Import libatrac from ffmpeg and modify librm to support ATRAC3.

The decoder is still in floating point.


git-svn-id: svn://svn.rockbox.org/rockbox/trunk@22235 a1c6a512-1295-4272-9138-f99709370657

apps/codecs/libatrac/Makefile.test

@@ -0,0 +1,11 @@
+CFLAGS = -Wall -O3 -DTEST -D"DEBUGF=printf"
+OBJS = atrac3.o dsputil.o bitstream.o fft.o mdct.o libavutil/log.o libavutil/mem.o ../librm/rm.o
+
+atractest: $(OBJS)
+	gcc -o atractest $(OBJS) -lm
+
+.c.o :
+	$(CC) $(CFLAGS) -c -o $@ $<
+
+clean:
+	rm -f atractest $(OBJS) *~ output.wav

apps/codecs/libatrac/README.rockbox

@@ -0,0 +1,20 @@
+Library: libatrac
+Imported by : Mohamed Tarek
+Import date : 10-August-2009
+
+LICENSING INFORMATION
+
+ffmpeg is licensed under the Lesser GNU General Public License.
+
+IMPORT DETAILS
+
+The decoder is based on ffmpeg-svn r18079. It still uses floating 
+point math and not suitable to be used in rockbox.
+
+TESTING
+
+The test program should compile in any Unix-like environment using the
+command "make -f Makefile.test".
+
+Running "./atractest file.rm" will decode the audio data to a WAV file
+called "output.wav" in the current directory.

apps/codecs/libatrac/atrac3data.h

@@ -0,0 +1,144 @@
+/*
+ * Atrac 3 compatible decoder data
+ * Copyright (c) 2006-2007 Maxim Poliakovski
+ * Copyright (c) 2006-2007 Benjamin Larsson
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavcodec/atrac3data.h
+ * Atrac 3 AKA RealAudio 8 compatible decoder data
+ */
+
+#ifndef AVCODEC_ATRAC3DATA_H
+#define AVCODEC_ATRAC3DATA_H
+
+#include <stdint.h>
+
+/* VLC tables */
+
+static const uint8_t huffcode1[9] = {
+  0x0,0x4,0x5,0xC,0xD,0x1C,0x1D,0x1E,0x1F,
+};
+
+static const uint8_t huffbits1[9] = {
+  1,3,3,4,4,5,5,5,5,
+};
+
+static const uint8_t huffcode2[5] = {
+  0x0,0x4,0x5,0x6,0x7,
+};
+
+static const uint8_t huffbits2[5] = {
+  1,3,3,3,3,
+};
+
+static const uint8_t huffcode3[7] = {
+0x0,0x4,0x5,0xC,0xD,0xE,0xF,
+};
+
+static const uint8_t huffbits3[7] = {
+  1,3,3,4,4,4,4,
+};
+
+static const uint8_t huffcode4[9] = {
+  0x0,0x4,0x5,0xC,0xD,0x1C,0x1D,0x1E,0x1F,
+};
+
+static const uint8_t huffbits4[9] = {
+  1,3,3,4,4,5,5,5,5,
+};
+
+static const uint8_t huffcode5[15] = {
+  0x0,0x2,0x3,0x8,0x9,0xA,0xB,0x1C,0x1D,0x3C,0x3D,0x3E,0x3F,0xC,0xD,
+};
+
+static const uint8_t huffbits5[15] = {
+  2,3,3,4,4,4,4,5,5,6,6,6,6,4,4
+};
+
+static const uint8_t huffcode6[31] = {
+  0x0,0x2,0x3,0x4,0x5,0x6,0x7,0x14,0x15,0x16,0x17,0x18,0x19,0x34,0x35,
+  0x36,0x37,0x38,0x39,0x3A,0x3B,0x78,0x79,0x7A,0x7B,0x7C,0x7D,0x7E,0x7F,0x8,0x9,
+};
+
+static const uint8_t huffbits6[31] = {
+  3,4,4,4,4,4,4,5,5,5,5,5,5,6,6,6,6,6,6,6,6,7,7,7,7,7,7,7,7,4,4
+};
+
+static const uint8_t huffcode7[63] = {
+  0x0,0x8,0x9,0xA,0xB,0xC,0xD,0xE,0xF,0x10,0x11,0x24,0x25,0x26,0x27,0x28,
+  0x29,0x2A,0x2B,0x2C,0x2D,0x2E,0x2F,0x30,0x31,0x32,0x33,0x68,0x69,0x6A,0x6B,0x6C,
+  0x6D,0x6E,0x6F,0x70,0x71,0x72,0x73,0x74,0x75,0xEC,0xED,0xEE,0xEF,0xF0,0xF1,0xF2,
+  0xF3,0xF4,0xF5,0xF6,0xF7,0xF8,0xF9,0xFA,0xFB,0xFC,0xFD,0xFE,0xFF,0x2,0x3,
+};
+
+static const uint8_t huffbits7[63] = {
+  3,5,5,5,5,5,5,5,5,5,5,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,7,7,7,7,7,
+  7,7,7,7,7,7,7,7,7,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,4,4
+};
+
+static const uint8_t huff_tab_sizes[7] = {
+  9, 5, 7, 9, 15, 31, 63,
+};
+
+static const uint8_t* const huff_codes[7] = {
+  huffcode1,huffcode2,huffcode3,huffcode4,huffcode5,huffcode6,huffcode7,
+};
+
+static const uint8_t* const huff_bits[7] = {
+  huffbits1,huffbits2,huffbits3,huffbits4,huffbits5,huffbits6,huffbits7,
+};
+
+static const uint16_t atrac3_vlc_offs[] = {
+    0,512,1024,1536,2048,2560,3072,3584,4096
+};
+
+/* selector tables */
+
+static const uint8_t CLCLengthTab[8] = {0, 4, 3, 3, 4, 4, 5, 6};
+static const int8_t seTab_0[4] = {0, 1, -2, -1};
+static const int8_t decTable1[18] = {0,0, 0,1, 0,-1, 1,0, -1,0, 1,1, 1,-1, -1,1, -1,-1};
+
+
+/* tables for the scalefactor decoding */
+
+static const float iMaxQuant[8] = {
+  0.0, 1.0/1.5, 1.0/2.5, 1.0/3.5, 1.0/4.5, 1.0/7.5, 1.0/15.5, 1.0/31.5
+};
+
+static const uint16_t subbandTab[33] = {
+  0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 176, 192, 224,
+  256, 288, 320, 352, 384, 416, 448, 480, 512, 576, 640, 704, 768, 896, 1024
+};
+
+/* transform data */
+
+static const float qmf_48tap_half[24] = {
+   -0.00001461907, -0.00009205479, -0.000056157569, 0.00030117269,
+    0.0002422519,-0.00085293897, -0.0005205574, 0.0020340169,
+    0.00078333891, -0.0042153862, -0.00075614988, 0.0078402944,
+   -0.000061169922, -0.01344162, 0.0024626821, 0.021736089,
+   -0.007801671, -0.034090221, 0.01880949, 0.054326009,
+   -0.043596379, -0.099384367, 0.13207909, 0.46424159
+};
+
+/* joint stereo related tables */
+static const float matrixCoeffs[8] = {0.0, 2.0, 2.0, 2.0, 0.0, 0.0, 1.0, 1.0};
+
+#endif /* AVCODEC_ATRAC3DATA_H */

apps/codecs/libatrac/bitstream.c

@@ -0,0 +1,276 @@
+/*
+ * Common bit i/o utils
+ * Copyright (c) 2000, 2001 Fabrice Bellard
+ * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * alternative bitstream reader & writer by Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "bitstream.h"
+
+#ifdef ROCKBOX
+#undef DEBUGF
+#define DEBUGF(...)
+#endif
+
+const uint8_t ff_log2_run[32]={
+ 0, 0, 0, 0, 1, 1, 1, 1,
+ 2, 2, 2, 2, 3, 3, 3, 3,
+ 4, 4, 5, 5, 6, 6, 7, 7,
+ 8, 9,10,11,12,13,14,15
+};
+
+/**
+ * Same as av_mallocz_static(), but does a realloc.
+ *
+ * @param[in] ptr The block of memory to reallocate.
+ * @param[in] size The requested size.
+ * @return Block of memory of requested size.
+ * @deprecated. Code which uses ff_realloc_static is broken/misdesigned
+ * and should correctly use static arrays
+ */
+
+
+void ff_put_string(PutBitContext * pbc, const char *s, int put_zero)
+{
+    while(*s){
+        put_bits(pbc, 8, *s);
+        s++;
+    }
+    if(put_zero)
+        put_bits(pbc, 8, 0);
+}
+
+/* VLC decoding */
+
+//#define DEBUG_VLC
+
+#define GET_DATA(v, table, i, wrap, size) \
+{\
+    const uint8_t *ptr = (const uint8_t *)table + i * wrap;\
+    switch(size) {\
+    case 1:\
+        v = *(const uint8_t *)ptr;\
+        break;\
+    case 2:\
+        v = *(const uint16_t *)ptr;\
+        break;\
+    default:\
+        v = *(const uint32_t *)ptr;\
+        break;\
+    }\
+}
+
+
+static int alloc_table(VLC *vlc, int size, int use_static)
+{
+    int index;
+    index = vlc->table_size;
+    vlc->table_size += size;
+    if (vlc->table_size > vlc->table_allocated) {
+        if(use_static>1){
+            DEBUGF("init_vlc() used with too little memory : table_size > allocated_memory\n");
+        }
+
+        if (!vlc->table)
+            return -1;
+    }
+    return index;
+}
+
+static int build_table(VLC *vlc, int table_nb_bits,
+                       int nb_codes,
+                       const void *bits, int bits_wrap, int bits_size,
+                       const void *codes, int codes_wrap, int codes_size,
+                       const void *symbols, int symbols_wrap, int symbols_size,
+                       uint32_t code_prefix, int n_prefix, int flags)
+{
+    int i, j, k, n, table_size, table_index, nb, n1, index, code_prefix2, symbol;
+    uint32_t code;
+    VLC_TYPE (*table)[2];
+
+    table_size = 1 << table_nb_bits;
+    table_index = alloc_table(vlc, table_size, flags & (INIT_VLC_USE_STATIC|INIT_VLC_USE_NEW_STATIC));
+#ifdef DEBUG_VLC
+    DEBUGF("new table index=%d size=%d code_prefix=%x n=%d\n",
+           table_index, table_size, code_prefix, n_prefix);
+#endif
+    if (table_index < 0)
+        return -1;
+    table = &vlc->table[table_index];
+
+    for(i=0;i<table_size;i++) {
+        table[i][1] = 0; //bits
+        table[i][0] = -1; //codes
+    }
+
+    /* first pass: map codes and compute auxillary table sizes */
+    for(i=0;i<nb_codes;i++) {
+        GET_DATA(n, bits, i, bits_wrap, bits_size);
+        GET_DATA(code, codes, i, codes_wrap, codes_size);
+        /* we accept tables with holes */
+        if (n <= 0)
+            continue;
+        if (!symbols)
+            symbol = i;
+        else
+            GET_DATA(symbol, symbols, i, symbols_wrap, symbols_size);
+#if defined(DEBUG_VLC) && 0
+        DEBUGF("i=%d n=%d code=0x%x\n", i, n, code);
+#endif
+        /* if code matches the prefix, it is in the table */
+        n -= n_prefix;
+        if(flags & INIT_VLC_LE)
+            code_prefix2= code & (n_prefix>=32 ? (int)0xffffffff : (1 << n_prefix)-1);
+        else
+            code_prefix2= code >> n;
+        if (n > 0 && code_prefix2 == (int)code_prefix) {
+            if (n <= table_nb_bits) {
+                /* no need to add another table */
+                j = (code << (table_nb_bits - n)) & (table_size - 1);
+                nb = 1 << (table_nb_bits - n);
+                for(k=0;k<nb;k++) {
+                    if(flags & INIT_VLC_LE)
+                        j = (code >> n_prefix) + (k<<n);
+#ifdef DEBUG_VLC
+                    DEBUGF("%4x: code=%d n=%d\n",
+                           j, i, n);
+#endif
+                    if (table[j][1] /*bits*/ != 0) {
+                        DEBUGF("incorrect codes\n");
+                        return -1;
+                    }
+                    table[j][1] = n; //bits
+                    table[j][0] = symbol;
+                    j++;
+                }
+            } else {
+                n -= table_nb_bits;
+                j = (code >> ((flags & INIT_VLC_LE) ? n_prefix : n)) & ((1 << table_nb_bits) - 1);
+#ifdef DEBUG_VLC
+                DEBUGF("%4x: n=%d (subtable)\n",
+                       j, n);
+#endif
+                /* compute table size */
+                n1 = -table[j][1]; //bits
+                if (n > n1)
+                    n1 = n;
+                table[j][1] = -n1; //bits
+            }
+        }
+    }
+
+    /* second pass : fill auxillary tables recursively */
+    for(i=0;i<table_size;i++) {
+        n = table[i][1]; //bits
+        if (n < 0) {
+            n = -n;
+            if (n > table_nb_bits) {
+                n = table_nb_bits;
+                table[i][1] = -n; //bits
+            }
+            index = build_table(vlc, n, nb_codes,
+                                bits, bits_wrap, bits_size,
+                                codes, codes_wrap, codes_size,
+                                symbols, symbols_wrap, symbols_size,
+                                (flags & INIT_VLC_LE) ? (code_prefix | (i << n_prefix)) : ((code_prefix << table_nb_bits) | i),
+                                n_prefix + table_nb_bits, flags);
+            if (index < 0)
+                return -1;
+            /* note: realloc has been done, so reload tables */
+            table = &vlc->table[table_index];
+            table[i][0] = index; //code
+        }
+    }
+    return table_index;
+}
+
+
+/* Build VLC decoding tables suitable for use with get_vlc().
+
+   'nb_bits' set thee decoding table size (2^nb_bits) entries. The
+   bigger it is, the faster is the decoding. But it should not be too
+   big to save memory and L1 cache. '9' is a good compromise.
+
+   'nb_codes' : number of vlcs codes
+
+   'bits' : table which gives the size (in bits) of each vlc code.
+
+   'codes' : table which gives the bit pattern of of each vlc code.
+
+   'symbols' : table which gives the values to be returned from get_vlc().
+
+   'xxx_wrap' : give the number of bytes between each entry of the
+   'bits' or 'codes' tables.
+
+   'xxx_size' : gives the number of bytes of each entry of the 'bits'
+   or 'codes' tables.
+
+   'wrap' and 'size' allows to use any memory configuration and types
+   (byte/word/long) to store the 'bits', 'codes', and 'symbols' tables.
+
+   'use_static' should be set to 1 for tables, which should be freed
+   with av_free_static(), 0 if free_vlc() will be used.
+*/
+int init_vlc_sparse(VLC *vlc, int nb_bits, int nb_codes,
+             const void *bits, int bits_wrap, int bits_size,
+             const void *codes, int codes_wrap, int codes_size,
+             const void *symbols, int symbols_wrap, int symbols_size,
+             int flags)
+{
+    vlc->bits = nb_bits;
+    if(flags & INIT_VLC_USE_NEW_STATIC){
+        if(vlc->table_size && vlc->table_size == vlc->table_allocated){
+            return 0;
+        }else if(vlc->table_size){
+            return -1; // fatal error, we are called on a partially initialized table
+        }
+    }else if(!(flags & INIT_VLC_USE_STATIC)) {
+        vlc->table = NULL;
+        vlc->table_allocated = 0;
+        vlc->table_size = 0;
+    } else {
+        /* Static tables are initially always NULL, return
+           if vlc->table != NULL to avoid double allocation */
+        if(vlc->table)
+            return 0;
+    }
+
+#ifdef DEBUG_VLC
+    DEBUGF("build table nb_codes=%d\n", nb_codes);
+#endif
+
+    if (build_table(vlc, nb_bits, nb_codes,
+                    bits, bits_wrap, bits_size,
+                    codes, codes_wrap, codes_size,
+                    symbols, symbols_wrap, symbols_size,
+                    0, 0, flags) < 0) {
+        //free(&vlc->table);
+        return -1;
+    }
+    /* Changed the following condition to be true if table_size > table_allocated. *
+     * This would be more sensible for static tables since we want warnings for    *
+     * memory shortages only.                                                      */
+#ifdef TEST
+    if((flags & INIT_VLC_USE_NEW_STATIC) && vlc->table_size > vlc->table_allocated)
+        DEBUGF("needed %d had %d\n", vlc->table_size, vlc->table_allocated);
+#endif
+    return 0;
+}
+

apps/codecs/libatrac/bitstream.h

@@ -0,0 +1,963 @@
+/*
+ * copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef BITSTREAM_H
+#define BITSTREAM_H
+
+#include <inttypes.h>
+#include <stdlib.h>
+//#include <assert.h>
+#include <string.h>
+#include <stdio.h>
+#include "bswap.h"
+
+/* The following 2 defines are taken from libavutil/intreadwrite.h */
+#define AV_RB32(x)  ((((const uint8_t*)(x))[0] << 24) | \
+                     (((const uint8_t*)(x))[1] << 16) | \
+                     (((const uint8_t*)(x))[2] <<  8) | \
+                      ((const uint8_t*)(x))[3])
+#define AV_WB32(p, d) do { \
+                    ((uint8_t*)(p))[3] = (d); \
+                    ((uint8_t*)(p))[2] = (d)>>8; \
+                    ((uint8_t*)(p))[1] = (d)>>16; \
+                    ((uint8_t*)(p))[0] = (d)>>24; } while(0)
+
+#if defined(ALT_BITSTREAM_READER_LE) && !defined(ALT_BITSTREAM_READER)
+#   define ALT_BITSTREAM_READER
+#endif
+
+//#define ALT_BITSTREAM_WRITER
+//#define ALIGNED_BITSTREAM_WRITER
+#if !defined(LIBMPEG2_BITSTREAM_READER) && !defined(A32_BITSTREAM_READER) && !defined(ALT_BITSTREAM_READER)
+#   if defined(ARCH_ARM)
+#       define A32_BITSTREAM_READER
+#   else
+#       define ALT_BITSTREAM_READER
+//#define LIBMPEG2_BITSTREAM_READER
+//#define A32_BITSTREAM_READER
+#   endif
+#endif
+
+extern const uint8_t ff_reverse[256];
+
+#if defined(ARCH_X86)
+// avoid +32 for shift optimization (gcc should do that ...)
+static inline  int32_t NEG_SSR32( int32_t a, int8_t s){
+    __asm__ ("sarl %1, %0\n\t"
+         : "+r" (a)
+         : "ic" ((uint8_t)(-s))
+    );
+    return a;
+}
+static inline uint32_t NEG_USR32(uint32_t a, int8_t s){
+    __asm__ ("shrl %1, %0\n\t"
+         : "+r" (a)
+         : "ic" ((uint8_t)(-s))
+    );
+    return a;
+}
+#else
+#    define NEG_SSR32(a,s) ((( int32_t)(a))>>(32-(s)))
+#    define NEG_USR32(a,s) (((uint32_t)(a))>>(32-(s)))
+#endif
+
+/* bit output */
+
+/* buf and buf_end must be present and used by every alternative writer. */
+typedef struct PutBitContext {
+#ifdef ALT_BITSTREAM_WRITER
+    uint8_t *buf, *buf_end;
+    int index;
+#else
+    uint32_t bit_buf;
+    int bit_left;
+    uint8_t *buf, *buf_ptr, *buf_end;
+#endif
+    int size_in_bits;
+} PutBitContext;
+
+static inline void init_put_bits(PutBitContext *s, uint8_t *buffer, int buffer_size)
+{
+    if(buffer_size < 0) {
+        buffer_size = 0;
+        buffer = NULL;
+    }
+
+    s->size_in_bits= 8*buffer_size;
+    s->buf = buffer;
+    s->buf_end = s->buf + buffer_size;
+#ifdef ALT_BITSTREAM_WRITER
+    s->index=0;
+    ((uint32_t*)(s->buf))[0]=0;
+//    memset(buffer, 0, buffer_size);
+#else
+    s->buf_ptr = s->buf;
+    s->bit_left=32;
+    s->bit_buf=0;
+#endif
+}
+
+/* return the number of bits output */
+static inline int put_bits_count(PutBitContext *s)
+{
+#ifdef ALT_BITSTREAM_WRITER
+    return s->index;
+#else
+    return (s->buf_ptr - s->buf) * 8 + 32 - s->bit_left;
+#endif
+}
+
+/* pad the end of the output stream with zeros */
+static inline void flush_put_bits(PutBitContext *s)
+{
+#ifdef ALT_BITSTREAM_WRITER
+    align_put_bits(s);
+#else
+#ifndef BITSTREAM_WRITER_LE
+    s->bit_buf<<= s->bit_left;
+#endif
+    while (s->bit_left < 32) {
+        /* XXX: should test end of buffer */
+#ifdef BITSTREAM_WRITER_LE
+        *s->buf_ptr++=s->bit_buf;
+        s->bit_buf>>=8;
+#else
+        *s->buf_ptr++=s->bit_buf >> 24;
+        s->bit_buf<<=8;
+#endif
+        s->bit_left+=8;
+    }
+    s->bit_left=32;
+    s->bit_buf=0;
+#endif
+}
+
+void align_put_bits(PutBitContext *s);
+void ff_put_string(PutBitContext * pbc, const char *s, int put_zero);
+void ff_copy_bits(PutBitContext *pb, const uint8_t *src, int length);
+
+/* bit input */
+/* buffer, buffer_end and size_in_bits must be present and used by every reader */
+typedef struct GetBitContext {
+    const uint8_t *buffer, *buffer_end;
+#ifdef ALT_BITSTREAM_READER
+    int index;
+#elif defined LIBMPEG2_BITSTREAM_READER
+    uint8_t *buffer_ptr;
+    uint32_t cache;
+    int bit_count;
+#elif defined A32_BITSTREAM_READER
+    uint32_t *buffer_ptr;
+    uint32_t cache0;
+    uint32_t cache1;
+    int bit_count;
+#endif
+    int size_in_bits;
+} GetBitContext;
+
+#define VLC_TYPE int16_t
+
+typedef struct VLC {
+    int bits;
+    VLC_TYPE (*table)[2]; ///< code, bits
+    int table_size, table_allocated;
+} VLC;
+
+typedef struct RL_VLC_ELEM {
+    int16_t level;
+    int8_t len;
+    uint8_t run;
+} RL_VLC_ELEM;
+
+#ifndef ALT_BITSTREAM_WRITER
+static inline void put_bits(PutBitContext *s, int n, unsigned int value)
+{
+    unsigned int bit_buf;
+    int bit_left;
+
+    //    printf("put_bits=%d %x\n", n, value);
+    //assert(n == 32 || value < (1U << n));
+
+    bit_buf = s->bit_buf;
+    bit_left = s->bit_left;
+
+    //    printf("n=%d value=%x cnt=%d buf=%x\n", n, value, bit_cnt, bit_buf);
+    /* XXX: optimize */
+#ifdef BITSTREAM_WRITER_LE
+    bit_buf |= value << (32 - bit_left);
+    if (n >= bit_left) {
+#if !HAVE_FAST_UNALIGNED
+        if (3 & (intptr_t) s->buf_ptr) {
+            AV_WL32(s->buf_ptr, bit_buf);
+        } else
+#endif
+        *(uint32_t *)s->buf_ptr = le2me_32(bit_buf);
+        s->buf_ptr+=4;
+        bit_buf = (bit_left==32)?0:value >> bit_left;
+        bit_left+=32;
+    }
+    bit_left-=n;
+#else
+    if (n < bit_left) {
+        bit_buf = (bit_buf<<n) | value;
+        bit_left-=n;
+    } else {
+        bit_buf<<=bit_left;
+        bit_buf |= value >> (n - bit_left);
+#if !defined(HAVE_FAST_UNALIGNED)
+        if (3 & (intptr_t) s->buf_ptr) {
+            AV_WB32(s->buf_ptr, bit_buf);
+        } else
+#endif
+        *(uint32_t *)s->buf_ptr = be2me_32(bit_buf);
+        //printf("bitbuf = %08x\n", bit_buf);
+        s->buf_ptr+=4;
+        bit_left+=32 - n;
+        bit_buf = value;
+    }
+#endif
+
+    s->bit_buf = bit_buf;
+    s->bit_left = bit_left;
+}
+#endif
+
+
+#ifdef ALT_BITSTREAM_WRITER
+static inline void put_bits(PutBitContext *s, int n, unsigned int value)
+{
+#    ifdef ALIGNED_BITSTREAM_WRITER
+#        if ARCH_X86
+    __asm__ volatile(
+        "movl %0, %%ecx                 \n\t"
+        "xorl %%eax, %%eax              \n\t"
+        "shrdl %%cl, %1, %%eax          \n\t"
+        "shrl %%cl, %1                  \n\t"
+        "movl %0, %%ecx                 \n\t"
+        "shrl $3, %%ecx                 \n\t"
+        "andl $0xFFFFFFFC, %%ecx        \n\t"
+        "bswapl %1                      \n\t"
+        "orl %1, (%2, %%ecx)            \n\t"
+        "bswapl %%eax                   \n\t"
+        "addl %3, %0                    \n\t"
+        "movl %%eax, 4(%2, %%ecx)       \n\t"
+        : "=&r" (s->index), "=&r" (value)
+        : "r" (s->buf), "r" (n), "0" (s->index), "1" (value<<(-n))
+        : "%eax", "%ecx"
+    );
+#        else
+    int index= s->index;
+    uint32_t *ptr= ((uint32_t *)s->buf)+(index>>5);
+
+    value<<= 32-n;
+
+    ptr[0] |= be2me_32(value>>(index&31));
+    ptr[1]  = be2me_32(value<<(32-(index&31)));
+//if(n>24) printf("%d %d\n", n, value);
+    index+= n;
+    s->index= index;
+#        endif
+#    else //ALIGNED_BITSTREAM_WRITER
+#        if ARCH_X86
+    __asm__ volatile(
+        "movl $7, %%ecx                 \n\t"
+        "andl %0, %%ecx                 \n\t"
+        "addl %3, %%ecx                 \n\t"
+        "negl %%ecx                     \n\t"
+        "shll %%cl, %1                  \n\t"
+        "bswapl %1                      \n\t"
+        "movl %0, %%ecx                 \n\t"
+        "shrl $3, %%ecx                 \n\t"
+        "orl %1, (%%ecx, %2)            \n\t"
+        "addl %3, %0                    \n\t"
+        "movl $0, 4(%%ecx, %2)          \n\t"
+        : "=&r" (s->index), "=&r" (value)
+        : "r" (s->buf), "r" (n), "0" (s->index), "1" (value)
+        : "%ecx"
+    );
+#        else
+    int index= s->index;
+    uint32_t *ptr= (uint32_t*)(((uint8_t *)s->buf)+(index>>3));
+
+    ptr[0] |= be2me_32(value<<(32-n-(index&7) ));
+    ptr[1] = 0;
+//if(n>24) printf("%d %d\n", n, value);
+    index+= n;
+    s->index= index;
+#        endif
+#    endif //!ALIGNED_BITSTREAM_WRITER
+}
+#endif
+
+static inline void put_sbits(PutBitContext *pb, int bits, int32_t val)
+{
+    //assert(bits >= 0 && bits <= 31);
+
+    put_bits(pb, bits, val & ((1<<bits)-1));
+}
+
+
+static inline uint8_t* pbBufPtr(PutBitContext *s)
+{
+#ifdef ALT_BITSTREAM_WRITER
+        return s->buf + (s->index>>3);
+#else
+        return s->buf_ptr;
+#endif
+}
+
+/**
+ *
+ * PutBitContext must be flushed & aligned to a byte boundary before calling this.
+ */
+static inline void skip_put_bytes(PutBitContext *s, int n){
+        //assert((put_bits_count(s)&7)==0);
+#ifdef ALT_BITSTREAM_WRITER
+        FIXME may need some cleaning of the buffer
+        s->index += n<<3;
+#else
+        //assert(s->bit_left==32);
+        s->buf_ptr += n;
+#endif
+}
+
+/**
+ * Skips the given number of bits.
+ * Must only be used if the actual values in the bitstream do not matter.
+ */
+static inline void skip_put_bits(PutBitContext *s, int n){
+#ifdef ALT_BITSTREAM_WRITER
+    s->index += n;
+#else
+    s->bit_left -= n;
+    s->buf_ptr-= s->bit_left>>5;
+    s->bit_left &= 31;
+#endif
+}
+
+/**
+ * Changes the end of the buffer.
+ */
+static inline void set_put_bits_buffer_size(PutBitContext *s, int size){
+    s->buf_end= s->buf + size;
+}
+
+/* Bitstream reader API docs:
+name
+    arbitrary name which is used as prefix for the internal variables
+
+gb
+    getbitcontext
+
+OPEN_READER(name, gb)
+    loads gb into local variables
+
+CLOSE_READER(name, gb)
+    stores local vars in gb
+
+UPDATE_CACHE(name, gb)
+    refills the internal cache from the bitstream
+    after this call at least MIN_CACHE_BITS will be available,
+
+GET_CACHE(name, gb)
+    will output the contents of the internal cache, next bit is MSB of 32 or 64 bit (FIXME 64bit)
+
+SHOW_UBITS(name, gb, num)
+    will return the next num bits
+
+SHOW_SBITS(name, gb, num)
+    will return the next num bits and do sign extension
+
+SKIP_BITS(name, gb, num)
+    will skip over the next num bits
+    note, this is equivalent to SKIP_CACHE; SKIP_COUNTER
+
+SKIP_CACHE(name, gb, num)
+    will remove the next num bits from the cache (note SKIP_COUNTER MUST be called before UPDATE_CACHE / CLOSE_READER)
+
+SKIP_COUNTER(name, gb, num)
+    will increment the internal bit counter (see SKIP_CACHE & SKIP_BITS)
+
+LAST_SKIP_CACHE(name, gb, num)
+    will remove the next num bits from the cache if it is needed for UPDATE_CACHE otherwise it will do nothing
+
+LAST_SKIP_BITS(name, gb, num)
+    is equivalent to SKIP_LAST_CACHE; SKIP_COUNTER
+
+for examples see get_bits, show_bits, skip_bits, get_vlc
+*/
+
+#ifdef ALT_BITSTREAM_READER
+#   define MIN_CACHE_BITS 25
+
+#   define OPEN_READER(name, gb)\
+        int name##_index= (gb)->index;\
+        int name##_cache= 0;\
+
+#   define CLOSE_READER(name, gb)\
+        (gb)->index= name##_index;\
+
+# ifdef ALT_BITSTREAM_READER_LE
+#   define UPDATE_CACHE(name, gb)\
+        name##_cache= AV_RL32( ((const uint8_t *)(gb)->buffer)+(name##_index>>3) ) >> (name##_index&0x07);\
+
+#   define SKIP_CACHE(name, gb, num)\
+        name##_cache >>= (num);
+# else
+#   define UPDATE_CACHE(name, gb)\
+        name##_cache= AV_RB32( ((const uint8_t *)(gb)->buffer)+(name##_index>>3) ) << (name##_index&0x07);\
+
+#   define SKIP_CACHE(name, gb, num)\
+        name##_cache <<= (num);
+# endif
+
+// FIXME name?
+#   define SKIP_COUNTER(name, gb, num)\
+        name##_index += (num);\
+
+#   define SKIP_BITS(name, gb, num)\
+        {\
+            SKIP_CACHE(name, gb, num)\
+            SKIP_COUNTER(name, gb, num)\
+        }\
+
+#   define LAST_SKIP_BITS(name, gb, num) SKIP_COUNTER(name, gb, num)
+#   define LAST_SKIP_CACHE(name, gb, num) ;
+
+# ifdef ALT_BITSTREAM_READER_LE
+#   define SHOW_UBITS(name, gb, num)\
+        ((name##_cache) & (NEG_USR32(0xffffffff,num)))
+
+#   define SHOW_SBITS(name, gb, num)\
+        NEG_SSR32((name##_cache)<<(32-(num)), num)
+# else
+#   define SHOW_UBITS(name, gb, num)\
+        NEG_USR32(name##_cache, num)
+
+#   define SHOW_SBITS(name, gb, num)\
+        NEG_SSR32(name##_cache, num)
+# endif
+
+#   define GET_CACHE(name, gb)\
+        ((uint32_t)name##_cache)
+
+static inline int get_bits_count(GetBitContext *s){
+    return s->index;
+}
+
+static inline void skip_bits_long(GetBitContext *s, int n){
+    s->index += n;
+}
+
+#elif defined LIBMPEG2_BITSTREAM_READER
+//libmpeg2 like reader
+
+#   define MIN_CACHE_BITS 17
+
+#   define OPEN_READER(name, gb)\
+        int name##_bit_count=(gb)->bit_count;\
+        int name##_cache= (gb)->cache;\
+        uint8_t * name##_buffer_ptr=(gb)->buffer_ptr;\
+
+#   define CLOSE_READER(name, gb)\
+        (gb)->bit_count= name##_bit_count;\
+        (gb)->cache= name##_cache;\
+        (gb)->buffer_ptr= name##_buffer_ptr;\
+
+#   define UPDATE_CACHE(name, gb)\
+    if(name##_bit_count >= 0){\
+        name##_cache+= AV_RB16(name##_buffer_ptr) << name##_bit_count; \
+        name##_buffer_ptr+=2;\
+        name##_bit_count-= 16;\
+    }\
+
+#   define SKIP_CACHE(name, gb, num)\
+        name##_cache <<= (num);\
+
+#   define SKIP_COUNTER(name, gb, num)\
+        name##_bit_count += (num);\
+
+#   define SKIP_BITS(name, gb, num)\
+        {\
+            SKIP_CACHE(name, gb, num)\
+            SKIP_COUNTER(name, gb, num)\
+        }\
+
+#   define LAST_SKIP_BITS(name, gb, num) SKIP_BITS(name, gb, num)
+#   define LAST_SKIP_CACHE(name, gb, num) SKIP_CACHE(name, gb, num)
+
+#   define SHOW_UBITS(name, gb, num)\
+        NEG_USR32(name##_cache, num)
+
+#   define SHOW_SBITS(name, gb, num)\
+        NEG_SSR32(name##_cache, num)
+
+#   define GET_CACHE(name, gb)\
+        ((uint32_t)name##_cache)
+
+static inline int get_bits_count(GetBitContext *s){
+    return (s->buffer_ptr - s->buffer)*8 - 16 + s->bit_count;
+}
+
+static inline void skip_bits_long(GetBitContext *s, int n){
+    OPEN_READER(re, s)
+    re_bit_count += n;
+    re_buffer_ptr += 2*(re_bit_count>>4);
+    re_bit_count &= 15;
+    re_cache = ((re_buffer_ptr[-2]<<8) + re_buffer_ptr[-1]) << (16+re_bit_count);
+    UPDATE_CACHE(re, s)
+    CLOSE_READER(re, s)
+}
+
+#elif defined A32_BITSTREAM_READER
+
+#   define MIN_CACHE_BITS 32
+
+#   define OPEN_READER(name, gb)\
+        int name##_bit_count=(gb)->bit_count;\
+        uint32_t name##_cache0= (gb)->cache0;\
+        uint32_t name##_cache1= (gb)->cache1;\
+        uint32_t * name##_buffer_ptr=(gb)->buffer_ptr;\
+
+#   define CLOSE_READER(name, gb)\
+        (gb)->bit_count= name##_bit_count;\
+        (gb)->cache0= name##_cache0;\
+        (gb)->cache1= name##_cache1;\
+        (gb)->buffer_ptr= name##_buffer_ptr;\
+
+#   define UPDATE_CACHE(name, gb)\
+    if(name##_bit_count > 0){\
+        const uint32_t next= be2me_32( *name##_buffer_ptr );\
+        name##_cache0 |= NEG_USR32(next,name##_bit_count);\
+        name##_cache1 |= next<<name##_bit_count;\
+        name##_buffer_ptr++;\
+        name##_bit_count-= 32;\
+    }\
+
+#if ARCH_X86
+#   define SKIP_CACHE(name, gb, num)\
+        __asm__(\
+            "shldl %2, %1, %0          \n\t"\
+            "shll %2, %1               \n\t"\
+            : "+r" (name##_cache0), "+r" (name##_cache1)\
+            : "Ic" ((uint8_t)(num))\
+           );
+#else
+#   define SKIP_CACHE(name, gb, num)\
+        name##_cache0 <<= (num);\
+        name##_cache0 |= NEG_USR32(name##_cache1,num);\
+        name##_cache1 <<= (num);
+#endif
+
+#   define SKIP_COUNTER(name, gb, num)\
+        name##_bit_count += (num);\
+
+#   define SKIP_BITS(name, gb, num)\
+        {\
+            SKIP_CACHE(name, gb, num)\
+            SKIP_COUNTER(name, gb, num)\
+        }\
+
+#   define LAST_SKIP_BITS(name, gb, num) SKIP_BITS(name, gb, num)
+#   define LAST_SKIP_CACHE(name, gb, num) SKIP_CACHE(name, gb, num)
+
+#   define SHOW_UBITS(name, gb, num)\
+        NEG_USR32(name##_cache0, num)
+
+#   define SHOW_SBITS(name, gb, num)\
+        NEG_SSR32(name##_cache0, num)
+
+#   define GET_CACHE(name, gb)\
+        (name##_cache0)
+
+static inline int get_bits_count(GetBitContext *s){
+    return ((uint8_t*)s->buffer_ptr - s->buffer)*8 - 32 + s->bit_count;
+}
+
+static inline void skip_bits_long(GetBitContext *s, int n){
+    OPEN_READER(re, s)
+    re_bit_count += n;
+    re_buffer_ptr += re_bit_count>>5;
+    re_bit_count &= 31;
+    re_cache0 = be2me_32( re_buffer_ptr[-1] ) << re_bit_count;
+    re_cache1 = 0;
+    UPDATE_CACHE(re, s)
+    CLOSE_READER(re, s)
+}
+
+#endif
+
+/**
+ * read mpeg1 dc style vlc (sign bit + mantisse with no MSB).
+ * if MSB not set it is negative
+ * @param n length in bits
+ * @author BERO
+ */
+static inline int get_xbits(GetBitContext *s, int n){
+    register int sign;
+    register int32_t cache;
+    OPEN_READER(re, s)
+    UPDATE_CACHE(re, s)
+    cache = GET_CACHE(re,s);
+    sign=(~cache)>>31;
+    LAST_SKIP_BITS(re, s, n)
+    CLOSE_READER(re, s)
+    return (NEG_USR32(sign ^ cache, n) ^ sign) - sign;
+}
+
+static inline int get_sbits(GetBitContext *s, int n){
+    register int tmp;
+    OPEN_READER(re, s)
+    UPDATE_CACHE(re, s)
+    tmp= SHOW_SBITS(re, s, n);
+    LAST_SKIP_BITS(re, s, n)
+    CLOSE_READER(re, s)
+    return tmp;
+}
+
+/**
+ * reads 1-17 bits.
+ * Note, the alt bitstream reader can read up to 25 bits, but the libmpeg2 reader can't
+ */
+static inline unsigned int get_bits(GetBitContext *s, int n){
+    register int tmp;
+    OPEN_READER(re, s)
+    UPDATE_CACHE(re, s)
+    tmp= SHOW_UBITS(re, s, n);
+    LAST_SKIP_BITS(re, s, n)
+    CLOSE_READER(re, s)
+    return tmp;
+}
+
+/**
+ * shows 1-17 bits.
+ * Note, the alt bitstream reader can read up to 25 bits, but the libmpeg2 reader can't
+ */
+static inline unsigned int show_bits(GetBitContext *s, int n){
+    register int tmp;
+    OPEN_READER(re, s)
+    UPDATE_CACHE(re, s)
+    tmp= SHOW_UBITS(re, s, n);
+//    CLOSE_READER(re, s)
+    return tmp;
+}
+
+static inline void skip_bits(GetBitContext *s, int n){
+ //Note gcc seems to optimize this to s->index+=n for the ALT_READER :))
+    OPEN_READER(re, s)
+    UPDATE_CACHE(re, s)
+    LAST_SKIP_BITS(re, s, n)
+    CLOSE_READER(re, s)
+}
+
+static inline unsigned int get_bits1(GetBitContext *s){
+#ifdef ALT_BITSTREAM_READER
+    int index= s->index;
+    uint8_t result= s->buffer[ index>>3 ];
+#ifdef ALT_BITSTREAM_READER_LE
+    result>>= (index&0x07);
+    result&= 1;
+#else
+    result<<= (index&0x07);
+    result>>= 8 - 1;
+#endif
+    index++;
+    s->index= index;
+
+    return result;
+#else
+    return get_bits(s, 1);
+#endif
+}
+
+static inline unsigned int show_bits1(GetBitContext *s){
+    return show_bits(s, 1);
+}
+
+static inline void skip_bits1(GetBitContext *s){
+    skip_bits(s, 1);
+}
+
+/**
+ * reads 0-32 bits.
+ */
+static inline unsigned int get_bits_long(GetBitContext *s, int n){
+    if(n<=17) return get_bits(s, n);
+    else{
+#ifdef ALT_BITSTREAM_READER_LE
+        int ret= get_bits(s, 16);
+        return ret | (get_bits(s, n-16) << 16);
+#else
+        int ret= get_bits(s, 16) << (n-16);
+        return ret | get_bits(s, n-16);
+#endif
+    }
+}
+
+#if 0
+/**
+ * reads 0-32 bits as a signed integer.
+ */
+static inline int get_sbits_long(GetBitContext *s, int n) {
+    return sign_extend(get_bits_long(s, n), n);
+}
+#endif
+
+/**
+ * shows 0-32 bits.
+ */
+static inline unsigned int show_bits_long(GetBitContext *s, int n){
+    if(n<=17) return show_bits(s, n);
+    else{
+        GetBitContext gb= *s;
+        return get_bits_long(&gb, n);
+    }
+}
+
+#if 0
+static inline int check_marker(GetBitContext *s, const char *msg)
+{
+    int bit= get_bits1(s);
+    if(!bit)
+        printf("Marker bit missing %s\n", msg);
+
+    return bit;
+}
+#endif
+
+/**
+ * init GetBitContext.
+ * @param buffer bitstream buffer, must be FF_INPUT_BUFFER_PADDING_SIZE bytes larger then the actual read bits
+ * because some optimized bitstream readers read 32 or 64 bit at once and could read over the end
+ * @param bit_size the size of the buffer in bits
+ */
+static inline void init_get_bits(GetBitContext *s,
+                   const uint8_t *buffer, int bit_size)
+{
+    int buffer_size= (bit_size+7)>>3;
+    if(buffer_size < 0 || bit_size < 0) {
+        buffer_size = bit_size = 0;
+        buffer = NULL;
+    }
+
+    s->buffer= buffer;
+    s->size_in_bits= bit_size;
+    s->buffer_end= buffer + buffer_size;
+#ifdef ALT_BITSTREAM_READER
+    s->index=0;
+#elif defined LIBMPEG2_BITSTREAM_READER
+    s->buffer_ptr = (uint8_t*)((intptr_t)buffer&(~1));
+    s->bit_count = 16 + 8*((intptr_t)buffer&1);
+    skip_bits_long(s, 0);
+#elif defined A32_BITSTREAM_READER
+    s->buffer_ptr = (uint32_t*)((intptr_t)buffer&(~3));
+    s->bit_count = 32 + 8*((intptr_t)buffer&3);
+    skip_bits_long(s, 0);
+#endif
+}
+
+static inline void align_get_bits(GetBitContext *s)
+{
+    int n= (-get_bits_count(s)) & 7;
+    if(n) skip_bits(s, n);
+}
+
+#define init_vlc(vlc, nb_bits, nb_codes,\
+                 bits, bits_wrap, bits_size,\
+                 codes, codes_wrap, codes_size,\
+                 flags)\
+        init_vlc_sparse(vlc, nb_bits, nb_codes,\
+                 bits, bits_wrap, bits_size,\
+                 codes, codes_wrap, codes_size,\
+                 NULL, 0, 0, flags)
+
+int init_vlc_sparse(VLC *vlc, int nb_bits, int nb_codes,
+             const void *bits, int bits_wrap, int bits_size,
+             const void *codes, int codes_wrap, int codes_size,
+             const void *symbols, int symbols_wrap, int symbols_size,
+             int flags);
+#define INIT_VLC_USE_STATIC 1 ///< VERY strongly deprecated and forbidden
+#define INIT_VLC_LE         2
+#define INIT_VLC_USE_NEW_STATIC 4
+void free_vlc(VLC *vlc);
+
+#define INIT_VLC_STATIC(vlc, bits, a,b,c,d,e,f,g, static_size)\
+{\
+    static VLC_TYPE table[static_size][2];\
+    (vlc)->table= table;\
+    (vlc)->table_allocated= static_size;\
+    init_vlc(vlc, bits, a,b,c,d,e,f,g, INIT_VLC_USE_NEW_STATIC);\
+}
+
+
+/**
+ *
+ * if the vlc code is invalid and max_depth=1 than no bits will be removed
+ * if the vlc code is invalid and max_depth>1 than the number of bits removed
+ * is undefined
+ */
+#define GET_VLC(code, name, gb, table, bits, max_depth)\
+{\
+    int n, index, nb_bits;\
+\
+    index= SHOW_UBITS(name, gb, bits);\
+    code = table[index][0];\
+    n    = table[index][1];\
+\
+    if(max_depth > 1 && n < 0){\
+        LAST_SKIP_BITS(name, gb, bits)\
+        UPDATE_CACHE(name, gb)\
+\
+        nb_bits = -n;\
+\
+        index= SHOW_UBITS(name, gb, nb_bits) + code;\
+        code = table[index][0];\
+        n    = table[index][1];\
+        if(max_depth > 2 && n < 0){\
+            LAST_SKIP_BITS(name, gb, nb_bits)\
+            UPDATE_CACHE(name, gb)\
+\
+            nb_bits = -n;\
+\
+            index= SHOW_UBITS(name, gb, nb_bits) + code;\
+            code = table[index][0];\
+            n    = table[index][1];\
+        }\
+    }\
+    SKIP_BITS(name, gb, n)\
+}
+
+#define GET_RL_VLC(level, run, name, gb, table, bits, max_depth, need_update)\
+{\
+    int n, index, nb_bits;\
+\
+    index= SHOW_UBITS(name, gb, bits);\
+    level = table[index].level;\
+    n     = table[index].len;\
+\
+    if(max_depth > 1 && n < 0){\
+        SKIP_BITS(name, gb, bits)\
+        if(need_update){\
+            UPDATE_CACHE(name, gb)\
+        }\
+\
+        nb_bits = -n;\
+\
+        index= SHOW_UBITS(name, gb, nb_bits) + level;\
+        level = table[index].level;\
+        n     = table[index].len;\
+    }\
+    run= table[index].run;\
+    SKIP_BITS(name, gb, n)\
+}
+
+
+/**
+ * parses a vlc code, faster then get_vlc()
+ * @param bits is the number of bits which will be read at once, must be
+ *             identical to nb_bits in init_vlc()
+ * @param max_depth is the number of times bits bits must be read to completely
+ *                  read the longest vlc code
+ *                  = (max_vlc_length + bits - 1) / bits
+ */
+static inline int get_vlc2(GetBitContext *s, VLC_TYPE (*table)[2],
+                                  int bits, int max_depth)
+{
+    int code;
+
+    OPEN_READER(re, s)
+    UPDATE_CACHE(re, s)
+
+    GET_VLC(code, re, s, table, bits, max_depth)
+
+    CLOSE_READER(re, s)
+    return code;
+}
+
+//#define TRACE
+
+#ifdef TRACE
+static inline void print_bin(int bits, int n){
+    int i;
+
+    for(i=n-1; i>=0; i--){
+        printf("%d", (bits>>i)&1);
+    }
+    for(i=n; i<24; i++)
+        printf(" ");
+}
+
+static inline int get_bits_trace(GetBitContext *s, int n, char *file, const char *func, int line){
+    int r= get_bits(s, n);
+
+    print_bin(r, n);
+    printf("%5d %2d %3d bit @%5d in %s %s:%d\n", r, n, r, get_bits_count(s)-n, file, func, line);
+    return r;
+}
+static inline int get_vlc_trace(GetBitContext *s, VLC_TYPE (*table)[2], int bits, int max_depth, char *file, const char *func, int line){
+    int show= show_bits(s, 24);
+    int pos= get_bits_count(s);
+    int r= get_vlc2(s, table, bits, max_depth);
+    int len= get_bits_count(s) - pos;
+    int bits2= show>>(24-len);
+
+    print_bin(bits2, len);
+
+    printf("%5d %2d %3d vlc @%5d in %s %s:%d\n", bits2, len, r, pos, file, func, line);
+    return r;
+}
+static inline int get_xbits_trace(GetBitContext *s, int n, char *file, const char *func, int line){
+    int show= show_bits(s, n);
+    int r= get_xbits(s, n);
+
+    print_bin(show, n);
+    printf("%5d %2d %3d xbt @%5d in %s %s:%d\n", show, n, r, get_bits_count(s)-n, file, func, line);
+    return r;
+}
+
+#define get_bits(s, n)  get_bits_trace(s, n, __FILE__, __PRETTY_FUNCTION__, __LINE__)
+#define get_bits1(s)    get_bits_trace(s, 1, __FILE__, __PRETTY_FUNCTION__, __LINE__)
+#define get_xbits(s, n) get_xbits_trace(s, n, __FILE__, __PRETTY_FUNCTION__, __LINE__)
+#define get_vlc(s, vlc)            get_vlc_trace(s, (vlc)->table, (vlc)->bits, 3, __FILE__, __PRETTY_FUNCTION__, __LINE__)
+#define get_vlc2(s, tab, bits, max) get_vlc_trace(s, tab, bits, max, __FILE__, __PRETTY_FUNCTION__, __LINE__)
+
+#define tprintf(p, ...) printf
+
+#else //TRACE
+#define tprintf(p, ...) {}
+#endif
+
+static inline int decode012(GetBitContext *gb){
+    int n;
+    n = get_bits1(gb);
+    if (n == 0)
+        return 0;
+    else
+        return get_bits1(gb) + 1;
+}
+
+static inline int decode210(GetBitContext *gb){
+    if (get_bits1(gb))
+        return 0;
+    else
+        return 2 - get_bits1(gb);
+}
+
+#endif /* BITSTREAM_H */

apps/codecs/libatrac/bswap.h

@@ -0,0 +1,150 @@
+/**
+ * @file bswap.h
+ * byte swap.
+ */
+
+#ifndef __BSWAP_H__
+#define __BSWAP_H__
+
+#ifdef HAVE_BYTESWAP_H
+#include <byteswap.h>
+#else
+
+#ifdef ROCKBOX
+#include "codecs.h"
+
+/* rockbox' optimised inline functions */
+#define bswap_16(x) swap16(x)
+#define bswap_32(x) swap32(x)
+
+static inline uint64_t ByteSwap64(uint64_t x)
+{
+    union { 
+        uint64_t ll;
+        struct {
+           uint32_t l,h;
+        } l;
+    } r;
+    r.l.l = bswap_32 (x);
+    r.l.h = bswap_32 (x>>32);
+    return r.ll;
+}
+#define bswap_64(x) ByteSwap64(x)
+
+#elif defined(ARCH_X86)
+static inline unsigned short ByteSwap16(unsigned short x)
+{
+  __asm("xchgb %b0,%h0"	:
+        "=q" (x)	:
+        "0" (x));
+    return x;
+}
+#define bswap_16(x) ByteSwap16(x)
+
+static inline unsigned int ByteSwap32(unsigned int x)
+{
+#if __CPU__ > 386
+ __asm("bswap	%0":
+      "=r" (x)     :
+#else
+ __asm("xchgb	%b0,%h0\n"
+      "	rorl	$16,%0\n"
+      "	xchgb	%b0,%h0":
+      "=q" (x)		:
+#endif
+      "0" (x));
+  return x;
+}
+#define bswap_32(x) ByteSwap32(x)
+
+static inline unsigned long long int ByteSwap64(unsigned long long int x)
+{
+  register union { __extension__ uint64_t __ll;
+          uint32_t __l[2]; } __x;
+  asm("xchgl	%0,%1":
+      "=r"(__x.__l[0]),"=r"(__x.__l[1]):
+      "0"(bswap_32((unsigned long)x)),"1"(bswap_32((unsigned long)(x>>32))));
+  return __x.__ll;
+}
+#define bswap_64(x) ByteSwap64(x)
+
+#elif defined(ARCH_SH4)
+
+static inline uint16_t ByteSwap16(uint16_t x) {
+	__asm__("swap.b %0,%0":"=r"(x):"0"(x));
+	return x;
+}
+
+static inline uint32_t ByteSwap32(uint32_t x) {
+	__asm__(
+	"swap.b %0,%0\n"
+	"swap.w %0,%0\n"
+	"swap.b %0,%0\n"
+	:"=r"(x):"0"(x));
+	return x;
+}
+
+#define bswap_16(x) ByteSwap16(x)
+#define bswap_32(x) ByteSwap32(x)
+
+static inline uint64_t ByteSwap64(uint64_t x)
+{
+    union { 
+        uint64_t ll;
+        struct {
+           uint32_t l,h;
+        } l;
+    } r;
+    r.l.l = bswap_32 (x);
+    r.l.h = bswap_32 (x>>32);
+    return r.ll;
+}
+#define bswap_64(x) ByteSwap64(x)
+
+#else
+
+#define bswap_16(x) (((x) & 0x00ff) << 8 | ((x) & 0xff00) >> 8)
+
+
+// code from bits/byteswap.h (C) 1997, 1998 Free Software Foundation, Inc.
+#define bswap_32(x) \
+     ((((x) & 0xff000000) >> 24) | (((x) & 0x00ff0000) >>  8) | \
+      (((x) & 0x0000ff00) <<  8) | (((x) & 0x000000ff) << 24))
+
+/*static inline uint64_t ByteSwap64(uint64_t x)
+{
+    union { 
+        uint64_t ll;
+        uint32_t l[2]; 
+    } w, r;
+    w.ll = x;
+    r.l[0] = bswap_32 (w.l[1]);
+    r.l[1] = bswap_32 (w.l[0]);
+    return r.ll;
+}*/
+#define bswap_64(x) ByteSwap64(x)
+
+#endif	/* !ARCH_X86 */
+
+#endif	/* !HAVE_BYTESWAP_H */
+
+// be2me ... BigEndian to MachineEndian
+// le2me ... LittleEndian to MachineEndian
+
+#ifdef WORDS_BIGENDIAN
+#define be2me_16(x) (x)
+#define be2me_32(x) (x)
+#define be2me_64(x) (x)
+#define le2me_16(x) bswap_16(x)
+#define le2me_32(x) bswap_32(x)
+#define le2me_64(x) bswap_64(x)
+#else
+#define be2me_16(x) bswap_16(x)
+#define be2me_32(x) bswap_32(x)
+#define be2me_64(x) bswap_64(x)
+#define le2me_16(x) (x)
+#define le2me_32(x) (x)
+#define le2me_64(x) (x)
+#endif
+
+#endif /* __BSWAP_H__ */

apps/codecs/libatrac/bytestream.h

@@ -0,0 +1,71 @@
+/*
+ * Bytestream functions
+ * copyright (c) 2006 Baptiste Coudurier <baptiste.coudurier@free.fr>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVCODEC_BYTESTREAM_H
+#define AVCODEC_BYTESTREAM_H
+
+#include <string.h>
+#include "libavutil/common.h"
+#include "libavutil/intreadwrite.h"
+
+#define DEF_T(type, name, bytes, read, write)                             \
+static av_always_inline type bytestream_get_ ## name(const uint8_t **b){\
+    (*b) += bytes;\
+    return read(*b - bytes);\
+}\
+static av_always_inline void bytestream_put_ ##name(uint8_t **b, const type value){\
+    write(*b, value);\
+    (*b) += bytes;\
+}
+
+#define DEF(name, bytes, read, write) \
+    DEF_T(unsigned int, name, bytes, read, write)
+#define DEF64(name, bytes, read, write) \
+    DEF_T(uint64_t, name, bytes, read, write)
+
+DEF64(le64, 8, AV_RL64, AV_WL64)
+DEF  (le32, 4, AV_RL32, AV_WL32)
+DEF  (le24, 3, AV_RL24, AV_WL24)
+DEF  (le16, 2, AV_RL16, AV_WL16)
+DEF64(be64, 8, AV_RB64, AV_WB64)
+DEF  (be32, 4, AV_RB32, AV_WB32)
+DEF  (be24, 3, AV_RB24, AV_WB24)
+DEF  (be16, 2, AV_RB16, AV_WB16)
+DEF  (byte, 1, AV_RB8 , AV_WB8 )
+
+#undef DEF
+#undef DEF64
+#undef DEF_T
+
+static av_always_inline unsigned int bytestream_get_buffer(const uint8_t **b, uint8_t *dst, unsigned int size)
+{
+    memcpy(dst, *b, size);
+    (*b) += size;
+    return size;
+}
+
+static av_always_inline void bytestream_put_buffer(uint8_t **b, const uint8_t *src, unsigned int size)
+{
+    memcpy(*b, src, size);
+    (*b) += size;
+}
+
+#endif /* AVCODEC_BYTESTREAM_H */

apps/codecs/libatrac/dsputil.h

@@ -0,0 +1,898 @@
+/*
+ * DSP utils
+ * Copyright (c) 2000, 2001, 2002 Fabrice Bellard
+ * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavcodec/dsputil.h
+ * DSP utils.
+ * note, many functions in here may use MMX which trashes the FPU state, it is
+ * absolutely necessary to call emms_c() between dsp & float/double code
+ */
+
+#ifndef AVCODEC_DSPUTIL_H
+#define AVCODEC_DSPUTIL_H
+
+#include "libavutil/intreadwrite.h"
+#include "avcodec.h"
+
+
+//#define DEBUG
+/* dct code */
+typedef short DCTELEM;
+typedef int DWTELEM;
+typedef short IDWTELEM;
+
+void fdct_ifast (DCTELEM *data);
+void fdct_ifast248 (DCTELEM *data);
+void ff_jpeg_fdct_islow (DCTELEM *data);
+void ff_fdct248_islow (DCTELEM *data);
+
+void j_rev_dct (DCTELEM *data);
+void j_rev_dct4 (DCTELEM *data);
+void j_rev_dct2 (DCTELEM *data);
+void j_rev_dct1 (DCTELEM *data);
+void ff_wmv2_idct_c(DCTELEM *data);
+
+void ff_fdct_mmx(DCTELEM *block);
+void ff_fdct_mmx2(DCTELEM *block);
+void ff_fdct_sse2(DCTELEM *block);
+
+void ff_h264_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride);
+void ff_h264_idct_add_c(uint8_t *dst, DCTELEM *block, int stride);
+void ff_h264_idct8_dc_add_c(uint8_t *dst, DCTELEM *block, int stride);
+void ff_h264_idct_dc_add_c(uint8_t *dst, DCTELEM *block, int stride);
+void ff_h264_lowres_idct_add_c(uint8_t *dst, int stride, DCTELEM *block);
+void ff_h264_lowres_idct_put_c(uint8_t *dst, int stride, DCTELEM *block);
+void ff_h264_idct_add16_c(uint8_t *dst, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
+void ff_h264_idct_add16intra_c(uint8_t *dst, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
+void ff_h264_idct8_add4_c(uint8_t *dst, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
+void ff_h264_idct_add8_c(uint8_t **dest, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
+
+void ff_vector_fmul_add_add_c(float *dst, const float *src0, const float *src1,
+                              const float *src2, int src3, int blocksize, int step);
+void ff_vector_fmul_window_c(float *dst, const float *src0, const float *src1,
+                             const float *win, float add_bias, int len);
+void ff_float_to_int16_c(int16_t *dst, const float *src, long len);
+void ff_float_to_int16_interleave_c(int16_t *dst, const float **src, long len, int channels);
+
+/* encoding scans */
+extern const uint8_t ff_alternate_horizontal_scan[64];
+extern const uint8_t ff_alternate_vertical_scan[64];
+extern const uint8_t ff_zigzag_direct[64];
+extern const uint8_t ff_zigzag248_direct[64];
+
+/* pixel operations */
+#define MAX_NEG_CROP 1024
+
+/* temporary */
+extern uint32_t ff_squareTbl[512];
+extern uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP];
+
+/* VP3 DSP functions */
+void ff_vp3_idct_c(DCTELEM *block/* align 16*/);
+void ff_vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
+void ff_vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
+
+void ff_vp3_v_loop_filter_c(uint8_t *src, int stride, int *bounding_values);
+void ff_vp3_h_loop_filter_c(uint8_t *src, int stride, int *bounding_values);
+
+/* VP6 DSP functions */
+void ff_vp6_filter_diag4_c(uint8_t *dst, uint8_t *src, int stride,
+                           const int16_t *h_weights, const int16_t *v_weights);
+
+/* 1/2^n downscaling functions from imgconvert.c */
+void ff_img_copy_plane(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
+void ff_shrink22(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
+void ff_shrink44(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
+void ff_shrink88(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
+
+void ff_gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
+              int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height);
+
+/* minimum alignment rules ;)
+If you notice errors in the align stuff, need more alignment for some ASM code
+for some CPU or need to use a function with less aligned data then send a mail
+to the ffmpeg-devel mailing list, ...
+
+!warning These alignments might not match reality, (missing attribute((align))
+stuff somewhere possible).
+I (Michael) did not check them, these are just the alignments which I think
+could be reached easily ...
+
+!future video codecs might need functions with less strict alignment
+*/
+
+/*
+void get_pixels_c(DCTELEM *block, const uint8_t *pixels, int line_size);
+void diff_pixels_c(DCTELEM *block, const uint8_t *s1, const uint8_t *s2, int stride);
+void put_pixels_clamped_c(const DCTELEM *block, uint8_t *pixels, int line_size);
+void add_pixels_clamped_c(const DCTELEM *block, uint8_t *pixels, int line_size);
+void clear_blocks_c(DCTELEM *blocks);
+*/
+
+/* add and put pixel (decoding) */
+// blocksizes for op_pixels_func are 8x4,8x8 16x8 16x16
+//h for op_pixels_func is limited to {width/2, width} but never larger than 16 and never smaller then 4
+typedef void (*op_pixels_func)(uint8_t *block/*align width (8 or 16)*/, const uint8_t *pixels/*align 1*/, int line_size, int h);
+typedef void (*tpel_mc_func)(uint8_t *block/*align width (8 or 16)*/, const uint8_t *pixels/*align 1*/, int line_size, int w, int h);
+typedef void (*qpel_mc_func)(uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);
+typedef void (*h264_chroma_mc_func)(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int srcStride, int h, int x, int y);
+typedef void (*h264_weight_func)(uint8_t *block, int stride, int log2_denom, int weight, int offset);
+typedef void (*h264_biweight_func)(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offset);
+
+#define DEF_OLD_QPEL(name)\
+void ff_put_        ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);\
+void ff_put_no_rnd_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);\
+void ff_avg_        ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);
+
+DEF_OLD_QPEL(qpel16_mc11_old_c)
+DEF_OLD_QPEL(qpel16_mc31_old_c)
+DEF_OLD_QPEL(qpel16_mc12_old_c)
+DEF_OLD_QPEL(qpel16_mc32_old_c)
+DEF_OLD_QPEL(qpel16_mc13_old_c)
+DEF_OLD_QPEL(qpel16_mc33_old_c)
+DEF_OLD_QPEL(qpel8_mc11_old_c)
+DEF_OLD_QPEL(qpel8_mc31_old_c)
+DEF_OLD_QPEL(qpel8_mc12_old_c)
+DEF_OLD_QPEL(qpel8_mc32_old_c)
+DEF_OLD_QPEL(qpel8_mc13_old_c)
+DEF_OLD_QPEL(qpel8_mc33_old_c)
+
+#define CALL_2X_PIXELS(a, b, n)\
+static void a(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
+    b(block  , pixels  , line_size, h);\
+    b(block+n, pixels+n, line_size, h);\
+}
+
+/* motion estimation */
+// h is limited to {width/2, width, 2*width} but never larger than 16 and never smaller then 2
+// although currently h<4 is not used as functions with width <8 are neither used nor implemented
+typedef int (*me_cmp_func)(void /*MpegEncContext*/ *s, uint8_t *blk1/*align width (8 or 16)*/, uint8_t *blk2/*align 1*/, int line_size, int h)/* __attribute__ ((const))*/;
+
+
+// for snow slices
+typedef struct slice_buffer_s slice_buffer;
+
+/**
+ * Scantable.
+ */
+typedef struct ScanTable{
+    const uint8_t *scantable;
+    uint8_t permutated[64];
+    uint8_t raster_end[64];
+#if ARCH_PPC
+                /** Used by dct_quantize_altivec to find last-non-zero */
+    DECLARE_ALIGNED(16, uint8_t, inverse[64]);
+#endif
+} ScanTable;
+
+void ff_init_scantable(uint8_t *, ScanTable *st, const uint8_t *src_scantable);
+
+void ff_emulated_edge_mc(uint8_t *buf, uint8_t *src, int linesize,
+                         int block_w, int block_h,
+                         int src_x, int src_y, int w, int h);
+
+/**
+ * DSPContext.
+ */
+typedef struct DSPContext {
+    /* pixel ops : interface with DCT */
+    void (*get_pixels)(DCTELEM *block/*align 16*/, const uint8_t *pixels/*align 8*/, int line_size);
+    void (*diff_pixels)(DCTELEM *block/*align 16*/, const uint8_t *s1/*align 8*/, const uint8_t *s2/*align 8*/, int stride);
+    void (*put_pixels_clamped)(const DCTELEM *block/*align 16*/, uint8_t *pixels/*align 8*/, int line_size);
+    void (*put_signed_pixels_clamped)(const DCTELEM *block/*align 16*/, uint8_t *pixels/*align 8*/, int line_size);
+    void (*add_pixels_clamped)(const DCTELEM *block/*align 16*/, uint8_t *pixels/*align 8*/, int line_size);
+    void (*add_pixels8)(uint8_t *pixels, DCTELEM *block, int line_size);
+    void (*add_pixels4)(uint8_t *pixels, DCTELEM *block, int line_size);
+    int (*sum_abs_dctelem)(DCTELEM *block/*align 16*/);
+    /**
+     * translational global motion compensation.
+     */
+    void (*gmc1)(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int srcStride, int h, int x16, int y16, int rounder);
+    /**
+     * global motion compensation.
+     */
+    void (*gmc )(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int ox, int oy,
+                    int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height);
+    void (*clear_block)(DCTELEM *block/*align 16*/);
+    void (*clear_blocks)(DCTELEM *blocks/*align 16*/);
+    int (*pix_sum)(uint8_t * pix, int line_size);
+    int (*pix_norm1)(uint8_t * pix, int line_size);
+// 16x16 8x8 4x4 2x2 16x8 8x4 4x2 8x16 4x8 2x4
+
+    me_cmp_func sad[6]; /* identical to pix_absAxA except additional void * */
+    me_cmp_func sse[6];
+    me_cmp_func hadamard8_diff[6];
+    me_cmp_func dct_sad[6];
+    me_cmp_func quant_psnr[6];
+    me_cmp_func bit[6];
+    me_cmp_func rd[6];
+    me_cmp_func vsad[6];
+    me_cmp_func vsse[6];
+    me_cmp_func nsse[6];
+    me_cmp_func w53[6];
+    me_cmp_func w97[6];
+    me_cmp_func dct_max[6];
+    me_cmp_func dct264_sad[6];
+
+    me_cmp_func me_pre_cmp[6];
+    me_cmp_func me_cmp[6];
+    me_cmp_func me_sub_cmp[6];
+    me_cmp_func mb_cmp[6];
+    me_cmp_func ildct_cmp[6]; //only width 16 used
+    me_cmp_func frame_skip_cmp[6]; //only width 8 used
+
+    int (*ssd_int8_vs_int16)(const int8_t *pix1, const int16_t *pix2,
+                             int size);
+
+    /**
+     * Halfpel motion compensation with rounding (a+b+1)>>1.
+     * this is an array[4][4] of motion compensation functions for 4
+     * horizontal blocksizes (8,16) and the 4 halfpel positions<br>
+     * *pixels_tab[ 0->16xH 1->8xH ][ xhalfpel + 2*yhalfpel ]
+     * @param block destination where the result is stored
+     * @param pixels source
+     * @param line_size number of bytes in a horizontal line of block
+     * @param h height
+     */
+    op_pixels_func put_pixels_tab[4][4];
+
+    /**
+     * Halfpel motion compensation with rounding (a+b+1)>>1.
+     * This is an array[4][4] of motion compensation functions for 4
+     * horizontal blocksizes (8,16) and the 4 halfpel positions<br>
+     * *pixels_tab[ 0->16xH 1->8xH ][ xhalfpel + 2*yhalfpel ]
+     * @param block destination into which the result is averaged (a+b+1)>>1
+     * @param pixels source
+     * @param line_size number of bytes in a horizontal line of block
+     * @param h height
+     */
+    op_pixels_func avg_pixels_tab[4][4];
+
+    /**
+     * Halfpel motion compensation with no rounding (a+b)>>1.
+     * this is an array[2][4] of motion compensation functions for 2
+     * horizontal blocksizes (8,16) and the 4 halfpel positions<br>
+     * *pixels_tab[ 0->16xH 1->8xH ][ xhalfpel + 2*yhalfpel ]
+     * @param block destination where the result is stored
+     * @param pixels source
+     * @param line_size number of bytes in a horizontal line of block
+     * @param h height
+     */
+    op_pixels_func put_no_rnd_pixels_tab[4][4];
+
+    /**
+     * Halfpel motion compensation with no rounding (a+b)>>1.
+     * this is an array[2][4] of motion compensation functions for 2
+     * horizontal blocksizes (8,16) and the 4 halfpel positions<br>
+     * *pixels_tab[ 0->16xH 1->8xH ][ xhalfpel + 2*yhalfpel ]
+     * @param block destination into which the result is averaged (a+b)>>1
+     * @param pixels source
+     * @param line_size number of bytes in a horizontal line of block
+     * @param h height
+     */
+    op_pixels_func avg_no_rnd_pixels_tab[4][4];
+
+    void (*put_no_rnd_pixels_l2[2])(uint8_t *block/*align width (8 or 16)*/, const uint8_t *a/*align 1*/, const uint8_t *b/*align 1*/, int line_size, int h);
+
+    /**
+     * Thirdpel motion compensation with rounding (a+b+1)>>1.
+     * this is an array[12] of motion compensation functions for the 9 thirdpe
+     * positions<br>
+     * *pixels_tab[ xthirdpel + 4*ythirdpel ]
+     * @param block destination where the result is stored
+     * @param pixels source
+     * @param line_size number of bytes in a horizontal line of block
+     * @param h height
+     */
+    tpel_mc_func put_tpel_pixels_tab[11]; //FIXME individual func ptr per width?
+    tpel_mc_func avg_tpel_pixels_tab[11]; //FIXME individual func ptr per width?
+
+    qpel_mc_func put_qpel_pixels_tab[2][16];
+    qpel_mc_func avg_qpel_pixels_tab[2][16];
+    qpel_mc_func put_no_rnd_qpel_pixels_tab[2][16];
+    qpel_mc_func avg_no_rnd_qpel_pixels_tab[2][16];
+    qpel_mc_func put_mspel_pixels_tab[8];
+
+    /**
+     * h264 Chroma MC
+     */
+    h264_chroma_mc_func put_h264_chroma_pixels_tab[3];
+    /* This is really one func used in VC-1 decoding */
+    h264_chroma_mc_func put_no_rnd_h264_chroma_pixels_tab[3];
+    h264_chroma_mc_func avg_h264_chroma_pixels_tab[3];
+
+    qpel_mc_func put_h264_qpel_pixels_tab[4][16];
+    qpel_mc_func avg_h264_qpel_pixels_tab[4][16];
+
+    qpel_mc_func put_2tap_qpel_pixels_tab[4][16];
+    qpel_mc_func avg_2tap_qpel_pixels_tab[4][16];
+
+    h264_weight_func weight_h264_pixels_tab[10];
+    h264_biweight_func biweight_h264_pixels_tab[10];
+
+    /* AVS specific */
+    qpel_mc_func put_cavs_qpel_pixels_tab[2][16];
+    qpel_mc_func avg_cavs_qpel_pixels_tab[2][16];
+    void (*cavs_filter_lv)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+    void (*cavs_filter_lh)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+    void (*cavs_filter_cv)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+    void (*cavs_filter_ch)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+    void (*cavs_idct8_add)(uint8_t *dst, DCTELEM *block, int stride);
+
+    me_cmp_func pix_abs[2][4];
+
+    /* huffyuv specific */
+    void (*add_bytes)(uint8_t *dst/*align 16*/, uint8_t *src/*align 16*/, int w);
+    void (*add_bytes_l2)(uint8_t *dst/*align 16*/, uint8_t *src1/*align 16*/, uint8_t *src2/*align 16*/, int w);
+    void (*diff_bytes)(uint8_t *dst/*align 16*/, uint8_t *src1/*align 16*/, uint8_t *src2/*align 1*/,int w);
+    /**
+     * subtract huffyuv's variant of median prediction
+     * note, this might read from src1[-1], src2[-1]
+     */
+    void (*sub_hfyu_median_prediction)(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w, int *left, int *left_top);
+    void (*add_hfyu_median_prediction)(uint8_t *dst, uint8_t *top, uint8_t *diff, int w, int *left, int *left_top);
+    /* this might write to dst[w] */
+    void (*add_png_paeth_prediction)(uint8_t *dst, uint8_t *src, uint8_t *top, int w, int bpp);
+    void (*bswap_buf)(uint32_t *dst, const uint32_t *src, int w);
+
+    void (*h264_v_loop_filter_luma)(uint8_t *pix/*align 16*/, int stride, int alpha, int beta, int8_t *tc0);
+    void (*h264_h_loop_filter_luma)(uint8_t *pix/*align 4 */, int stride, int alpha, int beta, int8_t *tc0);
+    /* v/h_loop_filter_luma_intra: align 16 */
+    void (*h264_v_loop_filter_luma_intra)(uint8_t *pix, int stride, int alpha, int beta);
+    void (*h264_h_loop_filter_luma_intra)(uint8_t *pix, int stride, int alpha, int beta);
+    void (*h264_v_loop_filter_chroma)(uint8_t *pix/*align 8*/, int stride, int alpha, int beta, int8_t *tc0);
+    void (*h264_h_loop_filter_chroma)(uint8_t *pix/*align 4*/, int stride, int alpha, int beta, int8_t *tc0);
+    void (*h264_v_loop_filter_chroma_intra)(uint8_t *pix/*align 8*/, int stride, int alpha, int beta);
+    void (*h264_h_loop_filter_chroma_intra)(uint8_t *pix/*align 8*/, int stride, int alpha, int beta);
+    // h264_loop_filter_strength: simd only. the C version is inlined in h264.c
+    void (*h264_loop_filter_strength)(int16_t bS[2][4][4], uint8_t nnz[40], int8_t ref[2][40], int16_t mv[2][40][2],
+                                      int bidir, int edges, int step, int mask_mv0, int mask_mv1, int field);
+
+    void (*h263_v_loop_filter)(uint8_t *src, int stride, int qscale);
+    void (*h263_h_loop_filter)(uint8_t *src, int stride, int qscale);
+
+    void (*h261_loop_filter)(uint8_t *src, int stride);
+
+    void (*x8_v_loop_filter)(uint8_t *src, int stride, int qscale);
+    void (*x8_h_loop_filter)(uint8_t *src, int stride, int qscale);
+
+    void (*vp3_v_loop_filter)(uint8_t *src, int stride, int *bounding_values);
+    void (*vp3_h_loop_filter)(uint8_t *src, int stride, int *bounding_values);
+
+    void (*vp6_filter_diag4)(uint8_t *dst, uint8_t *src, int stride,
+                             const int16_t *h_weights,const int16_t *v_weights);
+
+    /* assume len is a multiple of 4, and arrays are 16-byte aligned */
+    void (*vorbis_inverse_coupling)(float *mag, float *ang, int blocksize);
+    void (*ac3_downmix)(float (*samples)[256], float (*matrix)[2], int out_ch, int in_ch, int len);
+    /* no alignment needed */
+    void (*flac_compute_autocorr)(const int32_t *data, int len, int lag, double *autoc);
+    /* assume len is a multiple of 8, and arrays are 16-byte aligned */
+    void (*vector_fmul)(float *dst, const float *src, int len);
+    void (*vector_fmul_reverse)(float *dst, const float *src0, const float *src1, int len);
+    /* assume len is a multiple of 8, and src arrays are 16-byte aligned */
+    void (*vector_fmul_add_add)(float *dst, const float *src0, const float *src1, const float *src2, int src3, int len, int step);
+    /* assume len is a multiple of 4, and arrays are 16-byte aligned */
+    void (*vector_fmul_window)(float *dst, const float *src0, const float *src1, const float *win, float add_bias, int len);
+    /* assume len is a multiple of 8, and arrays are 16-byte aligned */
+    void (*int32_to_float_fmul_scalar)(float *dst, const int *src, float mul, int len);
+
+    /* C version: convert floats from the range [384.0,386.0] to ints in [-32768,32767]
+     * simd versions: convert floats from [-32768.0,32767.0] without rescaling and arrays are 16byte aligned */
+    void (*float_to_int16)(int16_t *dst, const float *src, long len);
+    void (*float_to_int16_interleave)(int16_t *dst, const float **src, long len, int channels);
+
+    /* (I)DCT */
+    void (*fdct)(DCTELEM *block/* align 16*/);
+    void (*fdct248)(DCTELEM *block/* align 16*/);
+
+    /* IDCT really*/
+    void (*idct)(DCTELEM *block/* align 16*/);
+
+    /**
+     * block -> idct -> clip to unsigned 8 bit -> dest.
+     * (-1392, 0, 0, ...) -> idct -> (-174, -174, ...) -> put -> (0, 0, ...)
+     * @param line_size size in bytes of a horizontal line of dest
+     */
+    void (*idct_put)(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
+
+    /**
+     * block -> idct -> add dest -> clip to unsigned 8 bit -> dest.
+     * @param line_size size in bytes of a horizontal line of dest
+     */
+    void (*idct_add)(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
+
+    /**
+     * idct input permutation.
+     * several optimized IDCTs need a permutated input (relative to the normal order of the reference
+     * IDCT)
+     * this permutation must be performed before the idct_put/add, note, normally this can be merged
+     * with the zigzag/alternate scan<br>
+     * an example to avoid confusion:
+     * - (->decode coeffs -> zigzag reorder -> dequant -> reference idct ->...)
+     * - (x -> referece dct -> reference idct -> x)
+     * - (x -> referece dct -> simple_mmx_perm = idct_permutation -> simple_idct_mmx -> x)
+     * - (->decode coeffs -> zigzag reorder -> simple_mmx_perm -> dequant -> simple_idct_mmx ->...)
+     */
+    uint8_t idct_permutation[64];
+    int idct_permutation_type;
+#define FF_NO_IDCT_PERM 1
+#define FF_LIBMPEG2_IDCT_PERM 2
+#define FF_SIMPLE_IDCT_PERM 3
+#define FF_TRANSPOSE_IDCT_PERM 4
+#define FF_PARTTRANS_IDCT_PERM 5
+#define FF_SSE2_IDCT_PERM 6
+
+    int (*try_8x8basis)(int16_t rem[64], int16_t weight[64], int16_t basis[64], int scale);
+    void (*add_8x8basis)(int16_t rem[64], int16_t basis[64], int scale);
+#define BASIS_SHIFT 16
+#define RECON_SHIFT 6
+
+    void (*draw_edges)(uint8_t *buf, int wrap, int width, int height, int w);
+#define EDGE_WIDTH 16
+
+    /* h264 functions */
+    /* NOTE!!! if you implement any of h264_idct8_add, h264_idct8_add4 then you must implement all of them
+       NOTE!!! if you implement any of h264_idct_add, h264_idct_add16, h264_idct_add16intra, h264_idct_add8 then you must implement all of them
+        The reason for above, is that no 2 out of one list may use a different permutation.
+    */
+    void (*h264_idct_add)(uint8_t *dst/*align 4*/, DCTELEM *block/*align 16*/, int stride);
+    void (*h264_idct8_add)(uint8_t *dst/*align 8*/, DCTELEM *block/*align 16*/, int stride);
+    void (*h264_idct_dc_add)(uint8_t *dst/*align 4*/, DCTELEM *block/*align 16*/, int stride);
+    void (*h264_idct8_dc_add)(uint8_t *dst/*align 8*/, DCTELEM *block/*align 16*/, int stride);
+    void (*h264_dct)(DCTELEM block[4][4]);
+    void (*h264_idct_add16)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
+    void (*h264_idct8_add4)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
+    void (*h264_idct_add8)(uint8_t **dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
+    void (*h264_idct_add16intra)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
+
+    /* snow wavelet */
+    void (*vertical_compose97i)(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, IDWTELEM *b3, IDWTELEM *b4, IDWTELEM *b5, int width);
+    void (*horizontal_compose97i)(IDWTELEM *b, int width);
+    void (*inner_add_yblock)(const uint8_t *obmc, const int obmc_stride, uint8_t * * block, int b_w, int b_h, int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8);
+
+    void (*prefetch)(void *mem, int stride, int h);
+
+    void (*shrink[4])(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
+
+    /* vc1 functions */
+    void (*vc1_inv_trans_8x8)(DCTELEM *b);
+    void (*vc1_inv_trans_8x4)(uint8_t *dest, int line_size, DCTELEM *block);
+    void (*vc1_inv_trans_4x8)(uint8_t *dest, int line_size, DCTELEM *block);
+    void (*vc1_inv_trans_4x4)(uint8_t *dest, int line_size, DCTELEM *block);
+    void (*vc1_v_overlap)(uint8_t* src, int stride);
+    void (*vc1_h_overlap)(uint8_t* src, int stride);
+    /* put 8x8 block with bicubic interpolation and quarterpel precision
+     * last argument is actually round value instead of height
+     */
+    op_pixels_func put_vc1_mspel_pixels_tab[16];
+
+    /* intrax8 functions */
+    void (*x8_spatial_compensation[12])(uint8_t *src , uint8_t *dst, int linesize);
+    void (*x8_setup_spatial_compensation)(uint8_t *src, uint8_t *dst, int linesize,
+           int * range, int * sum,  int edges);
+
+    /* ape functions */
+    /**
+     * Add contents of the second vector to the first one.
+     * @param len length of vectors, should be multiple of 16
+     */
+    void (*add_int16)(int16_t *v1/*align 16*/, int16_t *v2, int len);
+    /**
+     * Add contents of the second vector to the first one.
+     * @param len length of vectors, should be multiple of 16
+     */
+    void (*sub_int16)(int16_t *v1/*align 16*/, int16_t *v2, int len);
+    /**
+     * Calculate scalar product of two vectors.
+     * @param len length of vectors, should be multiple of 16
+     * @param shift number of bits to discard from product
+     */
+    int32_t (*scalarproduct_int16)(int16_t *v1, int16_t *v2/*align 16*/, int len, int shift);
+
+    /* rv30 functions */
+    qpel_mc_func put_rv30_tpel_pixels_tab[4][16];
+    qpel_mc_func avg_rv30_tpel_pixels_tab[4][16];
+
+    /* rv40 functions */
+    qpel_mc_func put_rv40_qpel_pixels_tab[4][16];
+    qpel_mc_func avg_rv40_qpel_pixels_tab[4][16];
+    h264_chroma_mc_func put_rv40_chroma_pixels_tab[3];
+    h264_chroma_mc_func avg_rv40_chroma_pixels_tab[3];
+} DSPContext;
+
+void dsputil_static_init(void);
+void dsputil_init(DSPContext* p);
+
+int ff_check_alignment(void);
+
+/**
+ * permute block according to permuatation.
+ * @param last last non zero element in scantable order
+ */
+void ff_block_permute(DCTELEM *block, uint8_t *permutation, const uint8_t *scantable, int last);
+
+void ff_set_cmp(DSPContext* c, me_cmp_func *cmp, int type);
+
+#define         BYTE_VEC32(c)   ((c)*0x01010101UL)
+
+static inline uint32_t rnd_avg32(uint32_t a, uint32_t b)
+{
+    return (a | b) - (((a ^ b) & ~BYTE_VEC32(0x01)) >> 1);
+}
+
+static inline uint32_t no_rnd_avg32(uint32_t a, uint32_t b)
+{
+    return (a & b) + (((a ^ b) & ~BYTE_VEC32(0x01)) >> 1);
+}
+
+static inline int get_penalty_factor(int lambda, int lambda2, int type){
+    switch(type&0xFF){
+    default:
+    case FF_CMP_SAD:
+        return lambda>>FF_LAMBDA_SHIFT;
+    case FF_CMP_DCT:
+        return (3*lambda)>>(FF_LAMBDA_SHIFT+1);
+    case FF_CMP_W53:
+        return (4*lambda)>>(FF_LAMBDA_SHIFT);
+    case FF_CMP_W97:
+        return (2*lambda)>>(FF_LAMBDA_SHIFT);
+    case FF_CMP_SATD:
+    case FF_CMP_DCT264:
+        return (2*lambda)>>FF_LAMBDA_SHIFT;
+    case FF_CMP_RD:
+    case FF_CMP_PSNR:
+    case FF_CMP_SSE:
+    case FF_CMP_NSSE:
+        return lambda2>>FF_LAMBDA_SHIFT;
+    case FF_CMP_BIT:
+        return 1;
+    }
+}
+
+/**
+ * Empty mmx state.
+ * this must be called between any dsp function and float/double code.
+ * for example sin(); dsp->idct_put(); emms_c(); cos()
+ */
+#define emms_c()
+
+/* should be defined by architectures supporting
+   one or more MultiMedia extension */
+int mm_support(void);
+
+void dsputil_init_alpha(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_arm(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_bfin(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_mlib(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_mmi(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx);
+void dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);
+
+#define DECLARE_ALIGNED_16(t, v) DECLARE_ALIGNED(16, t, v)
+
+#if HAVE_MMX
+
+#undef emms_c
+
+extern int mm_flags;
+
+void add_pixels_clamped_mmx(const DCTELEM *block, uint8_t *pixels, int line_size);
+void put_pixels_clamped_mmx(const DCTELEM *block, uint8_t *pixels, int line_size);
+void put_signed_pixels_clamped_mmx(const DCTELEM *block, uint8_t *pixels, int line_size);
+
+static inline void emms(void)
+{
+    __asm__ volatile ("emms;":::"memory");
+}
+
+
+#define emms_c() \
+{\
+    if (mm_flags & FF_MM_MMX)\
+        emms();\
+}
+
+void dsputil_init_pix_mmx(DSPContext* c, AVCodecContext *avctx);
+
+#elif ARCH_ARM
+
+extern int mm_flags;
+
+#if HAVE_NEON
+#   define DECLARE_ALIGNED_8(t, v) DECLARE_ALIGNED(16, t, v)
+#   define STRIDE_ALIGN 16
+#endif
+
+#elif ARCH_PPC
+
+extern int mm_flags;
+
+#define DECLARE_ALIGNED_8(t, v) DECLARE_ALIGNED(16, t, v)
+#define STRIDE_ALIGN 16
+
+#elif HAVE_MMI
+
+#define DECLARE_ALIGNED_8(t, v) DECLARE_ALIGNED(16, t, v)
+#define STRIDE_ALIGN 16
+
+#else
+
+#define mm_flags 0
+#define mm_support() 0
+
+#endif
+
+#ifndef DECLARE_ALIGNED_8
+#   define DECLARE_ALIGNED_8(t, v) DECLARE_ALIGNED(8, t, v)
+#endif
+
+#ifndef STRIDE_ALIGN
+#   define STRIDE_ALIGN 8
+#endif
+
+/* PSNR */
+void get_psnr(uint8_t *orig_image[3], uint8_t *coded_image[3],
+              int orig_linesize[3], int coded_linesize,
+              AVCodecContext *avctx);
+
+/* FFT computation */
+
+/* NOTE: soon integer code will be added, so you must use the
+   FFTSample type */
+typedef float FFTSample;
+
+struct MDCTContext;
+
+typedef struct FFTComplex {
+    FFTSample re, im;
+} FFTComplex;
+
+typedef struct FFTContext {
+    int nbits;
+    int inverse;
+    uint16_t *revtab;
+    FFTComplex *exptab;
+    FFTComplex *exptab1; /* only used by SSE code */
+    FFTComplex *tmp_buf;
+    void (*fft_permute)(struct FFTContext *s, FFTComplex *z);
+    void (*fft_calc)(struct FFTContext *s, FFTComplex *z);
+    void (*imdct_calc)(struct MDCTContext *s, FFTSample *output, const FFTSample *input);
+    void (*imdct_half)(struct MDCTContext *s, FFTSample *output, const FFTSample *input);
+} FFTContext;
+
+extern FFTSample* ff_cos_tabs[13];
+
+/**
+ * Sets up a complex FFT.
+ * @param nbits           log2 of the length of the input array
+ * @param inverse         if 0 perform the forward transform, if 1 perform the inverse
+ */
+int ff_fft_init(FFTContext *s, int nbits, int inverse);
+void ff_fft_permute_c(FFTContext *s, FFTComplex *z);
+void ff_fft_permute_sse(FFTContext *s, FFTComplex *z);
+void ff_fft_calc_c(FFTContext *s, FFTComplex *z);
+void ff_fft_calc_sse(FFTContext *s, FFTComplex *z);
+void ff_fft_calc_3dn(FFTContext *s, FFTComplex *z);
+void ff_fft_calc_3dn2(FFTContext *s, FFTComplex *z);
+void ff_fft_calc_altivec(FFTContext *s, FFTComplex *z);
+
+/**
+ * Do the permutation needed BEFORE calling ff_fft_calc().
+ */
+static inline void ff_fft_permute(FFTContext *s, FFTComplex *z)
+{
+    s->fft_permute(s, z);
+}
+/**
+ * Do a complex FFT with the parameters defined in ff_fft_init(). The
+ * input data must be permuted before. No 1.0/sqrt(n) normalization is done.
+ */
+static inline void ff_fft_calc(FFTContext *s, FFTComplex *z)
+{
+    s->fft_calc(s, z);
+}
+void ff_fft_end(FFTContext *s);
+
+/* MDCT computation */
+
+typedef struct MDCTContext {
+    int n;  /* size of MDCT (i.e. number of input data * 2) */
+    int nbits; /* n = 2^nbits */
+    /* pre/post rotation tables */
+    FFTSample *tcos;
+    FFTSample *tsin;
+    FFTContext fft;
+} MDCTContext;
+
+static inline void ff_imdct_calc(MDCTContext *s, FFTSample *output, const FFTSample *input)
+{
+    s->fft.imdct_calc(s, output, input);
+}
+static inline void ff_imdct_half(MDCTContext *s, FFTSample *output, const FFTSample *input)
+{
+    s->fft.imdct_half(s, output, input);
+}
+
+/**
+ * Generate a Kaiser-Bessel Derived Window.
+ * @param   window  pointer to half window
+ * @param   alpha   determines window shape
+ * @param   n       size of half window
+ */
+void ff_kbd_window_init(float *window, float alpha, int n);
+
+/**
+ * Generate a sine window.
+ * @param   window  pointer to half window
+ * @param   n       size of half window
+ */
+void ff_sine_window_init(float *window, int n);
+extern float ff_sine_128 [ 128];
+extern float ff_sine_256 [ 256];
+extern float ff_sine_512 [ 512];
+extern float ff_sine_1024[1024];
+extern float ff_sine_2048[2048];
+extern float ff_sine_4096[4096];
+extern float *ff_sine_windows[6];
+
+int ff_mdct_init(MDCTContext *s, int nbits, int inverse);
+void ff_imdct_calc_c(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_imdct_half_c(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_imdct_calc_3dn(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_imdct_half_3dn(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_imdct_calc_3dn2(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_imdct_half_3dn2(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_imdct_calc_sse(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_imdct_half_sse(MDCTContext *s, FFTSample *output, const FFTSample *input);
+void ff_mdct_calc(MDCTContext *s, FFTSample *out, const FFTSample *input);
+void ff_mdct_end(MDCTContext *s);
+
+/* Real Discrete Fourier Transform */
+
+enum RDFTransformType {
+    RDFT,
+    IRDFT,
+    RIDFT,
+    IRIDFT,
+};
+
+typedef struct {
+    int nbits;
+    int inverse;
+    int sign_convention;
+
+    /* pre/post rotation tables */
+    FFTSample *tcos;
+    FFTSample *tsin;
+    FFTContext fft;
+} RDFTContext;
+
+/**
+ * Sets up a real FFT.
+ * @param nbits           log2 of the length of the input array
+ * @param trans           the type of transform
+ */
+int ff_rdft_init(RDFTContext *s, int nbits, enum RDFTransformType trans);
+void ff_rdft_calc(RDFTContext *s, FFTSample *data);
+void ff_rdft_end(RDFTContext *s);
+
+#define WRAPPER8_16(name8, name16)\
+static int name16(void /*MpegEncContext*/ *s, uint8_t *dst, uint8_t *src, int stride, int h){\
+    return name8(s, dst           , src           , stride, h)\
+          +name8(s, dst+8         , src+8         , stride, h);\
+}
+
+#define WRAPPER8_16_SQ(name8, name16)\
+static int name16(void /*MpegEncContext*/ *s, uint8_t *dst, uint8_t *src, int stride, int h){\
+    int score=0;\
+    score +=name8(s, dst           , src           , stride, 8);\
+    score +=name8(s, dst+8         , src+8         , stride, 8);\
+    if(h==16){\
+        dst += 8*stride;\
+        src += 8*stride;\
+        score +=name8(s, dst           , src           , stride, 8);\
+        score +=name8(s, dst+8         , src+8         , stride, 8);\
+    }\
+    return score;\
+}
+
+
+static inline void copy_block2(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        AV_WN16(dst   , AV_RN16(src   ));
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+static inline void copy_block4(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        AV_WN32(dst   , AV_RN32(src   ));
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+static inline void copy_block8(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        AV_WN32(dst   , AV_RN32(src   ));
+        AV_WN32(dst+4 , AV_RN32(src+4 ));
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+static inline void copy_block9(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        AV_WN32(dst   , AV_RN32(src   ));
+        AV_WN32(dst+4 , AV_RN32(src+4 ));
+        dst[8]= src[8];
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+static inline void copy_block16(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        AV_WN32(dst   , AV_RN32(src   ));
+        AV_WN32(dst+4 , AV_RN32(src+4 ));
+        AV_WN32(dst+8 , AV_RN32(src+8 ));
+        AV_WN32(dst+12, AV_RN32(src+12));
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+static inline void copy_block17(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        AV_WN32(dst   , AV_RN32(src   ));
+        AV_WN32(dst+4 , AV_RN32(src+4 ));
+        AV_WN32(dst+8 , AV_RN32(src+8 ));
+        AV_WN32(dst+12, AV_RN32(src+12));
+        dst[16]= src[16];
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+#endif /* AVCODEC_DSPUTIL_H */

apps/codecs/libatrac/ffmpeg_config.h

@@ -0,0 +1,14 @@
+/* Automatically generated by configure - do not modify */
+#ifndef _FFMPEG_CONFIG_H
+#define _FFMPEG_CONFIG_H
+// CHECK THIS : #include "codecs.h"
+
+#ifdef CPU_ARM
+#define CONFIG_ALIGN 1
+#endif
+
+#ifdef ROCKBOX_BIG_ENDIAN
+#define WORDS_BIGENDIAN
+#endif
+
+#endif

apps/codecs/libatrac/fft.c

@@ -0,0 +1,374 @@
+/*
+ * FFT/IFFT transforms
+ * Copyright (c) 2008 Loren Merritt
+ * Copyright (c) 2002 Fabrice Bellard
+ * Partly based on libdjbfft by D. J. Bernstein
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavcodec/fft.c
+ * FFT/IFFT transforms.
+ */
+
+#include "dsputil.h"
+
+/* cos(2*pi*x/n) for 0<=x<=n/4, followed by its reverse */
+DECLARE_ALIGNED_16(FFTSample, ff_cos_16[8]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_32[16]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_64[32]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_128[64]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_256[128]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_512[256]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_1024[512]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_2048[1024]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_4096[2048]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_8192[4096]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_16384[8192]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_32768[16384]);
+DECLARE_ALIGNED_16(FFTSample, ff_cos_65536[32768]);
+FFTSample *ff_cos_tabs[] = {
+    ff_cos_16, ff_cos_32, ff_cos_64, ff_cos_128, ff_cos_256, ff_cos_512, ff_cos_1024,
+    ff_cos_2048, ff_cos_4096, ff_cos_8192, ff_cos_16384, ff_cos_32768, ff_cos_65536,
+};
+
+static int split_radix_permutation(int i, int n, int inverse)
+{
+    int m;
+    if(n <= 2) return i&1;
+    m = n >> 1;
+    if(!(i&m))            return split_radix_permutation(i, m, inverse)*2;
+    m >>= 1;
+    if(inverse == !(i&m)) return split_radix_permutation(i, m, inverse)*4 + 1;
+    else                  return split_radix_permutation(i, m, inverse)*4 - 1;
+}
+
+av_cold int ff_fft_init(FFTContext *s, int nbits, int inverse)
+{
+    int i, j, m, n;
+    float alpha, c1, s1, s2;
+    int split_radix = 1;
+    int av_unused has_vectors;
+
+    if (nbits < 2 || nbits > 16)
+        goto fail;
+    s->nbits = nbits;
+    n = 1 << nbits;
+
+    s->tmp_buf = NULL;
+    s->exptab  = av_malloc((n / 2) * sizeof(FFTComplex));
+    if (!s->exptab)
+        goto fail;
+    s->revtab = av_malloc(n * sizeof(uint16_t));
+    if (!s->revtab)
+        goto fail;
+    s->inverse = inverse;
+
+    s2 = inverse ? 1.0 : -1.0;
+
+    s->fft_permute = ff_fft_permute_c;
+    s->fft_calc    = ff_fft_calc_c;
+    s->imdct_calc  = ff_imdct_calc_c;
+    s->imdct_half  = ff_imdct_half_c;
+    s->exptab1     = NULL;
+
+#if HAVE_MMX && HAVE_YASM
+    has_vectors = mm_support();
+    if (has_vectors & FF_MM_SSE && HAVE_SSE) {
+        /* SSE for P3/P4/K8 */
+        s->imdct_calc  = ff_imdct_calc_sse;
+        s->imdct_half  = ff_imdct_half_sse;
+        s->fft_permute = ff_fft_permute_sse;
+        s->fft_calc    = ff_fft_calc_sse;
+    } else if (has_vectors & FF_MM_3DNOWEXT && HAVE_AMD3DNOWEXT) {
+        /* 3DNowEx for K7 */
+        s->imdct_calc = ff_imdct_calc_3dn2;
+        s->imdct_half = ff_imdct_half_3dn2;
+        s->fft_calc   = ff_fft_calc_3dn2;
+    } else if (has_vectors & FF_MM_3DNOW && HAVE_AMD3DNOW) {
+        /* 3DNow! for K6-2/3 */
+        s->imdct_calc = ff_imdct_calc_3dn;
+        s->imdct_half = ff_imdct_half_3dn;
+        s->fft_calc   = ff_fft_calc_3dn;
+    }
+#elif HAVE_ALTIVEC
+    has_vectors = mm_support();
+    if (has_vectors & FF_MM_ALTIVEC) {
+        s->fft_calc = ff_fft_calc_altivec;
+        split_radix = 0;
+    }
+#endif
+
+    if (split_radix) {
+        for(j=4; j<=nbits; j++) {
+            int m = 1<<j;
+            double freq = 2*M_PI/m;
+            FFTSample *tab = ff_cos_tabs[j-4];
+            for(i=0; i<=m/4; i++)
+                tab[i] = cos(i*freq);
+            for(i=1; i<m/4; i++)
+                tab[m/2-i] = tab[i];
+        }
+        for(i=0; i<n; i++)
+            s->revtab[-split_radix_permutation(i, n, s->inverse) & (n-1)] = i;
+        s->tmp_buf = av_malloc(n * sizeof(FFTComplex));
+    } else {
+        int np, nblocks, np2, l;
+        FFTComplex *q;
+
+        for(i=0; i<(n/2); i++) {
+            alpha = 2 * M_PI * (float)i / (float)n;
+            c1 = cos(alpha);
+            s1 = sin(alpha) * s2;
+            s->exptab[i].re = c1;
+            s->exptab[i].im = s1;
+        }
+
+        np = 1 << nbits;
+        nblocks = np >> 3;
+        np2 = np >> 1;
+        s->exptab1 = av_malloc(np * 2 * sizeof(FFTComplex));
+        if (!s->exptab1)
+            goto fail;
+        q = s->exptab1;
+        do {
+            for(l = 0; l < np2; l += 2 * nblocks) {
+                *q++ = s->exptab[l];
+                *q++ = s->exptab[l + nblocks];
+
+                q->re = -s->exptab[l].im;
+                q->im = s->exptab[l].re;
+                q++;
+                q->re = -s->exptab[l + nblocks].im;
+                q->im = s->exptab[l + nblocks].re;
+                q++;
+            }
+            nblocks = nblocks >> 1;
+        } while (nblocks != 0);
+        av_freep(&s->exptab);
+
+        /* compute bit reverse table */
+        for(i=0;i<n;i++) {
+            m=0;
+            for(j=0;j<nbits;j++) {
+                m |= ((i >> j) & 1) << (nbits-j-1);
+            }
+            s->revtab[i]=m;
+        }
+    }
+
+    return 0;
+ fail:
+    av_freep(&s->revtab);
+    av_freep(&s->exptab);
+    av_freep(&s->exptab1);
+    av_freep(&s->tmp_buf);
+    return -1;
+}
+
+void ff_fft_permute_c(FFTContext *s, FFTComplex *z)
+{
+    int j, k, np;
+    FFTComplex tmp;
+    const uint16_t *revtab = s->revtab;
+    np = 1 << s->nbits;
+
+    if (s->tmp_buf) {
+        /* TODO: handle split-radix permute in a more optimal way, probably in-place */
+        for(j=0;j<np;j++) s->tmp_buf[revtab[j]] = z[j];
+        memcpy(z, s->tmp_buf, np * sizeof(FFTComplex));
+        return;
+    }
+
+    /* reverse */
+    for(j=0;j<np;j++) {
+        k = revtab[j];
+        if (k < j) {
+            tmp = z[k];
+            z[k] = z[j];
+            z[j] = tmp;
+        }
+    }
+}
+
+av_cold void ff_fft_end(FFTContext *s)
+{
+    av_freep(&s->revtab);
+    av_freep(&s->exptab);
+    av_freep(&s->exptab1);
+    av_freep(&s->tmp_buf);
+}
+
+#define sqrthalf (float)M_SQRT1_2
+
+#define BF(x,y,a,b) {\
+    x = a - b;\
+    y = a + b;\
+}
+
+#define BUTTERFLIES(a0,a1,a2,a3) {\
+    BF(t3, t5, t5, t1);\
+    BF(a2.re, a0.re, a0.re, t5);\
+    BF(a3.im, a1.im, a1.im, t3);\
+    BF(t4, t6, t2, t6);\
+    BF(a3.re, a1.re, a1.re, t4);\
+    BF(a2.im, a0.im, a0.im, t6);\
+}
+
+// force loading all the inputs before storing any.
+// this is slightly slower for small data, but avoids store->load aliasing
+// for addresses separated by large powers of 2.
+#define BUTTERFLIES_BIG(a0,a1,a2,a3) {\
+    FFTSample r0=a0.re, i0=a0.im, r1=a1.re, i1=a1.im;\
+    BF(t3, t5, t5, t1);\
+    BF(a2.re, a0.re, r0, t5);\
+    BF(a3.im, a1.im, i1, t3);\
+    BF(t4, t6, t2, t6);\
+    BF(a3.re, a1.re, r1, t4);\
+    BF(a2.im, a0.im, i0, t6);\
+}
+
+#define TRANSFORM(a0,a1,a2,a3,wre,wim) {\
+    t1 = a2.re * wre + a2.im * wim;\
+    t2 = a2.im * wre - a2.re * wim;\
+    t5 = a3.re * wre - a3.im * wim;\
+    t6 = a3.im * wre + a3.re * wim;\
+    BUTTERFLIES(a0,a1,a2,a3)\
+}
+
+#define TRANSFORM_ZERO(a0,a1,a2,a3) {\
+    t1 = a2.re;\
+    t2 = a2.im;\
+    t5 = a3.re;\
+    t6 = a3.im;\
+    BUTTERFLIES(a0,a1,a2,a3)\
+}
+
+/* z[0...8n-1], w[1...2n-1] */
+#define PASS(name)\
+static void name(FFTComplex *z, const FFTSample *wre, unsigned int n)\
+{\
+    FFTSample t1, t2, t3, t4, t5, t6;\
+    int o1 = 2*n;\
+    int o2 = 4*n;\
+    int o3 = 6*n;\
+    const FFTSample *wim = wre+o1;\
+    n--;\
+\
+    TRANSFORM_ZERO(z[0],z[o1],z[o2],z[o3]);\
+    TRANSFORM(z[1],z[o1+1],z[o2+1],z[o3+1],wre[1],wim[-1]);\
+    do {\
+        z += 2;\
+        wre += 2;\
+        wim -= 2;\
+        TRANSFORM(z[0],z[o1],z[o2],z[o3],wre[0],wim[0]);\
+        TRANSFORM(z[1],z[o1+1],z[o2+1],z[o3+1],wre[1],wim[-1]);\
+    } while(--n);\
+}
+
+PASS(pass)
+#undef BUTTERFLIES
+#define BUTTERFLIES BUTTERFLIES_BIG
+PASS(pass_big)
+
+#define DECL_FFT(n,n2,n4)\
+static void fft##n(FFTComplex *z)\
+{\
+    fft##n2(z);\
+    fft##n4(z+n4*2);\
+    fft##n4(z+n4*3);\
+    pass(z,ff_cos_##n,n4/2);\
+}
+
+static void fft4(FFTComplex *z)
+{
+    FFTSample t1, t2, t3, t4, t5, t6, t7, t8;
+
+    BF(t3, t1, z[0].re, z[1].re);
+    BF(t8, t6, z[3].re, z[2].re);
+    BF(z[2].re, z[0].re, t1, t6);
+    BF(t4, t2, z[0].im, z[1].im);
+    BF(t7, t5, z[2].im, z[3].im);
+    BF(z[3].im, z[1].im, t4, t8);
+    BF(z[3].re, z[1].re, t3, t7);
+    BF(z[2].im, z[0].im, t2, t5);
+}
+
+static void fft8(FFTComplex *z)
+{
+    FFTSample t1, t2, t3, t4, t5, t6, t7, t8;
+
+    fft4(z);
+
+    BF(t1, z[5].re, z[4].re, -z[5].re);
+    BF(t2, z[5].im, z[4].im, -z[5].im);
+    BF(t3, z[7].re, z[6].re, -z[7].re);
+    BF(t4, z[7].im, z[6].im, -z[7].im);
+    BF(t8, t1, t3, t1);
+    BF(t7, t2, t2, t4);
+    BF(z[4].re, z[0].re, z[0].re, t1);
+    BF(z[4].im, z[0].im, z[0].im, t2);
+    BF(z[6].re, z[2].re, z[2].re, t7);
+    BF(z[6].im, z[2].im, z[2].im, t8);
+
+    TRANSFORM(z[1],z[3],z[5],z[7],sqrthalf,sqrthalf);
+}
+
+#if !CONFIG_SMALL
+static void fft16(FFTComplex *z)
+{
+    FFTSample t1, t2, t3, t4, t5, t6;
+
+    fft8(z);
+    fft4(z+8);
+    fft4(z+12);
+
+    TRANSFORM_ZERO(z[0],z[4],z[8],z[12]);
+    TRANSFORM(z[2],z[6],z[10],z[14],sqrthalf,sqrthalf);
+    TRANSFORM(z[1],z[5],z[9],z[13],ff_cos_16[1],ff_cos_16[3]);
+    TRANSFORM(z[3],z[7],z[11],z[15],ff_cos_16[3],ff_cos_16[1]);
+}
+#else
+DECL_FFT(16,8,4)
+#endif
+DECL_FFT(32,16,8)
+DECL_FFT(64,32,16)
+DECL_FFT(128,64,32)
+DECL_FFT(256,128,64)
+DECL_FFT(512,256,128)
+#if !CONFIG_SMALL
+#define pass pass_big
+#endif
+DECL_FFT(1024,512,256)
+DECL_FFT(2048,1024,512)
+DECL_FFT(4096,2048,1024)
+DECL_FFT(8192,4096,2048)
+DECL_FFT(16384,8192,4096)
+DECL_FFT(32768,16384,8192)
+DECL_FFT(65536,32768,16384)
+
+static void (*fft_dispatch[])(FFTComplex*) = {
+    fft4, fft8, fft16, fft32, fft64, fft128, fft256, fft512, fft1024,
+    fft2048, fft4096, fft8192, fft16384, fft32768, fft65536,
+};
+
+void ff_fft_calc_c(FFTContext *s, FFTComplex *z)
+{
+    fft_dispatch[s->nbits-2](z);
+}
+

apps/codecs/libatrac/libavutil/avutil.h

@@ -0,0 +1,63 @@
+/*
+ * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVUTIL_AVUTIL_H
+#define AVUTIL_AVUTIL_H
+
+/**
+ * @file libavutil/avutil.h
+ * external API header
+ */
+
+
+#define AV_STRINGIFY(s)         AV_TOSTRING(s)
+#define AV_TOSTRING(s) #s
+
+#define AV_VERSION_INT(a, b, c) (a<<16 | b<<8 | c)
+#define AV_VERSION_DOT(a, b, c) a ##.## b ##.## c
+#define AV_VERSION(a, b, c) AV_VERSION_DOT(a, b, c)
+
+#define LIBAVUTIL_VERSION_MAJOR 50
+#define LIBAVUTIL_VERSION_MINOR  0
+#define LIBAVUTIL_VERSION_MICRO  0
+
+#define LIBAVUTIL_VERSION_INT   AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \
+                                               LIBAVUTIL_VERSION_MINOR, \
+                                               LIBAVUTIL_VERSION_MICRO)
+#define LIBAVUTIL_VERSION       AV_VERSION(LIBAVUTIL_VERSION_MAJOR,     \
+                                           LIBAVUTIL_VERSION_MINOR,     \
+                                           LIBAVUTIL_VERSION_MICRO)
+#define LIBAVUTIL_BUILD         LIBAVUTIL_VERSION_INT
+
+#define LIBAVUTIL_IDENT         "Lavu" AV_STRINGIFY(LIBAVUTIL_VERSION)
+
+/**
+ * Returns the LIBAVUTIL_VERSION_INT constant.
+ */
+unsigned avutil_version(void);
+
+#include "common.h"
+//#include "mathematics.h"
+//#include "rational.h"
+//#include "intfloat_readwrite.h"
+#include "log.h"
+//#include "pixfmt.h"
+
+#endif /* AVUTIL_AVUTIL_H */

apps/codecs/libatrac/libavutil/bswap.h

@@ -0,0 +1,99 @@
+/*
+ * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavutil/bswap.h
+ * byte swapping routines
+ */
+
+#ifndef AVUTIL_BSWAP_H
+#define AVUTIL_BSWAP_H
+
+#include <stdint.h>
+//#include "ffmpeg_config.h"
+#include "common.h"
+
+#if   ARCH_ARM
+#   include "arm/bswap.h"
+#elif ARCH_BFIN
+#   include "bfin/bswap.h"
+#elif ARCH_SH4
+#   include "sh4/bswap.h"
+#elif ARCH_X86
+#   include "x86/bswap.h"
+#endif
+
+#ifndef bswap_16
+static av_always_inline av_const uint16_t bswap_16(uint16_t x)
+{
+    x= (x>>8) | (x<<8);
+    return x;
+}
+#endif
+
+#ifndef bswap_32
+static av_always_inline av_const uint32_t bswap_32(uint32_t x)
+{
+    x= ((x<<8)&0xFF00FF00) | ((x>>8)&0x00FF00FF);
+    x= (x>>16) | (x<<16);
+    return x;
+}
+#endif
+
+#ifndef bswap_64
+static inline uint64_t av_const bswap_64(uint64_t x)
+{
+#if 0
+    x= ((x<< 8)&0xFF00FF00FF00FF00ULL) | ((x>> 8)&0x00FF00FF00FF00FFULL);
+    x= ((x<<16)&0xFFFF0000FFFF0000ULL) | ((x>>16)&0x0000FFFF0000FFFFULL);
+    return (x>>32) | (x<<32);
+#else
+    union {
+        uint64_t ll;
+        uint32_t l[2];
+    } w, r;
+    w.ll = x;
+    r.l[0] = bswap_32 (w.l[1]);
+    r.l[1] = bswap_32 (w.l[0]);
+    return r.ll;
+#endif
+}
+#endif
+
+// be2me ... big-endian to machine-endian
+// le2me ... little-endian to machine-endian
+
+#ifdef WORDS_BIGENDIAN
+#define be2me_16(x) (x)
+#define be2me_32(x) (x)
+#define be2me_64(x) (x)
+#define le2me_16(x) bswap_16(x)
+#define le2me_32(x) bswap_32(x)
+#define le2me_64(x) bswap_64(x)
+#else
+#define be2me_16(x) bswap_16(x)
+#define be2me_32(x) bswap_32(x)
+#define be2me_64(x) bswap_64(x)
+#define le2me_16(x) (x)
+#define le2me_32(x) (x)
+#define le2me_64(x) (x)
+#endif
+
+#endif /* AVUTIL_BSWAP_H */

apps/codecs/libatrac/libavutil/common.h

@@ -0,0 +1,286 @@
+/*
+ * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavutil/common.h
+ * common internal and external API header
+ */
+
+#ifndef AVUTIL_COMMON_H
+#define AVUTIL_COMMON_H
+
+#include <ctype.h>
+#include <errno.h>
+#include <inttypes.h>
+#include <limits.h>
+#include <math.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#ifdef __GNUC__
+#    define AV_GCC_VERSION_AT_LEAST(x,y) (__GNUC__ > x || __GNUC__ == x && __GNUC_MINOR__ >= y)
+#else
+#    define AV_GCC_VERSION_AT_LEAST(x,y) 0
+#endif
+
+#ifndef av_always_inline
+#if AV_GCC_VERSION_AT_LEAST(3,1)
+#    define av_always_inline __attribute__((always_inline)) inline
+#else
+#    define av_always_inline inline
+#endif
+#endif
+
+#ifndef av_noinline
+#if AV_GCC_VERSION_AT_LEAST(3,1)
+#    define av_noinline __attribute__((noinline))
+#else
+#    define av_noinline
+#endif
+#endif
+
+#ifndef av_pure
+#if AV_GCC_VERSION_AT_LEAST(3,1)
+#    define av_pure __attribute__((pure))
+#else
+#    define av_pure
+#endif
+#endif
+
+#ifndef av_const
+#if AV_GCC_VERSION_AT_LEAST(2,6)
+#    define av_const __attribute__((const))
+#else
+#    define av_const
+#endif
+#endif
+
+#ifndef av_cold
+#if (!defined(__ICC) || __ICC > 1100) && AV_GCC_VERSION_AT_LEAST(4,3)
+#    define av_cold __attribute__((cold))
+#else
+#    define av_cold
+#endif
+#endif
+
+#ifndef av_flatten
+#if AV_GCC_VERSION_AT_LEAST(4,1)
+#    define av_flatten __attribute__((flatten))
+#else
+#    define av_flatten
+#endif
+#endif
+
+#ifndef attribute_deprecated
+#if AV_GCC_VERSION_AT_LEAST(3,1)
+#    define attribute_deprecated __attribute__((deprecated))
+#else
+#    define attribute_deprecated
+#endif
+#endif
+
+#ifndef av_unused
+#if defined(__GNUC__)
+#    define av_unused __attribute__((unused))
+#else
+#    define av_unused
+#endif
+#endif
+
+#ifndef av_uninit
+#if defined(__GNUC__) && !defined(__ICC)
+#    define av_uninit(x) x=x
+#else
+#    define av_uninit(x) x
+#endif
+#endif
+
+//rounded division & shift
+#define RSHIFT(a,b) ((a) > 0 ? ((a) + ((1<<(b))>>1))>>(b) : ((a) + ((1<<(b))>>1)-1)>>(b))
+/* assume b>0 */
+#define ROUNDED_DIV(a,b) (((a)>0 ? (a) + ((b)>>1) : (a) - ((b)>>1))/(b))
+#define FFABS(a) ((a) >= 0 ? (a) : (-(a)))
+#define FFSIGN(a) ((a) > 0 ? 1 : -1)
+
+#define FFMAX(a,b) ((a) > (b) ? (a) : (b))
+#define FFMAX3(a,b,c) FFMAX(FFMAX(a,b),c)
+#define FFMIN(a,b) ((a) > (b) ? (b) : (a))
+#define FFMIN3(a,b,c) FFMIN(FFMIN(a,b),c)
+
+#define FFSWAP(type,a,b) do{type SWAP_tmp= b; b= a; a= SWAP_tmp;}while(0)
+#define FF_ARRAY_ELEMS(a) (sizeof(a) / sizeof((a)[0]))
+
+/* misc math functions */
+extern const uint8_t ff_log2_tab[256];
+
+static inline av_const int av_log2(unsigned int v)
+{
+    int n = 0;
+    if (v & 0xffff0000) {
+        v >>= 16;
+        n += 16;
+    }
+    if (v & 0xff00) {
+        v >>= 8;
+        n += 8;
+    }
+    n += ff_log2_tab[v];
+
+    return n;
+}
+
+static inline av_const int av_log2_16bit(unsigned int v)
+{
+    int n = 0;
+    if (v & 0xff00) {
+        v >>= 8;
+        n += 8;
+    }
+    n += ff_log2_tab[v];
+
+    return n;
+}
+
+/**
+ * Clips a signed integer value into the amin-amax range.
+ * @param a value to clip
+ * @param amin minimum value of the clip range
+ * @param amax maximum value of the clip range
+ * @return clipped value
+ */
+static inline av_const int av_clip(int a, int amin, int amax)
+{
+    if      (a < amin) return amin;
+    else if (a > amax) return amax;
+    else               return a;
+}
+
+/**
+ * Clips a signed integer value into the 0-255 range.
+ * @param a value to clip
+ * @return clipped value
+ */
+static inline av_const uint8_t av_clip_uint8(int a)
+{
+    if (a&(~255)) return (-a)>>31;
+    else          return a;
+}
+
+/**
+ * Clips a signed integer value into the -32768,32767 range.
+ * @param a value to clip
+ * @return clipped value
+ */
+static inline av_const int16_t av_clip_int16(int a)
+{
+    if ((a+32768) & ~65535) return (a>>31) ^ 32767;
+    else                    return a;
+}
+
+/**
+ * Clips a float value into the amin-amax range.
+ * @param a value to clip
+ * @param amin minimum value of the clip range
+ * @param amax maximum value of the clip range
+ * @return clipped value
+ */
+static inline av_const float av_clipf(float a, float amin, float amax)
+{
+    if      (a < amin) return amin;
+    else if (a > amax) return amax;
+    else               return a;
+}
+
+#define MKTAG(a,b,c,d) (a | (b << 8) | (c << 16) | (d << 24))
+#define MKBETAG(a,b,c,d) (d | (c << 8) | (b << 16) | (a << 24))
+
+/*!
+ * \def GET_UTF8(val, GET_BYTE, ERROR)
+ * Converts a UTF-8 character (up to 4 bytes long) to its 32-bit UCS-4 encoded form
+ * \param val is the output and should be of type uint32_t. It holds the converted
+ * UCS-4 character and should be a left value.
+ * \param GET_BYTE gets UTF-8 encoded bytes from any proper source. It can be
+ * a function or a statement whose return value or evaluated value is of type
+ * uint8_t. It will be executed up to 4 times for values in the valid UTF-8 range,
+ * and up to 7 times in the general case.
+ * \param ERROR action that should be taken when an invalid UTF-8 byte is returned
+ * from GET_BYTE. It should be a statement that jumps out of the macro,
+ * like exit(), goto, return, break, or continue.
+ */
+#define GET_UTF8(val, GET_BYTE, ERROR)\
+    val= GET_BYTE;\
+    {\
+        int ones= 7 - av_log2(val ^ 255);\
+        if(ones==1)\
+            ERROR\
+        val&= 127>>ones;\
+        while(--ones > 0){\
+            int tmp= GET_BYTE - 128;\
+            if(tmp>>6)\
+                ERROR\
+            val= (val<<6) + tmp;\
+        }\
+    }
+
+/*!
+ * \def PUT_UTF8(val, tmp, PUT_BYTE)
+ * Converts a 32-bit Unicode character to its UTF-8 encoded form (up to 4 bytes long).
+ * \param val is an input-only argument and should be of type uint32_t. It holds
+ * a UCS-4 encoded Unicode character that is to be converted to UTF-8. If
+ * val is given as a function it is executed only once.
+ * \param tmp is a temporary variable and should be of type uint8_t. It
+ * represents an intermediate value during conversion that is to be
+ * output by PUT_BYTE.
+ * \param PUT_BYTE writes the converted UTF-8 bytes to any proper destination.
+ * It could be a function or a statement, and uses tmp as the input byte.
+ * For example, PUT_BYTE could be "*output++ = tmp;" PUT_BYTE will be
+ * executed up to 4 times for values in the valid UTF-8 range and up to
+ * 7 times in the general case, depending on the length of the converted
+ * Unicode character.
+ */
+#define PUT_UTF8(val, tmp, PUT_BYTE)\
+    {\
+        int bytes, shift;\
+        uint32_t in = val;\
+        if (in < 0x80) {\
+            tmp = in;\
+            PUT_BYTE\
+        } else {\
+            bytes = (av_log2(in) + 4) / 5;\
+            shift = (bytes - 1) * 6;\
+            tmp = (256 - (256 >> bytes)) | (in >> shift);\
+            PUT_BYTE\
+            while (shift >= 6) {\
+                shift -= 6;\
+                tmp = 0x80 | ((in >> shift) & 0x3f);\
+                PUT_BYTE\
+            }\
+        }\
+    }
+
+#include "mem.h"
+
+//#ifdef HAVE_AV_CONFIG_H
+//#    include "ffmpeg_config.h"
+#    include "internal.h"
+//#endif /* HAVE_AV_CONFIG_H */
+
+#endif /* AVUTIL_COMMON_H */

apps/codecs/libatrac/libavutil/internal.h

@@ -0,0 +1,328 @@
+/*
+ * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavutil/internal.h
+ * common internal API header
+ */
+
+#ifndef AVUTIL_INTERNAL_H
+#define AVUTIL_INTERNAL_H
+
+#if !defined(DEBUG) && !defined(NDEBUG)
+#    define NDEBUG
+#endif
+
+#include <limits.h>
+#include <stdint.h>
+#include <stddef.h>
+#include <assert.h>
+//#include "ffmpeg_config.h"
+#include "common.h"
+#include "mem.h"
+//#include "timer.h"
+
+#ifndef attribute_align_arg
+#if (!defined(__ICC) || __ICC > 1100) && AV_GCC_VERSION_AT_LEAST(4,2)
+#    define attribute_align_arg __attribute__((force_align_arg_pointer))
+#else
+#    define attribute_align_arg
+#endif
+#endif
+
+#ifndef attribute_used
+#if AV_GCC_VERSION_AT_LEAST(3,1)
+#    define attribute_used __attribute__((used))
+#else
+#    define attribute_used
+#endif
+#endif
+
+#ifndef INT16_MIN
+#define INT16_MIN       (-0x7fff-1)
+#endif
+
+#ifndef INT16_MAX
+#define INT16_MAX       0x7fff
+#endif
+
+#ifndef INT32_MIN
+#define INT32_MIN       (-0x7fffffff-1)
+#endif
+
+#ifndef INT32_MAX
+#define INT32_MAX       0x7fffffff
+#endif
+
+#ifndef UINT32_MAX
+#define UINT32_MAX      0xffffffff
+#endif
+
+#ifndef INT64_MIN
+#define INT64_MIN       (-0x7fffffffffffffffLL-1)
+#endif
+
+#ifndef INT64_MAX
+#define INT64_MAX INT64_C(9223372036854775807)
+#endif
+
+#ifndef UINT64_MAX
+#define UINT64_MAX UINT64_C(0xFFFFFFFFFFFFFFFF)
+#endif
+
+#ifndef INT_BIT
+#    define INT_BIT (CHAR_BIT * sizeof(int))
+#endif
+
+#if ( defined(__PIC__) || defined(__pic__) ) && ! defined(PIC)
+#    define PIC
+#endif
+
+#ifndef offsetof
+#    define offsetof(T,F) ((unsigned int)((char *)&((T *)0)->F))
+#endif
+
+// Use rip-relative addressing if compiling PIC code on x86-64.
+#if ARCH_X86_64 && defined(PIC)
+#    define LOCAL_MANGLE(a) #a "(%%rip)"
+#else
+#    define LOCAL_MANGLE(a) #a
+#endif
+
+#define MANGLE(a) EXTERN_PREFIX LOCAL_MANGLE(a)
+
+/* debug stuff */
+
+/* dprintf macros */
+#ifdef DEBUG
+#    define dprintf(pctx, ...) av_log(pctx, AV_LOG_DEBUG, __VA_ARGS__)
+#else
+#    define dprintf(pctx, ...)
+#endif
+
+#define av_abort()      do { av_log(NULL, AV_LOG_ERROR, "Abort at %s:%d\n", __FILE__, __LINE__); abort(); } while (0)
+
+/* math */
+
+extern const uint32_t ff_inverse[256];
+
+#if ARCH_X86
+#    define FASTDIV(a,b) \
+    ({\
+        int ret,dmy;\
+        __asm__ volatile(\
+            "mull %3"\
+            :"=d"(ret),"=a"(dmy)\
+            :"1"(a),"g"(ff_inverse[b])\
+            );\
+        ret;\
+    })
+#elif HAVE_ARMV6 && HAVE_INLINE_ASM
+static inline av_const int FASTDIV(int a, int b)
+{
+    int r, t;
+    __asm__ volatile("cmp     %3, #2               \n\t"
+                     "ldr     %1, [%4, %3, lsl #2] \n\t"
+                     "lsrle   %0, %2, #1           \n\t"
+                     "smmulgt %0, %1, %2           \n\t"
+                     : "=&r"(r), "=&r"(t) : "r"(a), "r"(b), "r"(ff_inverse));
+    return r;
+}
+#elif ARCH_ARM && HAVE_INLINE_ASM
+static inline av_const int FASTDIV(int a, int b)
+{
+    int r, t;
+    __asm__ volatile ("umull %1, %0, %2, %3"
+                      : "=&r"(r), "=&r"(t) : "r"(a), "r"(ff_inverse[b]));
+    return r;
+}
+#elif CONFIG_FASTDIV
+#    define FASTDIV(a,b)   ((uint32_t)((((uint64_t)a)*ff_inverse[b])>>32))
+#else
+#    define FASTDIV(a,b)   ((a)/(b))
+#endif
+
+extern const uint8_t ff_sqrt_tab[256];
+
+static inline av_const unsigned int ff_sqrt(unsigned int a)
+{
+    unsigned int b;
+
+    if(a<255) return (ff_sqrt_tab[a+1]-1)>>4;
+    else if(a<(1<<12)) b= ff_sqrt_tab[a>>4 ]>>2;
+#if !CONFIG_SMALL
+    else if(a<(1<<14)) b= ff_sqrt_tab[a>>6 ]>>1;
+    else if(a<(1<<16)) b= ff_sqrt_tab[a>>8 ]   ;
+#endif
+    else{
+        int s= av_log2_16bit(a>>16)>>1;
+        unsigned int c= a>>(s+2);
+        b= ff_sqrt_tab[c>>(s+8)];
+        b= FASTDIV(c,b) + (b<<s);
+    }
+
+    return b - (a<b*b);
+}
+
+#if ARCH_X86
+#define MASK_ABS(mask, level)\
+            __asm__ volatile(\
+                "cltd                   \n\t"\
+                "xorl %1, %0            \n\t"\
+                "subl %1, %0            \n\t"\
+                : "+a" (level), "=&d" (mask)\
+            );
+#else
+#define MASK_ABS(mask, level)\
+            mask= level>>31;\
+            level= (level^mask)-mask;
+#endif
+
+#if HAVE_CMOV
+#define COPY3_IF_LT(x,y,a,b,c,d)\
+__asm__ volatile (\
+    "cmpl %0, %3        \n\t"\
+    "cmovl %3, %0       \n\t"\
+    "cmovl %4, %1       \n\t"\
+    "cmovl %5, %2       \n\t"\
+    : "+&r" (x), "+&r" (a), "+r" (c)\
+    : "r" (y), "r" (b), "r" (d)\
+);
+#else
+#define COPY3_IF_LT(x,y,a,b,c,d)\
+if((y)<(x)){\
+     (x)=(y);\
+     (a)=(b);\
+     (c)=(d);\
+}
+#endif
+
+/* avoid usage of dangerous/inappropriate system functions */
+#undef  malloc
+#define malloc please_use_av_malloc
+#undef  free
+#define free please_use_av_free
+#undef  realloc
+#define realloc please_use_av_realloc
+#undef  time
+#define time time_is_forbidden_due_to_security_issues
+//#undef  rand
+//#define rand rand_is_forbidden_due_to_state_trashing_use_av_random
+//#undef  srand
+//#define srand srand_is_forbidden_due_to_state_trashing_use_av_random_init
+#undef  random
+#define random random_is_forbidden_due_to_state_trashing_use_av_random
+#undef  sprintf
+#define sprintf sprintf_is_forbidden_due_to_security_issues_use_snprintf
+#undef  strcat
+#define strcat strcat_is_forbidden_due_to_security_issues_use_av_strlcat
+#undef  exit
+#define exit exit_is_forbidden
+#ifndef LIBAVFORMAT_BUILD
+//#undef  printf
+//#define printf please_use_av_log_instead_of_printf
+#undef  fprintf
+#define fprintf please_use_av_log_instead_of_fprintf
+#undef  puts
+#define puts please_use_av_log_instead_of_puts
+#undef  perror
+#define perror please_use_av_log_instead_of_perror
+#endif
+
+#define CHECKED_ALLOCZ(p, size)\
+{\
+    p= av_mallocz(size);\
+    if(p==NULL && (size)!=0){\
+        av_log(NULL, AV_LOG_ERROR, "Cannot allocate memory.");\
+        goto fail;\
+    }\
+}
+
+#if defined(__ICC) || defined(__SUNPRO_C)
+    #define DECLARE_ALIGNED(n,t,v)      t v __attribute__ ((aligned (n)))
+    #define DECLARE_ASM_CONST(n,t,v)    const t __attribute__ ((aligned (n))) v
+#elif defined(__GNUC__)
+    #define DECLARE_ALIGNED(n,t,v)      t v __attribute__ ((aligned (n)))
+    #define DECLARE_ASM_CONST(n,t,v)    static const t v attribute_used __attribute__ ((aligned (n)))
+#elif defined(_MSC_VER)
+    #define DECLARE_ALIGNED(n,t,v)      __declspec(align(n)) t v
+    #define DECLARE_ASM_CONST(n,t,v)    __declspec(align(n)) static const t v
+#elif HAVE_INLINE_ASM
+    #error The asm code needs alignment, but we do not know how to do it for this compiler.
+#else
+    #define DECLARE_ALIGNED(n,t,v)      t v
+    #define DECLARE_ASM_CONST(n,t,v)    static const t v
+#endif
+
+
+#if !HAVE_LLRINT
+static av_always_inline av_const long long llrint(double x)
+{
+    return rint(x);
+}
+#endif /* HAVE_LLRINT */
+
+#if !HAVE_LRINT
+static av_always_inline av_const long int lrint(double x)
+{
+    return rint(x);
+}
+#endif /* HAVE_LRINT */
+
+#if !HAVE_LRINTF
+static av_always_inline av_const long int lrintf(float x)
+{
+    return (int)(rint(x));
+}
+#endif /* HAVE_LRINTF */
+
+#if !HAVE_ROUND
+static av_always_inline av_const double round(double x)
+{
+    return (x > 0) ? floor(x + 0.5) : ceil(x - 0.5);
+}
+#endif /* HAVE_ROUND */
+
+#if !HAVE_ROUNDF
+static av_always_inline av_const float roundf(float x)
+{
+    return (x > 0) ? floor(x + 0.5) : ceil(x - 0.5);
+}
+#endif /* HAVE_ROUNDF */
+
+#if !HAVE_TRUNCF
+static av_always_inline av_const float truncf(float x)
+{
+    return (x > 0) ? floor(x) : ceil(x);
+}
+#endif /* HAVE_TRUNCF */
+
+/**
+ * Returns NULL if CONFIG_SMALL is true, otherwise the argument
+ * without modification. Used to disable the definition of strings
+ * (for example AVCodec long_names).
+ */
+#if CONFIG_SMALL
+#   define NULL_IF_CONFIG_SMALL(x) NULL
+#else
+#   define NULL_IF_CONFIG_SMALL(x) x
+#endif
+
+#endif /* AVUTIL_INTERNAL_H */

apps/codecs/libatrac/libavutil/intreadwrite.h

@@ -0,0 +1,192 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVUTIL_INTREADWRITE_H
+#define AVUTIL_INTREADWRITE_H
+
+#include <stdint.h>
+//#include "ffmpeg_config.h"
+#include "bswap.h"
+
+#ifdef __GNUC__
+
+struct unaligned_64 { uint64_t l; } __attribute__((packed));
+struct unaligned_32 { uint32_t l; } __attribute__((packed));
+struct unaligned_16 { uint16_t l; } __attribute__((packed));
+
+#define AV_RN16(a) (((const struct unaligned_16 *) (a))->l)
+#define AV_RN32(a) (((const struct unaligned_32 *) (a))->l)
+#define AV_RN64(a) (((const struct unaligned_64 *) (a))->l)
+
+#define AV_WN16(a, b) (((struct unaligned_16 *) (a))->l) = (b)
+#define AV_WN32(a, b) (((struct unaligned_32 *) (a))->l) = (b)
+#define AV_WN64(a, b) (((struct unaligned_64 *) (a))->l) = (b)
+
+#elif defined(__DECC)
+
+#define AV_RN16(a) (*((const __unaligned uint16_t*)(a)))
+#define AV_RN32(a) (*((const __unaligned uint32_t*)(a)))
+#define AV_RN64(a) (*((const __unaligned uint64_t*)(a)))
+
+#define AV_WN16(a, b) *((__unaligned uint16_t*)(a)) = (b)
+#define AV_WN32(a, b) *((__unaligned uint32_t*)(a)) = (b)
+#define AV_WN64(a, b) *((__unaligned uint64_t*)(a)) = (b)
+
+#else
+
+#define AV_RN16(a) (*((const uint16_t*)(a)))
+#define AV_RN32(a) (*((const uint32_t*)(a)))
+#define AV_RN64(a) (*((const uint64_t*)(a)))
+
+#define AV_WN16(a, b) *((uint16_t*)(a)) = (b)
+#define AV_WN32(a, b) *((uint32_t*)(a)) = (b)
+#define AV_WN64(a, b) *((uint64_t*)(a)) = (b)
+
+#endif /* !__GNUC__ */
+
+/* endian macros */
+#define AV_RB8(x)     (((const uint8_t*)(x))[0])
+#define AV_WB8(p, d)  do { ((uint8_t*)(p))[0] = (d); } while(0)
+
+#define AV_RL8(x)     AV_RB8(x)
+#define AV_WL8(p, d)  AV_WB8(p, d)
+
+#if HAVE_FAST_UNALIGNED
+# ifdef WORDS_BIGENDIAN
+#  define AV_RB16(x)    AV_RN16(x)
+#  define AV_WB16(p, d) AV_WN16(p, d)
+
+#  define AV_RL16(x)    bswap_16(AV_RN16(x))
+#  define AV_WL16(p, d) AV_WN16(p, bswap_16(d))
+
+#  define AV_RB32(x)    AV_RN32(x)
+#  define AV_WB32(p, d) AV_WN32(p, d)
+
+#  define AV_RL32(x)    bswap_32(AV_RN32(x))
+#  define AV_WL32(p, d) AV_WN32(p, bswap_32(d))
+
+#  define AV_RB64(x)    AV_RN64(x)
+#  define AV_WB64(p, d) AV_WN64(p, d)
+
+#  define AV_RL64(x)    bswap_64(AV_RN64(x))
+#  define AV_WL64(p, d) AV_WN64(p, bswap_64(d))
+# else /* WORDS_BIGENDIAN */
+#  define AV_RB16(x)    bswap_16(AV_RN16(x))
+#  define AV_WB16(p, d) AV_WN16(p, bswap_16(d))
+
+#  define AV_RL16(x)    AV_RN16(x)
+#  define AV_WL16(p, d) AV_WN16(p, d)
+
+#  define AV_RB32(x)    bswap_32(AV_RN32(x))
+#  define AV_WB32(p, d) AV_WN32(p, bswap_32(d))
+
+#  define AV_RL32(x)    AV_RN32(x)
+#  define AV_WL32(p, d) AV_WN32(p, d)
+
+#  define AV_RB64(x)    bswap_64(AV_RN64(x))
+#  define AV_WB64(p, d) AV_WN64(p, bswap_64(d))
+
+#  define AV_RL64(x)    AV_RN64(x)
+#  define AV_WL64(p, d) AV_WN64(p, d)
+# endif
+#else /* HAVE_FAST_UNALIGNED */
+#define AV_RB16(x)  ((((const uint8_t*)(x))[0] << 8) | ((const uint8_t*)(x))[1])
+#define AV_WB16(p, d) do { \
+                    ((uint8_t*)(p))[1] = (d); \
+                    ((uint8_t*)(p))[0] = (d)>>8; } while(0)
+
+#define AV_RL16(x)  ((((const uint8_t*)(x))[1] << 8) | \
+                      ((const uint8_t*)(x))[0])
+#define AV_WL16(p, d) do { \
+                    ((uint8_t*)(p))[0] = (d); \
+                    ((uint8_t*)(p))[1] = (d)>>8; } while(0)
+
+#define AV_RB32(x)  ((((const uint8_t*)(x))[0] << 24) | \
+                     (((const uint8_t*)(x))[1] << 16) | \
+                     (((const uint8_t*)(x))[2] <<  8) | \
+                      ((const uint8_t*)(x))[3])
+#define AV_WB32(p, d) do { \
+                    ((uint8_t*)(p))[3] = (d); \
+                    ((uint8_t*)(p))[2] = (d)>>8; \
+                    ((uint8_t*)(p))[1] = (d)>>16; \
+                    ((uint8_t*)(p))[0] = (d)>>24; } while(0)
+
+#define AV_RL32(x) ((((const uint8_t*)(x))[3] << 24) | \
+                    (((const uint8_t*)(x))[2] << 16) | \
+                    (((const uint8_t*)(x))[1] <<  8) | \
+                     ((const uint8_t*)(x))[0])
+#define AV_WL32(p, d) do { \
+                    ((uint8_t*)(p))[0] = (d); \
+                    ((uint8_t*)(p))[1] = (d)>>8; \
+                    ((uint8_t*)(p))[2] = (d)>>16; \
+                    ((uint8_t*)(p))[3] = (d)>>24; } while(0)
+
+#define AV_RB64(x)  (((uint64_t)((const uint8_t*)(x))[0] << 56) | \
+                     ((uint64_t)((const uint8_t*)(x))[1] << 48) | \
+                     ((uint64_t)((const uint8_t*)(x))[2] << 40) | \
+                     ((uint64_t)((const uint8_t*)(x))[3] << 32) | \
+                     ((uint64_t)((const uint8_t*)(x))[4] << 24) | \
+                     ((uint64_t)((const uint8_t*)(x))[5] << 16) | \
+                     ((uint64_t)((const uint8_t*)(x))[6] <<  8) | \
+                      (uint64_t)((const uint8_t*)(x))[7])
+#define AV_WB64(p, d) do { \
+                    ((uint8_t*)(p))[7] = (d);     \
+                    ((uint8_t*)(p))[6] = (d)>>8;  \
+                    ((uint8_t*)(p))[5] = (d)>>16; \
+                    ((uint8_t*)(p))[4] = (d)>>24; \
+                    ((uint8_t*)(p))[3] = (d)>>32; \
+                    ((uint8_t*)(p))[2] = (d)>>40; \
+                    ((uint8_t*)(p))[1] = (d)>>48; \
+                    ((uint8_t*)(p))[0] = (d)>>56; } while(0)
+
+#define AV_RL64(x)  (((uint64_t)((const uint8_t*)(x))[7] << 56) | \
+                     ((uint64_t)((const uint8_t*)(x))[6] << 48) | \
+                     ((uint64_t)((const uint8_t*)(x))[5] << 40) | \
+                     ((uint64_t)((const uint8_t*)(x))[4] << 32) | \
+                     ((uint64_t)((const uint8_t*)(x))[3] << 24) | \
+                     ((uint64_t)((const uint8_t*)(x))[2] << 16) | \
+                     ((uint64_t)((const uint8_t*)(x))[1] <<  8) | \
+                      (uint64_t)((const uint8_t*)(x))[0])
+#define AV_WL64(p, d) do { \
+                    ((uint8_t*)(p))[0] = (d);     \
+                    ((uint8_t*)(p))[1] = (d)>>8;  \
+                    ((uint8_t*)(p))[2] = (d)>>16; \
+                    ((uint8_t*)(p))[3] = (d)>>24; \
+                    ((uint8_t*)(p))[4] = (d)>>32; \
+                    ((uint8_t*)(p))[5] = (d)>>40; \
+                    ((uint8_t*)(p))[6] = (d)>>48; \
+                    ((uint8_t*)(p))[7] = (d)>>56; } while(0)
+#endif  /* HAVE_FAST_UNALIGNED */
+
+#define AV_RB24(x)  ((((const uint8_t*)(x))[0] << 16) | \
+                     (((const uint8_t*)(x))[1] <<  8) | \
+                      ((const uint8_t*)(x))[2])
+#define AV_WB24(p, d) do { \
+                    ((uint8_t*)(p))[2] = (d); \
+                    ((uint8_t*)(p))[1] = (d)>>8; \
+                    ((uint8_t*)(p))[0] = (d)>>16; } while(0)
+
+#define AV_RL24(x)  ((((const uint8_t*)(x))[2] << 16) | \
+                     (((const uint8_t*)(x))[1] <<  8) | \
+                      ((const uint8_t*)(x))[0])
+#define AV_WL24(p, d) do { \
+                    ((uint8_t*)(p))[0] = (d); \
+                    ((uint8_t*)(p))[1] = (d)>>8; \
+                    ((uint8_t*)(p))[2] = (d)>>16; } while(0)
+
+#endif /* AVUTIL_INTREADWRITE_H */

apps/codecs/libatrac/libavutil/log.c

@@ -0,0 +1,89 @@
+/*
+ * log functions
+ * Copyright (c) 2003 Michel Bardiaux
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavutil/log.c
+ * logging functions
+ */
+
+#include "avutil.h"
+#include "log.h"
+
+int av_log_level = AV_LOG_INFO;
+
+void av_log_default_callback(void* ptr, int level, const char* fmt, va_list vl)
+{
+    static int print_prefix=1;
+    static int count;
+    static char line[1024], prev[1024];
+    AVClass* avc= ptr ? *(AVClass**)ptr : NULL;
+    if(level>av_log_level)
+        return;
+#undef fprintf
+    if(print_prefix && avc) {
+        snprintf(line, sizeof(line), "[%s @ %p]", avc->item_name(ptr), ptr);
+    }else
+        line[0]=0;
+
+    vsnprintf(line + strlen(line), sizeof(line) - strlen(line), fmt, vl);
+
+    print_prefix= line[strlen(line)-1] == '\n';
+    if(print_prefix && !strcmp(line, prev)){
+        count++;
+        return;
+    }
+    if(count>0){
+        fprintf(stderr, "    Last message repeated %d times\n", count);
+        count=0;
+    }
+    fputs(line, stderr);
+    strcpy(prev, line);
+}
+
+static void (*av_log_callback)(void*, int, const char*, va_list) = av_log_default_callback;
+
+void av_log(void* avcl, int level, const char *fmt, ...)
+{
+    va_list vl;
+    va_start(vl, fmt);
+    av_vlog(avcl, level, fmt, vl);
+    va_end(vl);
+}
+
+void av_vlog(void* avcl, int level, const char *fmt, va_list vl)
+{
+    av_log_callback(avcl, level, fmt, vl);
+}
+
+int av_log_get_level(void)
+{
+    return av_log_level;
+}
+
+void av_log_set_level(int level)
+{
+    av_log_level = level;
+}
+
+void av_log_set_callback(void (*callback)(void*, int, const char*, va_list))
+{
+    av_log_callback = callback;
+}

apps/codecs/libatrac/libavutil/log.h

@@ -0,0 +1,116 @@
+/*
+ * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVUTIL_LOG_H
+#define AVUTIL_LOG_H
+
+#include <stdarg.h>
+#include "avutil.h"
+
+/**
+ * Describes the class of an AVClass context structure. That is an
+ * arbitrary struct of which the first field is a pointer to an
+ * AVClass struct (e.g. AVCodecContext, AVFormatContext etc.).
+ */
+typedef struct AVCLASS AVClass;
+struct AVCLASS {
+    /**
+     * The name of the class; usually it is the same name as the
+     * context structure type to which the AVClass is associated.
+     */
+    const char* class_name;
+
+    /**
+     * A pointer to a function which returns the name of a context
+     * instance \p ctx associated with the class.
+     */
+    const char* (*item_name)(void* ctx);
+
+    /**
+     * a pointer to the first option specified in the class if any or NULL
+     *
+     * @see av_set_default_options()
+     */
+    const struct AVOption *option;
+};
+
+/* av_log API */
+
+#define AV_LOG_QUIET    -8
+
+/**
+ * Something went really wrong and we will crash now.
+ */
+#define AV_LOG_PANIC     0
+
+/**
+ * Something went wrong and recovery is not possible.
+ * For example, no header was found for a format which depends
+ * on headers or an illegal combination of parameters is used.
+ */
+#define AV_LOG_FATAL     8
+
+/**
+ * Something went wrong and cannot losslessly be recovered.
+ * However, not all future data is affected.
+ */
+#define AV_LOG_ERROR    16
+
+/**
+ * Something somehow does not look correct. This may or may not
+ * lead to problems. An example would be the use of '-vstrict -2'.
+ */
+#define AV_LOG_WARNING  24
+
+#define AV_LOG_INFO     32
+#define AV_LOG_VERBOSE  40
+
+/**
+ * Stuff which is only useful for libav* developers.
+ */
+#define AV_LOG_DEBUG    48
+
+/**
+ * Sends the specified message to the log if the level is less than or equal
+ * to the current av_log_level. By default, all logging messages are sent to
+ * stderr. This behavior can be altered by setting a different av_vlog callback
+ * function.
+ *
+ * @param avcl A pointer to an arbitrary struct of which the first field is a
+ * pointer to an AVClass struct.
+ * @param level The importance level of the message, lower values signifying
+ * higher importance.
+ * @param fmt The format string (printf-compatible) that specifies how
+ * subsequent arguments are converted to output.
+ * @see av_vlog
+ */
+#ifdef __GNUC__
+void av_log(void*, int level, const char *fmt, ...) __attribute__ ((__format__ (__printf__, 3, 4)));
+#else
+void av_log(void*, int level, const char *fmt, ...);
+#endif
+
+void av_vlog(void*, int level, const char *fmt, va_list);
+int av_log_get_level(void);
+void av_log_set_level(int);
+void av_log_set_callback(void (*)(void*, int, const char*, va_list));
+void av_log_default_callback(void* ptr, int level, const char* fmt, va_list vl);
+
+#endif /* AVUTIL_LOG_H */

apps/codecs/libatrac/libavutil/mem.c

@@ -0,0 +1,159 @@
+/*
+ * default memory allocator for libavutil
+ * Copyright (c) 2002 Fabrice Bellard
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavutil/mem.c
+ * default memory allocator for libavutil
+ */
+
+//#include "ffmpeg_config.h"
+
+#include <limits.h>
+#include <stdlib.h>
+#include <string.h>
+#if HAVE_MALLOC_H
+#include <malloc.h>
+#endif
+
+#include "mem.h"
+
+/* here we can use OS-dependent allocation functions */
+#undef free
+#undef malloc
+#undef realloc
+
+/* You can redefine av_malloc and av_free in your project to use your
+   memory allocator. You do not need to suppress this file because the
+   linker will do it automatically. */
+
+void *av_malloc(unsigned int size)
+{
+    void *ptr = NULL;
+#if CONFIG_MEMALIGN_HACK
+    long diff;
+#endif
+
+    /* let's disallow possible ambiguous cases */
+    if(size > (INT_MAX-16) )
+        return NULL;
+
+#if CONFIG_MEMALIGN_HACK
+    ptr = malloc(size+16);
+    if(!ptr)
+        return ptr;
+    diff= ((-(long)ptr - 1)&15) + 1;
+    ptr = (char*)ptr + diff;
+    ((char*)ptr)[-1]= diff;
+#elif HAVE_POSIX_MEMALIGN
+    if (posix_memalign(&ptr,16,size))
+        ptr = NULL;
+#elif HAVE_MEMALIGN
+    ptr = memalign(16,size);
+    /* Why 64?
+       Indeed, we should align it:
+         on 4 for 386
+         on 16 for 486
+         on 32 for 586, PPro - K6-III
+         on 64 for K7 (maybe for P3 too).
+       Because L1 and L2 caches are aligned on those values.
+       But I don't want to code such logic here!
+     */
+     /* Why 16?
+        Because some CPUs need alignment, for example SSE2 on P4, & most RISC CPUs
+        it will just trigger an exception and the unaligned load will be done in the
+        exception handler or it will just segfault (SSE2 on P4).
+        Why not larger? Because I did not see a difference in benchmarks ...
+     */
+     /* benchmarks with P3
+        memalign(64)+1          3071,3051,3032
+        memalign(64)+2          3051,3032,3041
+        memalign(64)+4          2911,2896,2915
+        memalign(64)+8          2545,2554,2550
+        memalign(64)+16         2543,2572,2563
+        memalign(64)+32         2546,2545,2571
+        memalign(64)+64         2570,2533,2558
+
+        BTW, malloc seems to do 8-byte alignment by default here.
+     */
+#else
+    ptr = malloc(size);
+#endif
+    return ptr;
+}
+
+void *av_realloc(void *ptr, unsigned int size)
+{
+#if CONFIG_MEMALIGN_HACK
+    int diff;
+#endif
+
+    /* let's disallow possible ambiguous cases */
+    if(size > (INT_MAX-16) )
+        return NULL;
+
+#if CONFIG_MEMALIGN_HACK
+    //FIXME this isn't aligned correctly, though it probably isn't needed
+    if(!ptr) return av_malloc(size);
+    diff= ((char*)ptr)[-1];
+    return (char*)realloc((char*)ptr - diff, size + diff) + diff;
+#else
+    return realloc(ptr, size);
+#endif
+}
+
+void av_free(void *ptr)
+{
+    /* XXX: this test should not be needed on most libcs */
+    if (ptr)
+#if CONFIG_MEMALIGN_HACK
+        free((char*)ptr - ((char*)ptr)[-1]);
+#else
+        free(ptr);
+#endif
+}
+
+void av_freep(void *arg)
+{
+    void **ptr= (void**)arg;
+    av_free(*ptr);
+    *ptr = NULL;
+}
+
+void *av_mallocz(unsigned int size)
+{
+    void *ptr = av_malloc(size);
+    if (ptr)
+        memset(ptr, 0, size);
+    return ptr;
+}
+
+char *av_strdup(const char *s)
+{
+    char *ptr= NULL;
+    if(s){
+        int len = strlen(s) + 1;
+        ptr = av_malloc(len);
+        if (ptr)
+            memcpy(ptr, s, len);
+    }
+    return ptr;
+}
+

apps/codecs/libatrac/libavutil/mem.h

@@ -0,0 +1,104 @@
+/*
+ * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file libavutil/mem.h
+ * memory handling functions
+ */
+
+#ifndef AVUTIL_MEM_H
+#define AVUTIL_MEM_H
+
+#include "common.h"
+
+#if AV_GCC_VERSION_AT_LEAST(3,1)
+    #define av_malloc_attrib __attribute__((__malloc__))
+#else
+    #define av_malloc_attrib
+#endif
+
+#if (!defined(__ICC) || __ICC > 1100) && AV_GCC_VERSION_AT_LEAST(4,3)
+    #define av_alloc_size(n) __attribute__((alloc_size(n)))
+#else
+    #define av_alloc_size(n)
+#endif
+
+/**
+ * Allocates a block of \p size bytes with alignment suitable for all
+ * memory accesses (including vectors if available on the CPU).
+ * @param size Size in bytes for the memory block to be allocated.
+ * @return Pointer to the allocated block, NULL if the block cannot
+ * be allocated.
+ * @see av_mallocz()
+ */
+void *av_malloc(unsigned int size) av_malloc_attrib av_alloc_size(1);
+
+/**
+ * Allocates or reallocates a block of memory.
+ * If \p ptr is NULL and \p size > 0, allocates a new block. If \p
+ * size is zero, frees the memory block pointed to by \p ptr.
+ * @param size Size in bytes for the memory block to be allocated or
+ * reallocated.
+ * @param ptr Pointer to a memory block already allocated with
+ * av_malloc(z)() or av_realloc() or NULL.
+ * @return Pointer to a newly reallocated block or NULL if the block
+ * cannot be reallocated or the function is used to free the memory block.
+ * @see av_fast_realloc()
+ */
+void *av_realloc(void *ptr, unsigned int size) av_alloc_size(2);
+
+/**
+ * Frees a memory block which has been allocated with av_malloc(z)() or
+ * av_realloc().
+ * @param ptr Pointer to the memory block which should be freed.
+ * @note ptr = NULL is explicitly allowed.
+ * @note It is recommended that you use av_freep() instead.
+ * @see av_freep()
+ */
+void av_free(void *ptr);
+
+/**
+ * Allocates a block of \p size bytes with alignment suitable for all
+ * memory accesses (including vectors if available on the CPU) and
+ * zeroes all the bytes of the block.
+ * @param size Size in bytes for the memory block to be allocated.
+ * @return Pointer to the allocated block, NULL if it cannot be allocated.
+ * @see av_malloc()
+ */
+void *av_mallocz(unsigned int size) av_malloc_attrib av_alloc_size(1);
+
+/**
+ * Duplicates the string \p s.
+ * @param s string to be duplicated
+ * @return Pointer to a newly allocated string containing a
+ * copy of \p s or NULL if the string cannot be allocated.
+ */
+char *av_strdup(const char *s) av_malloc_attrib;
+
+/**
+ * Frees a memory block which has been allocated with av_malloc(z)() or
+ * av_realloc() and set the pointer pointing to it to NULL.
+ * @param ptr Pointer to the pointer to the memory block which should
+ * be freed.
+ * @see av_free()
+ */
+void av_freep(void *ptr);
+
+#endif /* AVUTIL_MEM_H */

apps/codecs/libatrac/mdct.c

@@ -0,0 +1,245 @@
+/*
+ * MDCT/IMDCT transforms
+ * Copyright (c) 2002 Fabrice Bellard
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+#include "dsputil.h"
+
+#ifndef M_E
+#define M_E            2.7182818284590452354   /* e */
+#endif
+#ifndef M_LN2
+#define M_LN2          0.69314718055994530942  /* log_e 2 */
+#endif
+#ifndef M_LN10
+#define M_LN10         2.30258509299404568402  /* log_e 10 */
+#endif
+#ifndef M_PI
+#define M_PI           3.14159265358979323846  /* pi */
+#endif
+#ifndef M_SQRT1_2
+#define M_SQRT1_2      0.70710678118654752440  /* 1/sqrt(2) */
+#endif
+
+/**
+ * @file libavcodec/mdct.c
+ * MDCT/IMDCT transforms.
+ */
+
+// Generate a Kaiser-Bessel Derived Window.
+#define BESSEL_I0_ITER 50 // default: 50 iterations of Bessel I0 approximation
+av_cold void ff_kbd_window_init(float *window, float alpha, int n)
+{
+   int i, j;
+   double sum = 0.0, bessel, tmp;
+   double local_window[n];
+   double alpha2 = (alpha * M_PI / n) * (alpha * M_PI / n);
+
+   for (i = 0; i < n; i++) {
+       tmp = i * (n - i) * alpha2;
+       bessel = 1.0;
+       for (j = BESSEL_I0_ITER; j > 0; j--)
+           bessel = bessel * tmp / (j * j) + 1;
+       sum += bessel;
+       local_window[i] = sum;
+   }
+
+   sum++;
+   for (i = 0; i < n; i++)
+       window[i] = sqrt(local_window[i] / sum);
+}
+
+DECLARE_ALIGNED(16, float, ff_sine_128 [ 128]);
+DECLARE_ALIGNED(16, float, ff_sine_256 [ 256]);
+DECLARE_ALIGNED(16, float, ff_sine_512 [ 512]);
+DECLARE_ALIGNED(16, float, ff_sine_1024[1024]);
+DECLARE_ALIGNED(16, float, ff_sine_2048[2048]);
+DECLARE_ALIGNED(16, float, ff_sine_4096[4096]);
+float *ff_sine_windows[6] = {
+    ff_sine_128, ff_sine_256, ff_sine_512, ff_sine_1024, ff_sine_2048, ff_sine_4096
+};
+
+// Generate a sine window.
+av_cold void ff_sine_window_init(float *window, int n) {
+    int i;
+    for(i = 0; i < n; i++)
+        window[i] = sinf((i + 0.5) * (M_PI / (2.0 * n)));
+}
+
+/**
+ * init MDCT or IMDCT computation.
+ */
+av_cold int ff_mdct_init(MDCTContext *s, int nbits, int inverse)
+{
+    int n, n4, i;
+    double alpha;
+
+    memset(s, 0, sizeof(*s));
+    n = 1 << nbits;
+    s->nbits = nbits;
+    s->n = n;
+    n4 = n >> 2;
+    s->tcos = av_malloc(n4 * sizeof(FFTSample));
+    if (!s->tcos)
+        goto fail;
+    s->tsin = av_malloc(n4 * sizeof(FFTSample));
+    if (!s->tsin)
+        goto fail;
+
+    for(i=0;i<n4;i++) {
+        alpha = 2 * M_PI * (i + 1.0 / 8.0) / n;
+        s->tcos[i] = -cos(alpha);
+        s->tsin[i] = -sin(alpha);
+    }
+    if (ff_fft_init(&s->fft, s->nbits - 2, inverse) < 0)
+        goto fail;
+    return 0;
+ fail:
+    av_freep(&s->tcos);
+    av_freep(&s->tsin);
+    return -1;
+}
+
+/* complex multiplication: p = a * b */
+#define CMUL(pre, pim, are, aim, bre, bim) \
+{\
+    FFTSample _are = (are);\
+    FFTSample _aim = (aim);\
+    FFTSample _bre = (bre);\
+    FFTSample _bim = (bim);\
+    (pre) = _are * _bre - _aim * _bim;\
+    (pim) = _are * _bim + _aim * _bre;\
+}
+
+/**
+ * Compute the middle half of the inverse MDCT of size N = 2^nbits,
+ * thus excluding the parts that can be derived by symmetry
+ * @param output N/2 samples
+ * @param input N/2 samples
+ */
+void ff_imdct_half_c(MDCTContext *s, FFTSample *output, const FFTSample *input)
+{
+    int k, n8, n4, n2, n, j;
+    const uint16_t *revtab = s->fft.revtab;
+    const FFTSample *tcos = s->tcos;
+    const FFTSample *tsin = s->tsin;
+    const FFTSample *in1, *in2;
+    FFTComplex *z = (FFTComplex *)output;
+
+    n = 1 << s->nbits;
+    n2 = n >> 1;
+    n4 = n >> 2;
+    n8 = n >> 3;
+
+    /* pre rotation */
+    in1 = input;
+    in2 = input + n2 - 1;
+    for(k = 0; k < n4; k++) {
+        j=revtab[k];
+        CMUL(z[j].re, z[j].im, *in2, *in1, tcos[k], tsin[k]);
+        in1 += 2;
+        in2 -= 2;
+    }
+    ff_fft_calc(&s->fft, z);
+
+    /* post rotation + reordering */
+    output += n4;
+    for(k = 0; k < n8; k++) {
+        FFTSample r0, i0, r1, i1;
+        CMUL(r0, i1, z[n8-k-1].im, z[n8-k-1].re, tsin[n8-k-1], tcos[n8-k-1]);
+        CMUL(r1, i0, z[n8+k  ].im, z[n8+k  ].re, tsin[n8+k  ], tcos[n8+k  ]);
+        z[n8-k-1].re = r0;
+        z[n8-k-1].im = i0;
+        z[n8+k  ].re = r1;
+        z[n8+k  ].im = i1;
+    }
+}
+
+/**
+ * Compute inverse MDCT of size N = 2^nbits
+ * @param output N samples
+ * @param input N/2 samples
+ */
+void ff_imdct_calc_c(MDCTContext *s, FFTSample *output, const FFTSample *input)
+{
+    int k;
+    int n = 1 << s->nbits;
+    int n2 = n >> 1;
+    int n4 = n >> 2;
+
+    ff_imdct_half_c(s, output+n4, input);
+
+    for(k = 0; k < n4; k++) {
+        output[k] = -output[n2-k-1];
+        output[n-k-1] = output[n2+k];
+    }
+}
+
+/**
+ * Compute MDCT of size N = 2^nbits
+ * @param input N samples
+ * @param out N/2 samples
+ */
+void ff_mdct_calc(MDCTContext *s, FFTSample *out, const FFTSample *input)
+{
+    int i, j, n, n8, n4, n2, n3;
+    FFTSample re, im;
+    const uint16_t *revtab = s->fft.revtab;
+    const FFTSample *tcos = s->tcos;
+    const FFTSample *tsin = s->tsin;
+    FFTComplex *x = (FFTComplex *)out;
+
+    n = 1 << s->nbits;
+    n2 = n >> 1;
+    n4 = n >> 2;
+    n8 = n >> 3;
+    n3 = 3 * n4;
+
+    /* pre rotation */
+    for(i=0;i<n8;i++) {
+        re = -input[2*i+3*n4] - input[n3-1-2*i];
+        im = -input[n4+2*i] + input[n4-1-2*i];
+        j = revtab[i];
+        CMUL(x[j].re, x[j].im, re, im, -tcos[i], tsin[i]);
+
+        re = input[2*i] - input[n2-1-2*i];
+        im = -(input[n2+2*i] + input[n-1-2*i]);
+        j = revtab[n8 + i];
+        CMUL(x[j].re, x[j].im, re, im, -tcos[n8 + i], tsin[n8 + i]);
+    }
+
+    ff_fft_calc(&s->fft, x);
+
+    /* post rotation */
+    for(i=0;i<n8;i++) {
+        FFTSample r0, i0, r1, i1;
+        CMUL(i1, r0, x[n8-i-1].re, x[n8-i-1].im, -tsin[n8-i-1], -tcos[n8-i-1]);
+        CMUL(i0, r1, x[n8+i  ].re, x[n8+i  ].im, -tsin[n8+i  ], -tcos[n8+i  ]);
+        x[n8-i-1].re = r0;
+        x[n8-i-1].im = i0;
+        x[n8+i  ].re = r1;
+        x[n8+i  ].im = i1;
+    }
+}
+
+av_cold void ff_mdct_end(MDCTContext *s)
+{
+    av_freep(&s->tcos);
+    av_freep(&s->tsin);
+    ff_fft_end(&s->fft);
+}

apps/codecs/librm/rm.c

@@ -29,7 +29,7 @@
 
 #define SWAP(a, b) do{uint8_t SWAP_tmp= b; b= a; a= SWAP_tmp;}while(0)
 
-static void advance_buffer(uint8_t **buf, int val)
+void advance_buffer(uint8_t **buf, int val)
 {
     *buf += val;
 }
@@ -237,10 +237,40 @@ static int real_read_audio_stream_info(int fd, RMContext *rmctx)
            skipped += 1;
        }
   
-       read_uint32be(fd, &rmctx->extradata_size);
+       switch(fourcc) {
-       skipped += 4;
+           case FOURCC('c','o','o','k'):               
-       read(fd, rmctx->codec_extradata, rmctx->extradata_size);
+               rmctx->codec_type = CODEC_COOK;
-       skipped += rmctx->extradata_size;
+               read_uint32be(fd, &rmctx->extradata_size);
+               skipped += 4;
+               read(fd, rmctx->codec_extradata, rmctx->extradata_size);
+               skipped += rmctx->extradata_size;
+               break;
+
+           case FOURCC('a','t','r','c'):  
+DEBUGF("WERE HERE\n");             
+               rmctx->codec_type = CODEC_ATRAC;
+               read_uint32be(fd, &rmctx->extradata_size);
+               skipped += 4;
+               read(fd, rmctx->codec_extradata, rmctx->extradata_size);
+               skipped += rmctx->extradata_size;
+               break;
+
+           case FOURCC('r','a','a','c'):
+           case FOURCC('r','a','c','p'):
+               rmctx->codec_type = CODEC_AAC;
+               read_uint32be(fd, &rmctx->extradata_size);
+               skipped += 4;
+               read(fd, rmctx->codec_extradata, rmctx->extradata_size);
+               skipped += rmctx->extradata_size;
+               break;
+
+           case FOURCC('d','n','e','t'):
+               rmctx->codec_type = CODEC_AC3;
+               break;
+
+           default: /* Not a supported codec */
+               return -1;
+       }
        
        
        DEBUGF("        flavor = %d\n",flavor);
@@ -252,8 +282,10 @@ static int real_read_audio_stream_info(int fd, RMContext *rmctx)
        DEBUGF("        channels= %d\n",rmctx->nb_channels);
        DEBUGF("        fourcc = %s\n",fourcc2str(fourcc));
        DEBUGF("        codec_extra_data_length = %d\n",rmctx->extradata_size);
-       DEBUGF("        codec_extradata :\n");
+       if(rmctx->codec_type == CODEC_COOK) {
-       print_cook_extradata(rmctx);
+           DEBUGF("        cook_extradata :\n");
+           print_cook_extradata(rmctx);
+       }
 
     }
 
@@ -530,7 +562,7 @@ int rm_get_packet(uint8_t **src,RMContext *rmctx, RMPacket *pkt)
         
         advance_buffer(src,12);
         consumed += 12;
-        if (rmctx->codec_type == CODEC_COOK) {
+        if (rmctx->codec_type == CODEC_COOK || rmctx->codec_type == CODEC_ATRAC) {
             for(x = 0 ; x < w/sps; x++)
             {
                 place = sps*(h*x+((h+1)/2)*(y&1)+(y>>1)); 

apps/codecs/librm/rm.h

@@ -31,7 +31,8 @@
 enum codecs {
     CODEC_COOK, 
     CODEC_AAC,
-    CODEC_AC3
+    CODEC_AC3,
+    CODEC_ATRAC
 };
 
 typedef struct rm_packet