wmadeci.c

wma定点解码算法

}

#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;\

#   define UPDATE_CACHE(name, gb)\
        name##_cache= unaligned32_be( ((uint8_t *)(gb)->buffer)+(name##_index>>3) ) << (name##_index&0x07);\

#   define SKIP_CACHE(name, gb, num)\
        name##_cache <<= (num);\

#   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) ;

#   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->index;
}
#elif defined LIBMPEG2_BITSTREAM_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;\

#ifdef LIBMPEG2_BITSTREAM_READER_HACK

#   define UPDATE_CACHE(name, gb)\
    if(name##_bit_count >= 0){\
        name##_cache+= (int)be2me_16(*(uint16_t*)name##_buffer_ptr) << name##_bit_count;\
        ((uint16_t*)name##_buffer_ptr)++;\
        name##_bit_count-= 16;\
    }\

#else

#   define UPDATE_CACHE(name, gb)\
    if(name##_bit_count >= 0){\
        name##_cache+= ((name##_buffer_ptr[0]<<8) + name##_buffer_ptr[1]) << name##_bit_count;\
        name##_buffer_ptr+=2;\
        name##_bit_count-= 16;\
    }\

#endif

#   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;
}

#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;\
    }\

#ifdef 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;
}

#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 tmp;
    register int32_t cache;
    OPEN_READER(re, s)
    UPDATE_CACHE(re, s)
    cache = GET_CACHE(re,s);
    if ((int32_t)cache<0)
    {
        tmp = NEG_USR32(cache,n);
    }
    else
    {
        tmp = - NEG_USR32(~cache,n);
    }
    LAST_SKIP_BITS(re, s, n)
    CLOSE_READER(re, s)
    return tmp;
}

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 0-17 bits.
 * Note, the alt bitstream reader can read upto 25 bits, but the libmpeg2 reader cant
 */
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;
}


/**
 *
 * 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)\
}

static inline void skip_bits(GetBitContext *s, int n)
{
    OPEN_READER(re, s)
    UPDATE_CACHE(re, s)
    LAST_SKIP_BITS(re, s, n)
    CLOSE_READER(re, s)
}

void align_get_bits(GetBitContext *s)
{
    int n= (-get_bits_count(s)) & 7;
    if(n)
        skip_bits(s, n);
}

/**
 * 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
 */
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 = 0;
    }
    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
#ifdef LIBMPEG2_BITSTREAM_READER_HACK

    if ((int)buffer&1)
    {
        /* word alignment */
        s->cache = (*buffer++)<<24;
        s->buffer_ptr = buffer;
        s->bit_count = 16-8;
    }
    else
#endif

    {
        s->buffer_ptr = buffer;
        s->bit_count = 16;
        s->cache = 0;
    }
#elif defined A32_BITSTREAM_READER
    s->buffer_ptr = (uint32_t*)buffer;
    s->bit_count = 32;
    s->cache0 = 0;
    s->cache1 = 0;
#endif

    {
        OPEN_READER(re, s)
        UPDATE_CACHE(re, s)
        UPDATE_CACHE(re, s)
        CLOSE_READER(re, s)
    }
#ifdef A32_BITSTREAM_READER
    s->cache1 = 0;
#endif
}

/**
 * Memory allocation of size byte with alignment suitable for all
 * memory accesses (including vectors if available on the
 * CPU). av_malloc(0) must return a non NULL pointer.
 */
void *av_malloc(unsigned int size)
{
    void *ptr;

#if defined (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 didnt 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;
}

/**
 * av_realloc semantics (same as glibc): if ptr is NULL and size > 0,
 * identical to malloc(size). If size is zero, it is identical to
 * free(ptr) and NULL is returned.
 */
void *av_realloc(void *ptr, unsigned int size)
{
    return realloc(ptr, size);
}

/* NOTE: ptr = NULL is explicitly allowed */
void av_free(void *ptr)
{
    /* XXX: this test should not be needed on most libcs */
    if (ptr)
        free(ptr);
}

/* cannot call it directly because of 'void **' casting is not automatic */
void __av_freepy(void **ptr)
{
    av_free(*ptr);
    *ptr = NULL;
}
#define av_freep(p) __av_freepy((void **)(p))

/**
 * The size of the FFT is 2^nbits. If inverse is TRUE, inverse FFT is
 * done
 */
int fft_inits(FFTContext *s, int nbits, int inverse)
{
    int i, j, m, n;
    fixed32 alpha, c1, s1;
    int s2;

    s->nbits = nbits;
    n = 1 << nbits;

    s->exptab = av_malloc((n >> 1) * 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 : -1;

    for(i=0;i<(n/2);++i)
    {
        fixed32 ifix = itofix32(i);
        fixed32 nfix = itofix32(n);
        fixed32 res = fixdiv32(ifix,nfix);
        fixed32 pi2 = fixmul32(0x20000, M_PI_F);
        alpha = fixmul32(pi2, res);
        c1 = fixcos32(alpha);
        s1 = fixsin32(alpha) * s2;
        s->exptab[i].re = c1;
        s->exptab[i].im = s1;
    }

    s->fft_calc = fft_calc;
    s->exptab1 = NULL;
    /* compute constant table for HAVE_SSE version */
#if (defined(HAVE_MMX) && defined(HAVE_BUILTIN_VECTOR)) || defined(HAVE_ALTIVEC)

    {
        int has_vectors = 0;

#if defined(HAVE_MMX)

        has_vectors = mm_support() & MM_SSE;
#endif
#if defined(HAVE_ALTIVEC) && !defined(ALTIVEC_USE_REFERENCE_C_CODE)

        has_vectors = mm_support() & MM_ALTIVEC;
#endif

        if (has_vectors)
        {
            int np, nblocks, np2, l;
            FFTComplex *q;

            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);
#if defined(HAVE_MMX)

            s->fft_calc = fft_calc_sse;
#else

            s->fft_calc = fft_calc_altivec;
#endif

        }
    }
#endif