view eacmv.c @ 12197:fbf4d5b1b664 libavcodec

Remove FF_MM_SSE2/3 flags for CPUs where this is generally not faster than regular MMX code. Examples of this are the Core1 CPU. Instead, set a new flag, FF_MM_SSE2/3SLOW, which can be checked for particular SSE2/3 functions that have been checked specifically on such CPUs and are actually faster than their MMX counterparts. In addition, use this flag to enable particular VP8 and LPC SSE2 functions that are faster than their MMX counterparts. Based on a patch by Loren Merritt <lorenm AT u washington edu>.
author rbultje
date Mon, 19 Jul 2010 22:38:23 +0000
parents 7dd2a45249a9
children 914f484bb476
line wrap: on
line source

/*
 * Electronic Arts CMV Video Decoder
 * Copyright (c) 2007-2008 Peter Ross
 *
 * 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 St, Fifth Floor, Boston, MA  02110-1301  USA
 */

/**
 * @file
 * Electronic Arts CMV Video Decoder
 * by Peter Ross (pross@xvid.org)
 *
 * Technical details here:
 * http://wiki.multimedia.cx/index.php?title=Electronic_Arts_CMV
 */

#include "libavutil/intreadwrite.h"
#include "avcodec.h"

typedef struct CmvContext {
    AVCodecContext *avctx;
    AVFrame frame;        ///< current
    AVFrame last_frame;   ///< last
    AVFrame last2_frame;  ///< second-last
    int width, height;
    unsigned int palette[AVPALETTE_COUNT];
} CmvContext;

static av_cold int cmv_decode_init(AVCodecContext *avctx){
    CmvContext *s = avctx->priv_data;
    s->avctx = avctx;
    avctx->pix_fmt = PIX_FMT_PAL8;
    return 0;
}

static void cmv_decode_intra(CmvContext * s, const uint8_t *buf, const uint8_t *buf_end){
    unsigned char *dst = s->frame.data[0];
    int i;

    for (i=0; i < s->avctx->height && buf+s->avctx->width<=buf_end; i++) {
        memcpy(dst, buf, s->avctx->width);
        dst += s->frame.linesize[0];
        buf += s->avctx->width;
    }
}

static void cmv_motcomp(unsigned char *dst, int dst_stride,
                        const unsigned char *src, int src_stride,
                        int x, int y,
                        int xoffset, int yoffset,
                        int width, int height){
    int i,j;

    for(j=y;j<y+4;j++)
    for(i=x;i<x+4;i++)
    {
        if (i+xoffset>=0 && i+xoffset<width &&
            j+yoffset>=0 && j+yoffset<height) {
            dst[j*dst_stride + i] = src[(j+yoffset)*src_stride + i+xoffset];
        }else{
            dst[j*dst_stride + i] = 0;
        }
    }
}

static void cmv_decode_inter(CmvContext * s, const uint8_t *buf, const uint8_t *buf_end){
    const uint8_t *raw = buf + (s->avctx->width*s->avctx->height/16);
    int x,y,i;

    i = 0;
    for(y=0; y<s->avctx->height/4; y++)
    for(x=0; x<s->avctx->width/4 && buf+i<buf_end; x++) {
        if (buf[i]==0xFF) {
            unsigned char *dst = s->frame.data[0] + (y*4)*s->frame.linesize[0] + x*4;
            if (raw+16<buf_end && *raw==0xFF) { /* intra */
                raw++;
                memcpy(dst, raw, 4);
                memcpy(dst+s->frame.linesize[0], raw+4, 4);
                memcpy(dst+2*s->frame.linesize[0], raw+8, 4);
                memcpy(dst+3*s->frame.linesize[0], raw+12, 4);
                raw+=16;
            }else if(raw<buf_end) {  /* inter using second-last frame as reference */
                int xoffset = (*raw & 0xF) - 7;
                int yoffset = ((*raw >> 4)) - 7;
                cmv_motcomp(s->frame.data[0], s->frame.linesize[0],
                            s->last2_frame.data[0], s->last2_frame.linesize[0],
                            x*4, y*4, xoffset, yoffset, s->avctx->width, s->avctx->height);
                raw++;
            }
        }else{  /* inter using last frame as reference */
            int xoffset = (buf[i] & 0xF) - 7;
            int yoffset = ((buf[i] >> 4)) - 7;
            cmv_motcomp(s->frame.data[0], s->frame.linesize[0],
                      s->last_frame.data[0], s->last_frame.linesize[0],
                      x*4, y*4, xoffset, yoffset, s->avctx->width, s->avctx->height);
        }
        i++;
    }
}

static void cmv_process_header(CmvContext *s, const uint8_t *buf, const uint8_t *buf_end)
{
    int pal_start, pal_count, i;

    if(buf+16>=buf_end) {
        av_log(s->avctx, AV_LOG_WARNING, "truncated header\n");
        return;
    }

    s->width  = AV_RL16(&buf[4]);
    s->height = AV_RL16(&buf[6]);
    if (s->avctx->width!=s->width || s->avctx->height!=s->height)
        avcodec_set_dimensions(s->avctx, s->width, s->height);

    s->avctx->time_base.num = 1;
    s->avctx->time_base.den = AV_RL16(&buf[10]);

    pal_start = AV_RL16(&buf[12]);
    pal_count = AV_RL16(&buf[14]);

    buf += 16;
    for (i=pal_start; i<pal_start+pal_count && i<AVPALETTE_COUNT && buf+2<buf_end; i++) {
        s->palette[i] = AV_RB24(buf);
        buf += 3;
    }
}

#define EA_PREAMBLE_SIZE 8
#define MVIh_TAG MKTAG('M', 'V', 'I', 'h')

static int cmv_decode_frame(AVCodecContext *avctx,
                            void *data, int *data_size,
                            AVPacket *avpkt)
{
    const uint8_t *buf = avpkt->data;
    int buf_size = avpkt->size;
    CmvContext *s = avctx->priv_data;
    const uint8_t *buf_end = buf + buf_size;

    if (AV_RL32(buf)==MVIh_TAG||AV_RB32(buf)==MVIh_TAG) {
        cmv_process_header(s, buf+EA_PREAMBLE_SIZE, buf_end);
        return buf_size;
    }

    if (avcodec_check_dimensions(s->avctx, s->width, s->height))
        return -1;

    /* shuffle */
    if (s->last2_frame.data[0])
        avctx->release_buffer(avctx, &s->last2_frame);
    FFSWAP(AVFrame, s->last_frame, s->last2_frame);
    FFSWAP(AVFrame, s->frame, s->last_frame);

    s->frame.reference = 1;
    s->frame.buffer_hints = FF_BUFFER_HINTS_VALID;
    if (avctx->get_buffer(avctx, &s->frame)<0) {
        av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
        return -1;
    }

    memcpy(s->frame.data[1], s->palette, AVPALETTE_SIZE);

    buf += EA_PREAMBLE_SIZE;
    if ((buf[0]&1)) {  // subtype
        cmv_decode_inter(s, buf+2, buf_end);
        s->frame.key_frame = 0;
        s->frame.pict_type = FF_P_TYPE;
    }else{
        s->frame.key_frame = 1;
        s->frame.pict_type = FF_I_TYPE;
        cmv_decode_intra(s, buf+2, buf_end);
    }

    *data_size = sizeof(AVFrame);
    *(AVFrame*)data = s->frame;

    return buf_size;
}

static av_cold int cmv_decode_end(AVCodecContext *avctx){
    CmvContext *s = avctx->priv_data;
    if (s->frame.data[0])
        s->avctx->release_buffer(avctx, &s->frame);
    if (s->last_frame.data[0])
        s->avctx->release_buffer(avctx, &s->last_frame);
    if (s->last2_frame.data[0])
        s->avctx->release_buffer(avctx, &s->last2_frame);

    return 0;
}

AVCodec eacmv_decoder = {
    "eacmv",
    AVMEDIA_TYPE_VIDEO,
    CODEC_ID_CMV,
    sizeof(CmvContext),
    cmv_decode_init,
    NULL,
    cmv_decode_end,
    cmv_decode_frame,
    CODEC_CAP_DR1,
    .long_name = NULL_IF_CONFIG_SMALL("Electronic Arts CMV video"),
};