vlc_deinterlace.cpp

linux下实现视频播放的播放器

/*
 *  Copyright (C) 2005-2007  gulikoza
 *
 *  Taken from VideoLAN project, original copyright:
 *
 *  Copyright (C) 2000, 2001, 2002, 2003 the VideoLAN team
 *  Author: Sam Hocevar <[e-mail]>
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 */

/* $Id$ */

#include "iptv.h"
#include "VLC_deinterlace.h"

#ifdef HAVE_ALTIVEC_H
#include <altivec.h>
#endif

/*
#if (C_HOSTCPU == X86) || (C_HOSTCPU == X86_64)
#include "mmx.h"
#endif
*/

// Function pointers
MERGE Merge;
END_MERGE EndMerge;

void MergeGeneric( void *_p_dest, const void *_p_s1,
                          const void *_p_s2, size_t i_bytes )
{
    uint8_t* p_dest = (uint8_t*)_p_dest;
    const uint8_t *p_s1 = (const uint8_t *)_p_s1;
    const uint8_t *p_s2 = (const uint8_t *)_p_s2;
    uint8_t* p_end = p_dest + i_bytes - 8;

    while( p_dest < p_end )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }

    p_end += 8;

    while( p_dest < p_end )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }
}

#if (C_HOSTCPU == X86) || (C_HOSTCPU == X86_64)
void MergeMMXEXT( void *_p_dest, const void *_p_s1, const void *_p_s2,
                         size_t i_bytes )
{
    uint8_t* p_dest = (uint8_t*)_p_dest;
    const uint8_t *p_s1 = (const uint8_t *)_p_s1;
    const uint8_t *p_s2 = (const uint8_t *)_p_s2;
    uint8_t* p_end = p_dest + i_bytes - 8;
    while( p_dest < p_end )
    {
        __asm__  __volatile__( "movq %2,%%mm1;"
                               "pavgb %1, %%mm1;"
                               "movq %%mm1, %0" :"=m" (*p_dest):
                                                 "m" (*p_s1),
                                                 "m" (*p_s2) );
        p_dest += 8;
        p_s1 += 8;
        p_s2 += 8;
    }

    p_end += 8;

    while( p_dest < p_end )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }
}

void Merge3DNow( void *_p_dest, const void *_p_s1, const void *_p_s2,
                        size_t i_bytes )
{
    uint8_t* p_dest = (uint8_t*)_p_dest;
    const uint8_t *p_s1 = (const uint8_t *)_p_s1;
    const uint8_t *p_s2 = (const uint8_t *)_p_s2;
    uint8_t* p_end = p_dest + i_bytes - 8;
    while( p_dest < p_end )
    {
        __asm__  __volatile__( "movq %2,%%mm1;"
                               "pavgusb %1, %%mm1;"
                               "movq %%mm1, %0" :"=m" (*p_dest):
                                                 "m" (*p_s1),
                                                 "m" (*p_s2) );
        p_dest += 8;
        p_s1 += 8;
        p_s2 += 8;
    }

    p_end += 8;

    while( p_dest < p_end )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }
}

void MergeSSE2( void *_p_dest, const void *_p_s1, const void *_p_s2,
                       size_t i_bytes )
{
    uint8_t* p_dest = (uint8_t*)_p_dest;
    const uint8_t *p_s1 = (const uint8_t *)_p_s1;
    const uint8_t *p_s2 = (const uint8_t *)_p_s2;
    uint8_t* p_end;

    /* Use C until the first 16-bytes aligned source pixel */
    while( (intptr_t)p_s1 & 0xF )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }
    p_end = p_dest + i_bytes - 16;
    while( p_dest < p_end )
    {
        __asm__  __volatile__( "movdqu %2,%%xmm1;"
                               "pavgb %1, %%xmm1;"
                               "movdqu %%xmm1, %0" :"=m" (*p_dest):
                                                 "m" (*p_s1),
                                                 "m" (*p_s2) );
        p_dest += 16;
        p_s1 += 16;
        p_s2 += 16;
    }

    p_end += 16;

    while( p_dest < p_end )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }
}

void EndMMX( void )
{
    __asm__ __volatile__( "emms" :: );
}

void End3DNow( void )
{
    __asm__ __volatile__( "femms" :: );
}
#endif

#ifdef HAVE_ALTIVEC_H
void MergeAltivec( void *_p_dest, const void *_p_s1,
                          const void *_p_s2, size_t i_bytes )
{
    uint8_t *p_dest = (uint8_t *)_p_dest;
    uint8_t *p_s1   = (uint8_t *)_p_s1;
    uint8_t *p_s2   = (uint8_t *)_p_s2;
    uint8_t *p_end  = p_dest + i_bytes - 15;

    /* Use C until the first 16-bytes aligned destination pixel */
    while( (int)p_dest & 0xF )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }

    if( ( (int)p_s1 & 0xF ) | ( (int)p_s2 & 0xF ) )
    {
        /* Unaligned source */
        vector unsigned char s1v, s2v, destv;
        vector unsigned char s1oldv, s2oldv, s1newv, s2newv;
        vector unsigned char perm1v, perm2v;

        perm1v = vec_lvsl( 0, p_s1 );
        perm2v = vec_lvsl( 0, p_s2 );
        s1oldv = vec_ld( 0, p_s1 );
        s2oldv = vec_ld( 0, p_s2 );

        while( p_dest < p_end )
        {
            s1newv = vec_ld( 16, p_s1 );
            s2newv = vec_ld( 16, p_s2 );
            s1v    = vec_perm( s1oldv, s1newv, perm1v );
            s2v    = vec_perm( s2oldv, s2newv, perm2v );
            s1oldv = s1newv;
            s2oldv = s2newv;
            destv  = vec_avg( s1v, s2v );
            vec_st( destv, 0, p_dest );

            p_s1   += 16;
            p_s2   += 16;
            p_dest += 16;
        }
    }
    else
    {
        /* Aligned source */
        vector unsigned char s1v, s2v, destv;

        while( p_dest < p_end )
        {
            s1v   = vec_ld( 0, p_s1 );
            s2v   = vec_ld( 0, p_s2 );
            destv = vec_avg( s1v, s2v );
            vec_st( destv, 0, p_dest );

            p_s1   += 16;
            p_s2   += 16;
            p_dest += 16;
        }
    }

    p_end += 15;

    while( p_dest < p_end )
    {
        *p_dest++ = ( (uint16_t)(*p_s1++) + (uint16_t)(*p_s2++) ) >> 1;
    }
}
#endif