODE2(0,3,5), CODE2(0,3,5), CODE2(0,3,5),
+ CODE2(0,-3,5), CODE2(0,-3,5), CODE2(0,-3,5), CODE2(0,-3,5),
+ CODE2(4,1,5), CODE2(4,1,5), CODE2(4,1,5), CODE2(4,1,5),
+ CODE2(4,-1,5), CODE2(4,-1,5), CODE2(4,-1,5), CODE2(4,-1,5),
+ CODE2(3,1,5), CODE2(3,1,5), CODE2(3,1,5), CODE2(3,1,5),
+ CODE2(3,-1,5), CODE2(3,-1,5), CODE2(3,-1,5), CODE2(3,-1,5)
+};
+
+/* Table B-14, DCT coefficients table zero,
+* codes 0000001000 ... 0000001111
+*/
+static const u_long VLCtab1[8*2] = {
+ CODE(16,1,10), CODE(5,2,10), CODE(0,7,10), CODE(2,3,10),
+ CODE(1,4,10), CODE(15,1,10), CODE(14,1,10), CODE(4,2,10)
+};
+
+/* Table B-14/15, DCT coefficients table zero / one,
+* codes 000000010000 ... 000000011111
+*/
+static const u_long VLCtab2[16*2] = {
+ CODE(0,11,12), CODE(8,2,12), CODE(4,3,12), CODE(0,10,12),
+ CODE(2,4,12), CODE(7,2,12), CODE(21,1,12), CODE(20,1,12),
+ CODE(0,9,12), CODE(19,1,12), CODE(18,1,12), CODE(1,5,12),
+ CODE(3,3,12), CODE(0,8,12), CODE(6,2,12), CODE(17,1,12)
+};
+
+/* Table B-14/15, DCT coefficients table zero / one,
+* codes 0000000010000 ... 0000000011111
+*/
+static const u_long VLCtab3[16*2] = {
+ CODE(10,2,13), CODE(9,2,13), CODE(5,3,13), CODE(3,4,13),
+ CODE(2,5,13), CODE(1,7,13), CODE(1,6,13), CODE(0,15,13),
+ CODE(0,14,13), CODE(0,13,13), CODE(0,12,13), CODE(26,1,13),
+ CODE(25,1,13), CODE(24,1,13), CODE(23,1,13), CODE(22,1,13)
+};
+
+/* Table B-14/15, DCT coefficients table zero / one,
+* codes 00000000010000 ... 00000000011111
+*/
+static const u_long VLCtab4[16*2] = {
+ CODE(0,31,14), CODE(0,30,14), CODE(0,29,14), CODE(0,28,14),
+ CODE(0,27,14), CODE(0,26,14), CODE(0,25,14), CODE(0,24,14),
+ CODE(0,23,14), CODE(0,22,14), CODE(0,21,14), CODE(0,20,14),
+ CODE(0,19,14), CODE(0,18,14), CODE(0,17,14), CODE(0,16,14)
+};
+
+/* Table B-14/15, DCT coefficients table zero / one,
+* codes 000000000010000 ... 000000000011111
+*/
+static const u_long VLCtab5[16*2] = {
+ CODE(0,40,15), CODE(0,39,15), CODE(0,38,15), CODE(0,37,15),
+ CODE(0,36,15), CODE(0,35,15), CODE(0,34,15), CODE(0,33,15),
+ CODE(0,32,15), CODE(1,14,15), CODE(1,13,15), CODE(1,12,15),
+ CODE(1,11,15), CODE(1,10,15), CODE(1,9,15), CODE(1,8,15)
+};
+
+/* Table B-14/15, DCT coefficients table zero / one,
+* codes 0000000000010000 ... 0000000000011111
+*/
+static const u_long VLCtab6[16*2] = {
+ CODE(1,18,16), CODE(1,17,16), CODE(1,16,16), CODE(1,15,16),
+ CODE(6,3,16), CODE(16,2,16), CODE(15,2,16), CODE(14,2,16),
+ CODE(13,2,16), CODE(12,2,16), CODE(11,2,16), CODE(31,1,16),
+ CODE(30,1,16), CODE(29,1,16), CODE(28,1,16), CODE(27,1,16)
+};
+
+/*
+ DC code
+ Y U,V
+0 100 00 0
+1 00x 01x -1,1
+2 01xx 10xx -3,-2,2,3
+3 101xxx 110xxx -7..-4,4..7
+4 110xxxx 1110xxxx -15..-8,8..15
+5 1110xxxxx 11110xxxxx -31..-16,16..31
+6 11110xxxxxx 111110xxxxxx -63..-32,32..63
+7 111110xxxxxxx 1111110xxxxxxx -127..-64,64..127
+8 1111110xxxxxxxx 11111110xxxxxxxx -255..-128,128..255
+*/
+
+static const u_long DC_Ytab0[48] = {
+ CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),
+ CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),
+ CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),
+ CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),
+
+ CODE1(0,-3,4),CODE1(0,-3,4),CODE1(0,-3,4),CODE1(0,-3,4),
+ CODE1(0,-2,4),CODE1(0,-2,4),CODE1(0,-2,4),CODE1(0,-2,4),
+ CODE1(0,2,4),CODE1(0,2,4),CODE1(0,2,4),CODE1(0,2,4),
+ CODE1(0,3,4),CODE1(0,3,4),CODE1(0,3,4),CODE1(0,3,4),
+
+ CODE1(0,0,3),CODE1(0,0,3),CODE1(0,0,3),CODE1(0,0,3),
+ CODE1(0,0,3),CODE1(0,0,3),CODE1(0,0,3),CODE1(0,0,3),
+ CODE1(0,-7,6),CODE1(0,-6,6),CODE1(0,-5,6),CODE1(0,-4,6),
+ CODE1(0,4,6),CODE1(0,5,6),CODE1(0,6,6),CODE1(0,7,6),
+
+};
+
+static const u_long DC_UVtab0[56] = {
+ CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),
+ CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),
+ CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),
+ CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),CODE1(0,0,2),
+
+ CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),
+ CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),CODE1(0,-1,3),
+ CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),
+ CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),CODE1(0,1,3),
+
+ CODE1(0,-3,4),CODE1(0,-3,4),CODE1(0,-3,4),CODE1(0,-3,4),
+ CODE1(0,-2,4),CODE1(0,-2,4),CODE1(0,-2,4),CODE1(0,-2,4),
+ CODE1(0,2,4),CODE1(0,2,4),CODE1(0,2,4),CODE1(0,2,4),
+ CODE1(0,3,4),CODE1(0,3,4),CODE1(0,3,4),CODE1(0,3,4),
+
+ CODE1(0,-7,6),CODE1(0,-6,6),CODE1(0,-5,6),CODE1(0,-4,6),
+ CODE1(0,4,6),CODE1(0,5,6),CODE1(0,6,6),CODE1(0,7,6),
+};
+
+#define DCTSIZE2 64
+
+/* decode one intra coded MPEG-1 block */
+
+#define Show_Bits(N) (bitbuf>>(32-(N)))
+/* 最小有効bit 17 bit*/
+
+#define Flush_Buffer(N) {bitbuf <<=(N);incnt +=(N);while(incnt>=0) {bitbuf |= Get_Word()<<incnt;incnt-=16;}}
+
+#define Init_Buffer() {bitbuf = (mdec_bs[0]<<16)|(mdec_bs[1]);mdec_bs+=2;incnt = -16;}
+
+#define Get_Word() (*mdec_bs++)
+#define Printf printf
+
+
+int DecDCTvlc(u_short *mdec_bs,u_short *mdec_rl)
+{
+// u_short *mdec_bs = mdecbs,*mdec_rl = mdecrl
+ u_short *rl_end;
+ u_long bitbuf;
+ int incnt; /* 16-有効bit数 x86=char risc = long */
+ int q_code;
+ int type,n;
+ int last_dc[3];
+
+/* BS_HDR u_short rlsize,magic,ver,q_scale */
+
+ //printf("%04x,%04x,",mdec_bs[0],mdec_bs[1]);
+ *(long*)mdec_rl=*(long*)mdec_bs;
+ mdec_rl+=2;
+ rl_end = mdec_rl+(int)mdec_bs[0]*2;
+ q_code = (mdec_bs[2]<<10); /* code = q */
+ type = mdec_bs[3];
+ mdec_bs+=4;
+
+ Init_Buffer();
+
+ n = 0;
+ last_dc[0]=last_dc[1]=last_dc[2] = 0;
+ while(mdec_rl<rl_end) {
+ u_long code2;
+ /* DC */
+ if (type==2) {
+ code2 = Show_Bits(10)|(10<<16); /* DC code */
+ } else {
+ code2 = Show_Bits(6);
+ if (n>=2) {
+ /* Y */
+ if (code2<48) {
+ code2 = DC_Ytab0[code2];
+ code2 = (code2&0xffff0000)|((last_dc[2]+=VALOF(code2)*4)&0x3ff);
+ } else {
+ int nbit,val;
+ int bit = 3;
+ while(Show_Bits(bit)&1) { bit++;}
+ bit++;
+ nbit = bit*2-1;
+ val = Show_Bits(nbit)&((1<<bit)-1);
+ if ((val&(1<<(bit-1)))==0)
+ val -= (1<<bit)-1;
+ val = (last_dc[2]+=val*4);
+ code2 = (nbit<<16) | (val&0x3ff);
+ }
+ //printf("%d ",last_dc[2]);
+ } else {
+ /* U,V */
+ if (code2<56) {
+ code2 = DC_UVtab0[code2];
+ code2 = (code2&0xffff0000)|((last_dc[n]+=VALOF(code2)*4)&0x3ff);
+ } else {
+ int nbit,val;
+ int bit = 4;
+ while(Show_Bits(bit)&1) { bit++;}
+ nbit = bit*2;
+ val = Show_Bits(nbit)&((1<<bit)-1);
+ if ((val&(1<<(bit-1)))==0)
+ val -= (1<<bit)-1;
+ val = (last_dc[n]+=val*4);
+ code2 = (nbit<<16) | (val&0x3ff);
+ }
+ //printf("%d ",last_dc[n]);
+ }
+ if (++n==6) n=0;
+ }
+ // printf("%d ",VALOF(code2));
+ code2 |= q_code;
+
+ /* AC */
+ for(;;){
+// u_long code;
+#define code code2
+#define SBIT 17
+ *mdec_rl++=code2;
+ Flush_Buffer(BITOF(code2));
+ code = Show_Bits(SBIT);
+ if (code>=1<<(SBIT- 2)) {
+ code2 = VLCtabnext[(code>>12)-8];
+ if (code2==EOB_CODE) break;
+ }
+ else if (code>=1<<(SBIT- 6)) {
+ code2 = VLCtab0[(code>>8)-8];
+ if (code2==ESCAPE_CODE) {
+ Flush_Buffer(6); /* ESCAPE len */
+ code2 = Show_Bits(16)| (16<<16);
+ }
+ }
+ else if (code>=1<<(SBIT- 7)) code2 = VLCtab1[(code>>6)-16];
+ else if (code>=1<<(SBIT- 8)) code2 = VLCtab2[(code>>4)-32];
+ else if (code>=1<<(SBIT- 9)) code2 = VLCtab3[(code>>3)-32];
+ else if (code>=1<<(SBIT-10)) code2 = VLCtab4[(code>>2)-32];
+ else if (code>=1<<(SBIT-11)) code2 = VLCtab5[(code>>1)-32];
+ else if (code>=1<<(SBIT-12)) code2 = VLCtab6[(code>>0)-32];
+ else {
+ do {
+ *mdec_rl++=EOB;
+ } while(mdec_rl<rl_end);
+ return 0;
+ }
+ }
+ *mdec_rl++=code2; /* EOB code */
+ Flush_Buffer(2); /* EOB bitlen */
+ }
+ return 0;
+}
+
+
+
+/* this table is based on djpeg by Independent Jpeg Group */
+
+static const int aanscales[DCTSIZE2] = {
+ /* precomputed values scaled up by 14 bits */
+ 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520,
+ 22725, 31521, 29692, 26722, 22725, 17855, 12299, 6270,
+ 21407, 29692, 27969, 25172, 21407, 16819, 11585, 5906,
+ 19266, 26722, 25172, 22654, 19266, 15137, 10426, 5315,
+ 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520,
+ 12873, 17855, 16819, 15137, 12873, 10114, 6967, 3552,
+ 8867, 12299, 11585, 10426, 8867, 6967, 4799, 2446,
+ 4520, 6270, 5906, 5315, 4520, 3552, 2446, 1247
+};
+
+extern unsigned char zscan[DCTSIZE2];
+
+typedef struct {
+ int iqtab[DCTSIZE2];
+ unsigned char *iq_y;
+ u_short *mdec_rl,*rl_end;
+ int mdec_mode;
+} bs_context_t;
+
+void iqtab_init(bs_context_t *ctxt)
+{
+#define CONST_BITS 14
+#define IFAST_SCALE_BITS 2
+ int i;
+ for(i=0;i<DCTSIZE2;i++) {
+ ctxt->iqtab[i] =ctxt->iq_y[i]*aanscales[i]>>(CONST_BITS-IFAST_SCALE_BITS);
+ }
+}
+
+#define BLOCK long
+
+extern IDCT(BLOCK *blk,int k);
+
+u_short* rl2blk(bs_context_t *ctxt, BLOCK *blk,u_short *mdec_rl)
+{
+ int i,k,q_scale,rl;
+ memset(blk,0,6*DCTSIZE2*sizeof(BLOCK));
+ for(i=0;i<6;i++) {
+ rl = *mdec_rl++;
+ q_scale = RUNOF(rl);
+ blk[0] = ctxt->iqtab[0]*VALOF(rl);
+ k = 0;
+ for(;;) {
+ rl = *mdec_rl++;
+ if (rl==EOB) break;
+ k += RUNOF(rl)+1;
+ blk[zscan[k]] = ctxt->iqtab[zscan[k]]*q_scale*VALOF(rl)/8;
+ }
+
+ IDCT(blk,k+1);
+
+ blk+=DCTSIZE2;
+ }
+ return mdec_rl;
+}
+
+#define RGB15(r,g,b) ( (((b)&0xf8)<<7)|(((g)&0xf8)<<2)|((r)>>3) )
+
+#define ROUND(r) bs_roundtbl[(r)+256]
+#if 1
+#define SHIFT 12
+#define toFIX(a) (int)((a)*(1<<SHIFT))
+#define toINT(a) ((a)>>SHIFT)
+#define FIX_1 toFIX(1)
+#define MULR(a) toINT((a)*toFIX(1.402))
+#define MULG(a) toINT((a)*toFIX(-0.3437))
+#define MULG2(a) toINT((a)*toFIX(-0.7143))
+#define MULB(a) toINT((a)*toFIX(1.772))
+#else
+#define MULR(a) 0
+#define MULG(a) 0
+#define MULG2(a) 0
+#define MULB(a) 0
+#endif
+
+
+/*
+int ROUND(int r)
+{
+ if (r<0) return 0;
+ else if (r>255) return 255;
+ else return r;
+}
+*/
+
+extern u_char bs_roundtbl[256*3];
+
+static void yuv2rgb15(BLOCK *blk,u_short *image)
+{
+ int x,yy;
+ BLOCK *yblk = blk+DCTSIZE2*2;
+ for(yy=0;yy<16;yy+=2,blk+=4,yblk+=8,image+=8+16) {
+ if (yy==8) yblk+=DCTSIZE2;
+ for(x=0;x<4;x++,blk++,yblk+=2,image+=2) {
+ int r0,b0,g0,y;
+ r0 = MULR(blk[DCTSIZE2]); /* cr */
+ g0 = MULG(blk[0])+MULG2(blk[DCTSIZE2]);
+ b0 = MULB(blk[0]); /* cb */
+ y = yblk[0]+128;
+ image[0] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ y = yblk[1]+128+4;
+ image[1] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ y = yblk[8]+128+6;
+ image[16] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ y = yblk[9]+128+2;
+ image[17] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ r0 = MULR(blk[4+DCTSIZE2]);
+ g0 = MULG(blk[4])+MULG2(blk[4+DCTSIZE2]);
+ b0 = MULB(blk[4]);
+ y = yblk[DCTSIZE2+0]+128;
+ image[8+0] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ y = yblk[DCTSIZE2+1]+128+4;
+ image[8+1] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ y = yblk[DCTSIZE2+8]+128+6;
+ image[8+16] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ y = yblk[DCTSIZE2+9]+128+2;
+ image[8+17] = RGB15(ROUND(r0+y),ROUND(g0+y),ROUND(b0+y));
+ }
+ }
+}
+
+enum {B,G,R};
+
+static void yuv2rgb24(BLOCK *blk,u_char image[][3])
+{
+ int x,yy;
+ BLOCK *yblk = blk+DCTSIZE2*2;
+ for(yy=0;yy<16;yy+=2,blk+=4,yblk+=8,image+=8+16) {
+ if (yy==8) yblk+=DCTSIZE2;
+ for(x=0;x<4;x++,blk++,yblk+=2,image+=2) {
+ int r0,b0,g0,y;
+ r0 = MULR(blk[DCTSIZE2]); /* cr */
+ g0 = MULG(blk[0])+MULG2(blk[DCTSIZE2]);
+ b0 = MULB(blk[0]); /* cb */
+ y = yblk[0]+128;
+ image[0][R] = ROUND(r0+y);
+ image[0][G] = ROUND(g0+y);
+ image[0][B] = ROUND(b0+y);
+ y = yblk[1]+128;
+ image[1][R] = ROUND(r0+y);
+ image[1][G] = ROUND(g0+y);
+ image[1][B] = ROUND(b0+y);
+ y = yblk[8]+128;
+ image[16][R] = ROUND(r0+y);
+ image[16][G] = ROUND(g0+y);
+ image[16][B] = ROUND(b0+y);
+ y = yblk[9]+128;
+ image[17][R] = ROUND(r0+y);
+ image[17][G] = ROUND(g0+y);
+ image[17][B] = ROUND(b0+y);
+
+ r0 = MULR(blk[4+DCTSIZE2]);
+ g0 = MULG(blk[4])+MULG2(blk[4+DCTSIZE2]);
+ b0 = MULB(blk[4]);
+ y = yblk[DCTSIZE2+0]+128;
+ image[8+0][R] = ROUND(r0+y);
+ image[8+0][G] = ROUND(g0+y);
+ image[8+0][B] = ROUND(b0+y);
+ y = yblk[DCTSIZE2+1]+128;
+ image[8+1][R] = ROUND(r0+y);
+ image[8+1][G] = ROUND(g0+y);
+ image[8+1][B] = ROUND(b0+y);
+ y = yblk[DCTSIZE2+8]+128;
+ image[8+16][R] = ROUND(r0+y);
+ image[8+16][G] = ROUND(g0+y);
+ image[8+16][B] = ROUND(b0+y);
+ y = yblk[DCTSIZE2+9]+128;
+ image[8+17][R] = ROUND(r0+y);
+ image[8+17][G] = ROUND(g0+y);
+ image[8+17][B] = ROUND(b0+y);
+ }
+ }
+}
+
+static void DecDCTReset(bs_context_t *ctxt, int mode)
+{
+ iqtab_init(ctxt);
+}
+
+static void DecDCTin(bs_context_t *ctxt, u_short *mdecrl,int mode)
+{
+ mdecrl+=2;
+ ctxt->mdec_rl = mdecrl;
+ ctxt->rl_end = mdecrl+mdecrl[-2]*2;
+ ctxt->mdec_mode = mode;
+}
+
+static void DecDCTout(bs_context_t *ctxt, u_short *image,int size)
+{
+ BLOCK blk[DCTSIZE2*6];
+ int blocksize=16*16;
+ if (ctxt->mdec_mode) blocksize = 16*16*3/2;
+ for(;size>0;size-=blocksize/2,image+=blocksize) {
+ ctxt->mdec_rl = rl2blk(ctxt,blk,ctxt->mdec_rl);
+ if (ctxt->mdec_mode==0) yuv2rgb15(blk,image);
+ else yuv2rgb24(blk,image);
+ }
+}
+
+void bs_decode_rgb24 (
+ unsigned char *outbuf, // output RGB bytes (width*height*3)
+ bs_header_t *img, // input BS image
+ int width, int height, // dimension of BS image
+ unsigned char *myiqtab
+ )
+{
+ unsigned short *buf2 = (unsigned short *) outbuf;
+ unsigned short *bufp = (unsigned short *) img;
+ bs_context_t ctxt;
+ unsigned short *rl,*image;
+ int slice,rlsize;
+ int mode;
+ int x,y;
+ int height2 = (height+15)&~15;
+ int w;
+
+ ctxt.iq_y = myiqtab ? myiqtab : bs_iqtab();
+ mode=1;
+ w=24;
+ width = width*3/2;
+
+ image = (unsigned short *) malloc (height2*w*sizeof(short));
+ rl = (unsigned short *) malloc ((bufp[0]+2)*sizeof(long));
+
+ DecDCTReset(&ctxt,0);
+ DecDCTvlc(bufp,rl);
+ DecDCTin(&ctxt,rl,mode);
+
+ slice = height2*w/2;
+
+ for(x=0;x<width;x+=w)
+ {
+ short *dst,*src;
+ DecDCTout(&ctxt,image,slice);
+ src = image;
+ dst = buf2+x+(0)*width;
+ for(y=height-1;y>=0;y--)
+ {
+ memcpy(dst,src,w*2);
+ src+=w;
+ dst+=width;
+ }
+ }
+
+ free (image);
+ free (rl);
+}
+
+void bs_decode_rgb15 (
+ unsigned short *outbuf, // output RGB bytes (width*height*2)
+ bs_header_t *img, // input BS image
+ int width, int height, // dimension of BS image
+ unsigned char *myiqtab
+ )
+{
+ unsigned short *buf2 = (unsigned short *) outbuf;
+ unsigned short *bufp = (unsigned short *) img;
+ bs_context_t ctxt;
+ unsigned short *rl,*image;
+ int slice,rlsize;
+ int mode;
+ int x,y;
+ int height2 = (height+15)&~15;
+ int w;
+
+ ctxt.iq_y = myiqtab ? myiqtab : bs_iqtab();
+ mode=0;
+ w=24;
+
+ image = (unsigned short *) malloc (height2*w*sizeof(short));
+ rl = (unsigned short *) malloc ((bufp[0]+2)*sizeof(long));
+
+ DecDCTReset(&ctxt,0);
+ DecDCTvlc(bufp,rl);
+ DecDCTin(&ctxt,rl,mode);
+
+ slice = height2*w/2;
+
+ for(x=0;x<width;x+=w)
+ {
+ short *dst,*src;
+ DecDCTout(&ctxt,image,slice);
+ src = image;
+ dst = buf2+x+(height-1)*width;
+ for(y=height-1;y>=0;y--)
+ {
+ memcpy(dst,src,w*2);
+ src+=w;
+ dst-=width;
+ }
+ }
+
+ free (image);
+ free (rl);
+}
diff --git a/str-util.cpp b/str-util.cpp
index cf3f6ff..3588707 100644
--- a/str-util.cpp
+++ b/str-util.cpp
@@ -1,5 +1,6 @@
#include <stdlib.h>
#include <string.h>
+//#include "psxdev/bs.h"
#include "fileutils.h"
#include "generic.h"
#include "cdutils.h"
@@ -24,73 +25,6 @@ Channels : 2 bytes 30
*/
-int DCT[8][8] = {
- { 4096, 4096, 4096, 4096, 4096, 4096, 4096, 4096, },
- { 5681, 4816, 3218, 1130, -1130, -3218, -4816, -5681, },
- { 5352, 2217, -2217, -5532, -5532, -2217, 2217, 5352, },
- { 4816, -1130, -5681, -3218, 3218, 5681, 1130, -4816, },
- { 4096, -4096, 4096, -4096, 4096, -4096, 4096, -4096, },
- { 3218, -5681, 1130, 4816, -4816, -1130, 5681, -3218, },
- { 2217, -5352, 5352, -2217, -2217, 5352, -5352, 2217, },
- { 1130, -3218, 4816, -5681, 5681, -4816, 3218, -1130, },
-};
-
-int iDCT[8][8] = {
- { 4096, 5681, 5352, 4816, 4096, 3218, 2217, 1130, },
- { 4096, 4816, 2217, -1130, -4096, -5681, -5352, -3218, },
- { 4096, 3218, -2217, -5681, 4096, 1130, 5352, 4816, },
- { 4096, 1130, -5532, -3218, -4096, 4816, -2217, -5681, },
- { 4096, -1130, -5532, 3218, 4096, -4816, -2217, 5681, },
- { 4096, -3218, -2217, 5681, -4096, -1130, 5352, -4816, },
- { 4096, -4816, 2217, 1130, 4096, 5681, -5352, 3218, },
- { 4096, -5681, 5352, -4816, -4096, -3218, 2217, -1130, },
-};
-
-int zscan[] = {
- 0, 1, 8, 16, 9, 2, 3, 10,
- 17, 24, 32, 25, 18, 11, 4, 5,
- 12, 19, 26, 33, 40, 48, 41, 34,
- 27, 20, 13, 6, 7, 14, 21, 28,
- 35, 42, 49, 56, 57, 50, 43, 36,
- 29, 22, 15, 23, 30, 37, 44, 51,
- 58, 59, 52, 45, 38, 31, 39, 46,
- 53, 60, 61, 54, 47, 55, 62, 63,
-};
-
-int izscan[] = {
- 0, 1, 5, 6, 14, 15, 27, 28,
- 2, 4, 7, 13, 16, 26, 29, 42,
- 3, 8, 12, 17, 25, 30, 41, 43,
- 9, 11, 18, 24, 31, 40, 44, 53,
- 10, 19, 23, 32, 39, 45, 52, 54,
- 20, 22, 33, 38, 46, 51, 55, 60,
- 21, 34, 37, 47, 50, 56, 59, 61,
- 35, 36, 48, 49, 57, 58, 62, 63,
-};
-
-int iqtab[] = {
- 2, 16, 19, 22, 26, 27, 29, 34,
- 16, 16, 22, 24, 27, 29, 34, 37,
- 19, 22, 26, 27, 29, 34, 34, 38,
- 22, 22, 26, 27, 29, 34, 37, 40,
- 22, 26, 27, 29, 32, 35, 40, 48,
- 26, 27, 29, 32, 35, 40, 48, 58,
- 26, 27, 29, 34, 38, 46, 56, 69,
- 27, 29, 35, 38, 46, 56, 69, 83,
-};
-
-float RGB2YUV[3][3] = {
- { 1.00000, 0.00000, 1.40200, },
- { 1.00000, -0.34370, -0.71430, },
- { 1.00000, 1.77200, 0.00000, },
-};
-
-float YUV2RGB[3][3] = {
- { 0.22900, 0.58700, 0.11400, },
- { -0.16871, -0.33130, 0.50000, },
- { 0.50000, -0.41870, -0.08130, },
-};
-
struct STR_Header {
Uint16 StSTATUS;
Uint16 StTYPE;
@@ -107,46 +41,6 @@ struct STR_Header {
Byte * video = 0, * audio = 0;
-void print_bits(Uint16 w, int size = 16) {
- int i;
-
- for (i = size - 1; i >= 0; i--) {
- printm(M_BARE, "%c", w & (1 << i) ? '1' : '0');
- }
-}
-
-Uint32 get_bits(Byte * datas, int size, Uint32 * pointer) {
- static int internal = 0;
- static Uint16 w = 0;
- Uint32 r = 0;
- int os = size;
-
- if (internal < size) {
- r = w;
- size -= internal;
- w = *((Uint16 *) (video + *pointer));
- printm(M_BARE, "get_bits: needs food, got 0x%04x\n", w);
- *pointer += 2;
-// w = swap_word(w);
- internal = 16;
- }
-
- r <<= size;
- r |= ((w >> (16 - size)) & ((1 << size) - 1));
-
- w <<= size;
-
- printm(M_BARE, "Read bits: ");
- print_bits(r, os);
- printm(M_BARE, "\n");
-
- internal -= size;
-
- r &= ((1 << size) - 1);
-
- return r;
-}
-
void process_one_sector(FILE * f) {
Byte sector[2336];
STR_Header * h;
@@ -179,75 +73,9 @@ void process_one_sector(FILE * f) {
memcpy(video + h->StSECTOR_OFFSET * 2016, sector + 40, 2016);
if (h->StSECTOR_SIZE == (h->StSECTOR_OFFSET + 1)) {
- Uint32 pointer = 0;
- Uint32 macrobloc = 0;
// Frame finished.
printm(M_BARE, "End of Frame.\n");
-// fwrite(video, 1, h->StSECTOR_SIZE * 2016, stdout);
-
- Uint16 rl_size = *((Uint16 *) (video + pointer));
- pointer += 2;
- if (*((Uint16 *) (video + pointer)) != 0x3800) {
- printm(M_ERROR, "Broken frame\n");
-// break;
- }
- pointer += 2;
-
- Uint16 quant = *((Uint16 *) (video + pointer));
- pointer += 2;
- Uint16 ver = *((Uint16 *) (video + pointer));
- pointer += 2;
- int rls = 0;
- while (rls <= rl_size) {
- printm(M_BARE, "Processing macrobloc %i, rlsize = %i, quant = %i, ver = %i, pointer = 0x%08x\n", macrobloc, rl_size, quant, ver, pointer);
-
- int16 Cb[64], Cr[64], Y0[64], Y1[64], Y2[64], Y3[64];
- int16 * blocs[6] = {Cb, Cr, Y0, Y1, Y2, Y3};
- int bloc;
-
- for (bloc = 0; bloc < 6; bloc++) {
- int i;
-
- for (i = 0; i < 64; i++) {
- blocs[bloc][i] = 0;
- }
-
- int16 DC = get_bits(video, 10, &pointer);
- if (DC & 0x20)
- DC |= 0xc0;
-
- printm(M_BARE, "Processing subbloc %i, DC = %i, Quant = %i, pointer = 0x%08x\n", bloc, DC, quant, pointer);
-
- int inbloc = 1;
- blocs[bloc][0] = DC;
-
- while (inbloc < 64) {
- Uint8 run = get_bits(video, 6, &pointer);
- int16 level;
-
- level = get_bits(video, 8, &pointer);
-
- if (level == 0) {
- level = get_bits(video, 8, &pointer);
- } else if (level & 0x80) {
- if (level == 0x80)
- level = get_bits(video, 8, &pointer);
- level |= 0xff00;
- }
-
- printm(M_BARE, "Read a runlevel: (%i, %i) = (%x %x), inbloc = %i, (pointer is %i)\n", run, level, run, level, inbloc, pointer);
- rls++;
-
- blocs[bloc][inbloc] = level;
- inbloc += run + 1;
- }
-
- printm(M_BARE, "Finished the bloc with inbloc = %i\n", inbloc);
- }
- macrobloc++;
- }
- exit(-1);
free(video);
}
@@ -259,12 +87,8 @@ void process_one_sector(FILE * f) {
printm(M_BARE, "---------------------------------\n\n");
}
-#ifdef STR_MAIN
-
int main(void) {
while (!feof(stdin)) {
process_one_sector(stdin);
}
}
-
-#endif
diff --git a/yazedc.cpp b/yazedc.cpp
index 9d5b838..f4bced0 100644
--- a/yazedc.cpp
+++ b/yazedc.cpp
@@ -856,162 +856,3 @@ int set_sector_type(int st)
}
/* ------------- --------------*/
-#ifdef MAIN
-
-#define DO_L1 1
-#define DO_L2 2
-#define DO_SUB 4
-
-static const unsigned sect_size[8][2] = {
-/* nothing */
-{0,0},
-/* Layer 1 decode/encode */
-{ (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR, L1_RAW*FRAMES_PER_SECTOR},
-/* Layer 2 decode/encode */
-{ 16+L2_RAW+12+L2_Q+L2_P, L2_RAW},
-/* Layer 1 and 2 decode/encode */
-{ (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR, L1_RAW*FRAMES_PER_SECTOR},
-/* Subchannel decode/encode */
-{ (LSUB_RAW+LSUB_Q+LSUB_P)*PACKETS_PER_SUBCHANNELFRAME,
- LSUB_RAW*PACKETS_PER_SUBCHANNELFRAME},
-/* Layer 1 and subchannel decode/encode */
-{ (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR +
- (LSUB_RAW+LSUB_Q+LSUB_P)*PACKETS_PER_SUBCHANNELFRAME,
- LSUB_RAW*PACKETS_PER_SUBCHANNELFRAME +
- L1_RAW*FRAMES_PER_SECTOR},
-/* Layer 2 and subchannel decode/encode */
-{ L2_RAW+L2_Q+L2_P+
- (LSUB_RAW+LSUB_Q+LSUB_P)*PACKETS_PER_SUBCHANNELFRAME,
- LSUB_RAW*PACKETS_PER_SUBCHANNELFRAME +
- L2_RAW},
-/* Layer 1, 2 and subchannel decode/encode */
-{ (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR +
- (LSUB_RAW+LSUB_Q+LSUB_P)*PACKETS_PER_SUBCHANNELFRAME,
- LSUB_RAW*PACKETS_PER_SUBCHANNELFRAME +
- L1_RAW*FRAMES_PER_SECTOR},
-};
-
-int main(int argc, char **argv)
-{
- int encode = 1;
- int mask = DO_L2;
- FILE * infp;
- FILE * outfp;
- unsigned address = 0;
- unsigned char *l1_inbuf;
- unsigned char *l1_outbuf;
- unsigned char *l2_inbuf;
- unsigned char *l2_outbuf;
- unsigned char *sub_inbuf;
- unsigned char *sub_outbuf;
- unsigned char *last_outbuf;
- unsigned char inbuf[(LSUB_RAW+LSUB_Q+LSUB_P)*PACKETS_PER_SUBCHANNELFRAME +
- (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR];
- unsigned char outbuf[(LSUB_RAW+LSUB_Q+LSUB_P)*PACKETS_PER_SUBCHANNELFRAME +
- (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR];
- unsigned load_offset;
-
- l1_inbuf = l2_inbuf = sub_inbuf = inbuf;
- l1_outbuf = l2_outbuf = sub_outbuf = last_outbuf = outbuf;
-
- infp = fopen("sectors_in", "rb");
- outfp = fopen("sectors_out", "wb");
-
- sectortype= MODE_2_FORM_1;
- address = 0 + 75*2;
-
- switch (sectortype) {
- case MODE_1:
- case MODE_2:
- load_offset = 16;
- break;
- case MODE_2_FORM_1:
- case MODE_2_FORM_2:
- load_offset = 24;
- break;
- default:
- load_offset = 0;
- }
- while(1) {
-
- if (1 != fread(inbuf+load_offset,
- sect_size[mask][encode], 1, infp)) { perror(""); break; }
- if (encode == 1) {
- if (mask & DO_L2) {
- switch (sectortype) {
- case MODE_0:
- break;
- case MODE_1:
- break;
- case MODE_2:
- if (1 !=
- fread(inbuf+load_offset+
- sect_size[mask][encode],
- 2336 - sect_size[mask][encode],
- 1, infp)) { perror(""); break; }
- break;
- case MODE_2_FORM_1:
- break;
- case MODE_2_FORM_2:
- if (1 !=
- fread(inbuf+load_offset+
- sect_size[mask][encode],
- 2324 - sect_size[mask][encode],
- 1, infp)) { perror(""); break; }
- break;
- default:
- if (1 !=
- fread(inbuf+load_offset+
- sect_size[mask][encode],
- 2448 - sect_size[mask][encode],
- 1, infp)) { perror(""); break; }
- memset(inbuf,0,16);
- /*memset(inbuf+16+2048,0,12+272);*/
- break;
- }
- do_encode_L2(l2_inbuf, MODE_1, address);
- if (0) scramble_L2(l2_inbuf);
- last_outbuf = l1_inbuf = l2_inbuf;
- l1_outbuf = l2_inbuf;
- sub_inbuf = l2_inbuf + L2_RAW;
- sub_outbuf = l2_outbuf + 12 + 4+ L2_RAW+4+ 8+ L2_Q+L2_P;
- }
- if (mask & DO_L1) {
- do_encode_L1(l1_inbuf, l1_outbuf,1,1,1,1);
- last_outbuf = l1_outbuf;
- sub_inbuf = l1_inbuf + L1_RAW*FRAMES_PER_SECTOR;
- sub_outbuf = l1_outbuf + (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR;
- }
- if (mask & DO_SUB) {
- do_encode_sub(sub_inbuf, sub_outbuf, 0, 0);
- }
- } else {
- if (mask & DO_L1) {
- do_decode_L1(l1_inbuf, l1_outbuf,1,1,1,1);
- last_outbuf = l2_inbuf = l1_outbuf;
- l2_outbuf = l1_inbuf;
- sub_inbuf = l1_inbuf + (L1_RAW+L1_Q+L1_P)*FRAMES_PER_SECTOR;
- sub_outbuf = l1_outbuf + L1_RAW*FRAMES_PER_SECTOR;
- }
- if (mask & DO_L2) {
- do_decode_L2(l2_inbuf, l2_outbuf);
- last_outbuf = l2_outbuf;
- sub_inbuf = l2_inbuf + L2_RAW+L2_Q+L2_P;
- sub_outbuf = l2_outbuf + L2_RAW;
- }
- if (mask & DO_SUB) {
- do_decode_sub(sub_inbuf, sub_outbuf, 1, 1);
- }
- }
- if (1 != fwrite(last_outbuf, sect_size[mask][1 - encode], 1, outfp)) {
- perror("");
- break;
- }
- address++;
- }
-#if 0
- /* flush the data from the delay lines with zeroed sectors, if necessary */
-#endif
- return 0;
-}
-#endif