annotate dctref.c @ 11032:01bd040f8607 libavcodec

Unroll main loop so the edge==0 case is seperate. This allows many things to be simplified away. h264 decoder is overall 1% faster with a mbaff sample and 0.1% slower with the cathedral sample, probably because the slow loop filter code must be loaded into the code cache for each first MB of each row but isnt used for the following MBs.

/*
 * reference discrete cosine transform (double precision)
 * Copyright (C) 2009 Dylan Yudaken
 *
 * 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/dctref.c
 * reference discrete cosine transform (double precision)
 *
 * @author Dylan Yudaken (dyudaken at gmail)
 *
 * @note This file could be optimized a lot, but is for
 * reference and so readability is better.
 */

#include "libavutil/mathematics.h"
static double coefficients[8 * 8];

/**
 * Initialize the double precision discrete cosine transform
 * functions fdct & idct.
 */
av_cold void ff_ref_dct_init(void)
{
    unsigned int i, j;

    for (j = 0; j < 8; ++j) {
        coefficients[j] = sqrt(0.125);
        for (i = 8; i < 64; i += 8) {
            coefficients[i + j] = 0.5 * cos(i * (j + 0.5) * M_PI / 64.0);
        }
    }
}

/**
 * Transform 8x8 block of data with a double precision forward DCT <br>
 * This is a reference implementation.
 *
 * @param block pointer to 8x8 block of data to transform
 */
void ff_ref_fdct(short *block)
{
    /* implement the equation: block = coefficients * block * coefficients' */

    unsigned int i, j, k;
    double out[8 * 8];

    /* out = coefficients * block */
    for (i = 0; i < 64; i += 8) {
        for (j = 0; j < 8; ++j) {
            double tmp = 0;
            for (k = 0; k < 8; ++k) {
                tmp += coefficients[i + k] * block[k * 8 + j];
            }
            out[i + j] = tmp * 8;
        }
    }

    /* block = out * (coefficients') */
    for (j = 0; j < 8; ++j) {
        for (i = 0; i < 64; i += 8) {
            double tmp = 0;
            for (k = 0; k < 8; ++k) {
                tmp += out[i + k] * coefficients[j * 8 + k];
            }
            block[i + j] = floor(tmp + 0.499999999999);
        }
    }
}

/**
 * Transform 8x8 block of data with a double precision inverse DCT <br>
 * This is a reference implementation.
 *
 * @param block pointer to 8x8 block of data to transform
 */
void ff_ref_idct(short *block)
{
    /* implement the equation: block = (coefficients') * block * coefficients */

    unsigned int i, j, k;
    double out[8 * 8];

    /* out = block * coefficients */
    for (i = 0; i < 64; i += 8) {
        for (j = 0; j < 8; ++j) {
            double tmp = 0;
            for (k = 0; k < 8; ++k) {
                tmp += block[i + k] * coefficients[k * 8 + j];
            }
            out[i + j] = tmp;
        }
    }

    /* block = (coefficients') * out */
    for (i = 0; i < 8; ++i) {
        for (j = 0; j < 8; ++j) {
            double tmp = 0;
            for (k = 0; k < 64; k += 8) {
                tmp += coefficients[k + i] * out[k + j];
            }
            block[i * 8 + j] = floor(tmp + 0.5);
        }
    }
}