ffmpeg/libavfilter/vf_hqx.c

/*
 * Copyright (c) 2014 Clément Bœsch
 *
 * This file is part of FFmpeg.
 *
 * Permission to use, copy, modify, and/or distribute this software for any
 * purpose with or without fee is hereby granted, provided that the above
 * copyright notice and this permission notice appear in all copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */

/**
 * @file
 * hqx magnification filters (hq2x, hq3x, hq4x)
 *
 * Originally designed by Maxim Stephin.
 *
 * @see http://en.wikipedia.org/wiki/Hqx
 * @see http://web.archive.org/web/20131114143602/http://www.hiend3d.com/hq3x.html
 * @see http://blog.pkh.me/p/19-butchering-hqx-scaling-filters.html
 */

#include "libavutil/opt.h"
#include "libavutil/avassert.h"
#include "libavutil/pixdesc.h"
#include "internal.h"

typedef int (*hqxfunc_t)(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs);

typedef struct HQXContext {
    const AVClass *class;
    int n;
    hqxfunc_t func;
    uint32_t rgbtoyuv[1<<24];
} HQXContext;

typedef struct ThreadData {
    AVFrame *in, *out;
    const uint32_t *rgbtoyuv;
} ThreadData;

#define OFFSET(x) offsetof(HQXContext, x)
#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
static const AVOption hqx_options[] = {
    { "n", "set scale factor", OFFSET(n), AV_OPT_TYPE_INT, {.i64 = 3}, 2, 4, .flags = FLAGS },
    { NULL }
};

AVFILTER_DEFINE_CLASS(hqx);

static av_always_inline uint32_t rgb2yuv(const uint32_t *r2y, uint32_t c)
{
    return r2y[c & 0xffffff];
}

static av_always_inline int yuv_diff(uint32_t yuv1, uint32_t yuv2)
{
#define YMASK 0xff0000
#define UMASK 0x00ff00
#define VMASK 0x0000ff
#define ABSDIFF(a,b) (abs((int)(a)-(int)(b)))

    return ABSDIFF(yuv1 & YMASK, yuv2 & YMASK) > (48 << 16) ||
           ABSDIFF(yuv1 & UMASK, yuv2 & UMASK) > ( 7 <<  8) ||
           ABSDIFF(yuv1 & VMASK, yuv2 & VMASK) > ( 6 <<  0);
}

/* (c1*w1 + c2*w2) >> s */
static av_always_inline uint32_t interp_2px(uint32_t c1, int w1, uint32_t c2, int w2, int s)
{
    return (((((c1 & 0xff00ff00) >> 8) * w1 + ((c2 & 0xff00ff00) >> 8) * w2) << (8 - s)) & 0xff00ff00) |
           (((((c1 & 0x00ff00ff)     ) * w1 + ((c2 & 0x00ff00ff)     ) * w2) >>      s ) & 0x00ff00ff);
}

/* (c1*w1 + c2*w2 + c3*w3) >> s */
static av_always_inline uint32_t interp_3px(uint32_t c1, int w1, uint32_t c2, int w2, uint32_t c3, int w3, int s)
{
    return (((((c1 & 0xff00ff00) >> 8) * w1 + ((c2 & 0xff00ff00) >> 8) * w2 + ((c3 & 0xff00ff00) >> 8) * w3) << (8 - s)) & 0xff00ff00) |
           (((((c1 & 0x00ff00ff)     ) * w1 + ((c2 & 0x00ff00ff)     ) * w2 + ((c3 & 0x00ff00ff)     ) * w3) >>      s ) & 0x00ff00ff);
}

/* m is the mask of diff with the center pixel that matters in the pattern, and
 * r is the expected result (bit set to 1 if there is difference with the
 * center, 0 otherwise) */
#define P(m, r) ((k_shuffled & (m)) == (r))

/* adjust 012345678 to 01235678: the mask doesn't contain the (null) diff
 * between the center/current pixel and itself */
#define DROP4(z) ((z) > 4 ? (z)-1 : (z))

/* shuffle the input mask: move bit n (4-adjusted) to position stored in p<n> */
#define SHF(x, rot, n) (((x) >> ((rot) ? 7-DROP4(n) : DROP4(n)) & 1) << DROP4(p##n))

/* used to check if there is YUV difference between 2 pixels */
#define WDIFF(c1, c2) yuv_diff(rgb2yuv(r2y, c1), rgb2yuv(r2y, c2))

/* bootstrap template for every interpolation code. It defines the shuffled
 * masks and surrounding pixels. The rot flag is used to indicate if it's a
 * rotation; its basic effect is to shuffle k using p8..p0 instead of p0..p8 */
#define INTERP_BOOTSTRAP(rot)                                           \
    const int k_shuffled = SHF(k,rot,0) | SHF(k,rot,1) | SHF(k,rot,2)   \
                         | SHF(k,rot,3) |       0      | SHF(k,rot,5)   \
                         | SHF(k,rot,6) | SHF(k,rot,7) | SHF(k,rot,8);  \
                                                                        \
    const uint32_t w0 = w[p0], w1 = w[p1],                              \
                   w3 = w[p3], w4 = w[p4], w5 = w[p5],                  \
                               w7 = w[p7]

/* Assuming p0..p8 is mapped to pixels 0..8, this function interpolates the
 * top-left pixel in the total of the 2x2 pixels to interpolates. The function
 * is also used for the 3 other pixels */
static av_always_inline uint32_t hq2x_interp_1x1(const uint32_t *r2y, int k,
                                                 const uint32_t *w,
                                                 int p0, int p1, int p2,
                                                 int p3, int p4, int p5,
                                                 int p6, int p7, int p8)
{
    INTERP_BOOTSTRAP(0);

    if ((P(0xbf,0x37) || P(0xdb,0x13)) && WDIFF(w1, w5))
        return interp_2px(w4, 3, w3, 1, 2);
    if ((P(0xdb,0x49) || P(0xef,0x6d)) && WDIFF(w7, w3))
        return interp_2px(w4, 3, w1, 1, 2);
    if ((P(0x0b,0x0b) || P(0xfe,0x4a) || P(0xfe,0x1a)) && WDIFF(w3, w1))
        return w4;
    if ((P(0x6f,0x2a) || P(0x5b,0x0a) || P(0xbf,0x3a) || P(0xdf,0x5a) ||
         P(0x9f,0x8a) || P(0xcf,0x8a) || P(0xef,0x4e) || P(0x3f,0x0e) ||
         P(0xfb,0x5a) || P(0xbb,0x8a) || P(0x7f,0x5a) || P(0xaf,0x8a) ||
         P(0xeb,0x8a)) && WDIFF(w3, w1))
        return interp_2px(w4, 3, w0, 1, 2);
    if (P(0x0b,0x08))
        return interp_3px(w4, 2, w0, 1, w1, 1, 2);
    if (P(0x0b,0x02))
        return interp_3px(w4, 2, w0, 1, w3, 1, 2);
    if (P(0x2f,0x2f))
        return interp_3px(w4, 14, w3, 1, w1, 1, 4);
    if (P(0xbf,0x37) || P(0xdb,0x13))
        return interp_3px(w4, 5, w1, 2, w3, 1, 3);
    if (P(0xdb,0x49) || P(0xef,0x6d))
        return interp_3px(w4, 5, w3, 2, w1, 1, 3);
    if (P(0x1b,0x03) || P(0x4f,0x43) || P(0x8b,0x83) || P(0x6b,0x43))
        return interp_2px(w4, 3, w3, 1, 2);
    if (P(0x4b,0x09) || P(0x8b,0x89) || P(0x1f,0x19) || P(0x3b,0x19))
        return interp_2px(w4, 3, w1, 1, 2);
    if (P(0x7e,0x2a) || P(0xef,0xab) || P(0xbf,0x8f) || P(0x7e,0x0e))
        return interp_3px(w4, 2, w3, 3, w1, 3, 3);
    if (P(0xfb,0x6a) || P(0x6f,0x6e) || P(0x3f,0x3e) || P(0xfb,0xfa) ||
        P(0xdf,0xde) || P(0xdf,0x1e))
        return interp_2px(w4, 3, w0, 1, 2);
    if (P(0x0a,0x00) || P(0x4f,0x4b) || P(0x9f,0x1b) || P(0x2f,0x0b) ||
        P(0xbe,0x0a) || P(0xee,0x0a) || P(0x7e,0x0a) || P(0xeb,0x4b) ||
        P(0x3b,0x1b))
        return interp_3px(w4, 2, w3, 1, w1, 1, 2);
    return interp_3px(w4, 6, w3, 1, w1, 1, 3);
}

/* Assuming p0..p8 is mapped to pixels 0..8, this function interpolates the
 * top-left and top-center pixel in the total of the 3x3 pixels to
 * interpolates. The function is also used for the 3 other couples of pixels
 * defining the outline. The center pixel is not defined through this function,
 * since it's just the same as the original value. */
static av_always_inline void hq3x_interp_2x1(uint32_t *dst, int dst_linesize,
                                             const uint32_t *r2y, int k,
                                             const uint32_t *w,
                                             int pos00, int pos01,
                                             int p0, int p1, int p2,
                                             int p3, int p4, int p5,
                                             int p6, int p7, int p8,
                                             int rotate)
{
    INTERP_BOOTSTRAP(rotate);

    uint32_t *dst00 = &dst[dst_linesize*(pos00>>1) + (pos00&1)];
    uint32_t *dst01 = &dst[dst_linesize*(pos01>>1) + (pos01&1)];

    if ((P(0xdb,0x49) || P(0xef,0x6d)) && WDIFF(w7, w3))
        *dst00 = interp_2px(w4, 3, w1, 1, 2);
    else if ((P(0xbf,0x37) || P(0xdb,0x13)) && WDIFF(w1, w5))
        *dst00 = interp_2px(w4, 3, w3, 1, 2);
    else if ((P(0x0b,0x0b) || P(0xfe,0x4a) || P(0xfe,0x1a)) && WDIFF(w3, w1))
        *dst00 = w4;
    else if ((P(0x6f,0x2a) || P(0x5b,0x0a) || P(0xbf,0x3a) || P(0xdf,0x5a) ||
              P(0x9f,0x8a) || P(0xcf,0x8a) || P(0xef,0x4e) || P(0x3f,0x0e) ||
              P(0xfb,0x5a) || P(0xbb,0x8a) || P(0x7f,0x5a) || P(0xaf,0x8a) ||
              P(0xeb,0x8a)) && WDIFF(w3, w1))
        *dst00 = interp_2px(w4, 3, w0, 1, 2);
    else if (P(0x4b,0x09) || P(0x8b,0x89) || P(0x1f,0x19) || P(0x3b,0x19))
        *dst00 = interp_2px(w4, 3, w1, 1, 2);
    else if (P(0x1b,0x03) || P(0x4f,0x43) || P(0x8b,0x83) || P(0x6b,0x43))
        *dst00 = interp_2px(w4, 3, w3, 1, 2);
    else if (P(0x7e,0x2a) || P(0xef,0xab) || P(0xbf,0x8f) || P(0x7e,0x0e))
        *dst00 = interp_2px(w3, 1, w1, 1, 1);
    else if (P(0x4f,0x4b) || P(0x9f,0x1b) || P(0x2f,0x0b) || P(0xbe,0x0a) ||
             P(0xee,0x0a) || P(0x7e,0x0a) || P(0xeb,0x4b) || P(0x3b,0x1b))
        *dst00 = interp_3px(w4, 2, w3, 7, w1, 7, 4);
    else if (P(0x0b,0x08) || P(0xf9,0x68) || P(0xf3,0x62) || P(0x6d,0x6c) ||
             P(0x67,0x66) || P(0x3d,0x3c) || P(0x37,0x36) || P(0xf9,0xf8) ||
             P(0xdd,0xdc) || P(0xf3,0xf2) || P(0xd7,0xd6) || P(0xdd,0x1c) ||
             P(0xd7,0x16) || P(0x0b,0x02))
        *dst00 = interp_2px(w4, 3, w0, 1, 2);
    else
        *dst00 = interp_3px(w4, 2, w3, 1, w1, 1, 2);

    if ((P(0xfe,0xde) || P(0x9e,0x16) || P(0xda,0x12) || P(0x17,0x16) ||
         P(0x5b,0x12) || P(0xbb,0x12)) && WDIFF(w1, w5))
        *dst01 = w4;
    else if ((P(0x0f,0x0b) || P(0x5e,0x0a) || P(0xfb,0x7b) || P(0x3b,0x0b) ||
              P(0xbe,0x0a) || P(0x7a,0x0a)) && WDIFF(w3, w1))
        *dst01 = w4;
    else if (P(0xbf,0x8f) || P(0x7e,0x0e) || P(0xbf,0x37) || P(0xdb,0x13))
        *dst01 = interp_2px(w1, 3, w4, 1, 2);
    else if (P(0x02,0x00) || P(0x7c,0x28) || P(0xed,0xa9) || P(0xf5,0xb4) ||
             P(0xd9,0x90))
        *dst01 = interp_2px(w4, 3, w1, 1, 2);
    else if (P(0x4f,0x4b) || P(0xfb,0x7b) || P(0xfe,0x7e) || P(0x9f,0x1b) ||
             P(0x2f,0x0b) || P(0xbe,0x0a) || P(0x7e,0x0a) || P(0xfb,0x4b) ||
             P(0xfb,0xdb) || P(0xfe,0xde) || P(0xfe,0x56) || P(0x57,0x56) ||
             P(0x97,0x16) || P(0x3f,0x1e) || P(0xdb,0x12) || P(0xbb,0x12))
        *dst01 = interp_2px(w4, 7, w1, 1, 3);
    else
        *dst01 = w4;
}

/* Assuming p0..p8 is mapped to pixels 0..8, this function interpolates the
 * top-left block of 2x2 pixels in the total of the 4x4 pixels (or 4 blocks) to
 * interpolates. The function is also used for the 3 other blocks of 2x2
 * pixels. */
static av_always_inline void hq4x_interp_2x2(uint32_t *dst, int dst_linesize,
                                             const uint32_t *r2y, int k,
                                             const uint32_t *w,
                                             int pos00, int pos01,
                                             int pos10, int pos11,
                                             int p0, int p1, int p2,
                                             int p3, int p4, int p5,
                                             int p6, int p7, int p8)
{
    INTERP_BOOTSTRAP(0);

    uint32_t *dst00 = &dst[dst_linesize*(pos00>>1) + (pos00&1)];
    uint32_t *dst01 = &dst[dst_linesize*(pos01>>1) + (pos01&1)];
    uint32_t *dst10 = &dst[dst_linesize*(pos10>>1) + (pos10&1)];
    uint32_t *dst11 = &dst[dst_linesize*(pos11>>1) + (pos11&1)];

    const int cond00 = (P(0xbf,0x37) || P(0xdb,0x13)) && WDIFF(w1, w5);
    const int cond01 = (P(0xdb,0x49) || P(0xef,0x6d)) && WDIFF(w7, w3);
    const int cond02 = (P(0x6f,0x2a) || P(0x5b,0x0a) || P(0xbf,0x3a) ||
                        P(0xdf,0x5a) || P(0x9f,0x8a) || P(0xcf,0x8a) ||
                        P(0xef,0x4e) || P(0x3f,0x0e) || P(0xfb,0x5a) ||
                        P(0xbb,0x8a) || P(0x7f,0x5a) || P(0xaf,0x8a) ||
                        P(0xeb,0x8a)) && WDIFF(w3, w1);
    const int cond03 = P(0xdb,0x49) || P(0xef,0x6d);
    const int cond04 = P(0xbf,0x37) || P(0xdb,0x13);
    const int cond05 = P(0x1b,0x03) || P(0x4f,0x43) || P(0x8b,0x83) ||
                       P(0x6b,0x43);
    const int cond06 = P(0x4b,0x09) || P(0x8b,0x89) || P(0x1f,0x19) ||
                       P(0x3b,0x19);
    const int cond07 = P(0x0b,0x08) || P(0xf9,0x68) || P(0xf3,0x62) ||
                       P(0x6d,0x6c) || P(0x67,0x66) || P(0x3d,0x3c) ||
                       P(0x37,0x36) || P(0xf9,0xf8) || P(0xdd,0xdc) ||
                       P(0xf3,0xf2) || P(0xd7,0xd6) || P(0xdd,0x1c) ||
                       P(0xd7,0x16) || P(0x0b,0x02);
    const int cond08 = (P(0x0f,0x0b) || P(0x2b,0x0b) || P(0xfe,0x4a) ||
                        P(0xfe,0x1a)) && WDIFF(w3, w1);
    const int cond09 = P(0x2f,0x2f);
    const int cond10 = P(0x0a,0x00);
    const int cond11 = P(0x0b,0x09);
    const int cond12 = P(0x7e,0x2a) || P(0xef,0xab);
    const int cond13 = P(0xbf,0x8f) || P(0x7e,0x0e);
    const int cond14 = P(0x4f,0x4b) || P(0x9f,0x1b) || P(0x2f,0x0b) ||
                       P(0xbe,0x0a) || P(0xee,0x0a) || P(0x7e,0x0a) ||
                       P(0xeb,0x4b) || P(0x3b,0x1b);
    const int cond15 = P(0x0b,0x03);

    if (cond00)
        *dst00 = interp_2px(w4, 5, w3, 3, 3);
    else if (cond01)
        *dst00 = interp_2px(w4, 5, w1, 3, 3);
    else if ((P(0x0b,0x0b) || P(0xfe,0x4a) || P(0xfe,0x1a)) && WDIFF(w3, w1))
        *dst00 = w4;
    else if (cond02)
        *dst00 = interp_2px(w4, 5, w0, 3, 3);
    else if (cond03)
        *dst00 = interp_2px(w4, 3, w3, 1, 2);
    else if (cond04)
        *dst00 = interp_2px(w4, 3, w1, 1, 2);
    else if (cond05)
        *dst00 = interp_2px(w4, 5, w3, 3, 3);
    else if (cond06)
        *dst00 = interp_2px(w4, 5, w1, 3, 3);
    else if (P(0x0f,0x0b) || P(0x5e,0x0a) || P(0x2b,0x0b) || P(0xbe,0x0a) ||
             P(0x7a,0x0a) || P(0xee,0x0a))
        *dst00 = interp_2px(w1, 1, w3, 1, 1);
    else if (cond07)
        *dst00 = interp_2px(w4, 5, w0, 3, 3);
    else
        *dst00 = interp_3px(w4, 2, w1, 1, w3, 1, 2);

    if (cond00)
        *dst01 = interp_2px(w4, 7, w3, 1, 3);
    else if (cond08)
        *dst01 = w4;
    else if (cond02)
        *dst01 = interp_2px(w4, 3, w0, 1, 2);
    else if (cond09)
        *dst01 = w4;
    else if (cond10)
        *dst01 = interp_3px(w4, 5, w1, 2, w3, 1, 3);
    else if (P(0x0b,0x08))
        *dst01 = interp_3px(w4, 5, w1, 2, w0, 1, 3);
    else if (cond11)
        *dst01 = interp_2px(w4, 5, w1, 3, 3);
    else if (cond04)
        *dst01 = interp_2px(w1, 3, w4, 1, 2);
    else if (cond12)
        *dst01 = interp_3px(w1, 2, w4, 1, w3, 1, 2);
    else if (cond13)
        *dst01 = interp_2px(w1, 5, w3, 3, 3);
    else if (cond05)
        *dst01 = interp_2px(w4, 7, w3, 1, 3);
    else if (P(0xf3,0x62) || P(0x67,0x66) || P(0x37,0x36) || P(0xf3,0xf2) ||
             P(0xd7,0xd6) || P(0xd7,0x16) || P(0x0b,0x02))
        *dst01 = interp_2px(w4, 3, w0, 1, 2);
    else if (cond14)
        *dst01 = interp_2px(w1, 1, w4, 1, 1);
    else
        *dst01 = interp_2px(w4, 3, w1, 1, 2);

    if (cond01)
        *dst10 = interp_2px(w4, 7, w1, 1, 3);
    else if (cond08)
        *dst10 = w4;
    else if (cond02)
        *dst10 = interp_2px(w4, 3, w0, 1, 2);
    else if (cond09)
        *dst10 = w4;
    else if (cond10)
        *dst10 = interp_3px(w4, 5, w3, 2, w1, 1, 3);
    else if (P(0x0b,0x02))
        *dst10 = interp_3px(w4, 5, w3, 2, w0, 1, 3);
    else if (cond15)
        *dst10 = interp_2px(w4, 5, w3, 3, 3);
    else if (cond03)
        *dst10 = interp_2px(w3, 3, w4, 1, 2);
    else if (cond13)
        *dst10 = interp_3px(w3, 2, w4, 1, w1, 1, 2);
    else if (cond12)
        *dst10 = interp_2px(w3, 5, w1, 3, 3);
    else if (cond06)
        *dst10 = interp_2px(w4, 7, w1, 1, 3);
    else if (P(0x0b,0x08) || P(0xf9,0x68) || P(0x6d,0x6c) || P(0x3d,0x3c) ||
             P(0xf9,0xf8) || P(0xdd,0xdc) || P(0xdd,0x1c))
        *dst10 = interp_2px(w4, 3, w0, 1, 2);
    else if (cond14)
        *dst10 = interp_2px(w3, 1, w4, 1, 1);
    else
        *dst10 = interp_2px(w4, 3, w3, 1, 2);

    if ((P(0x7f,0x2b) || P(0xef,0xab) || P(0xbf,0x8f) || P(0x7f,0x0f)) &&
         WDIFF(w3, w1))
        *dst11 = w4;
    else if (cond02)
        *dst11 = interp_2px(w4, 7, w0, 1, 3);
    else if (cond15)
        *dst11 = interp_2px(w4, 7, w3, 1, 3);
    else if (cond11)
        *dst11 = interp_2px(w4, 7, w1, 1, 3);
    else if (P(0x0a,0x00) || P(0x7e,0x2a) || P(0xef,0xab) || P(0xbf,0x8f) ||
             P(0x7e,0x0e))
        *dst11 = interp_3px(w4, 6, w3, 1, w1, 1, 3);
    else if (cond07)
        *dst11 = interp_2px(w4, 7, w0, 1, 3);
    else
        *dst11 = w4;
}

static av_always_inline void hqx_filter(const ThreadData *td, int jobnr, int nb_jobs, int n)
{
    int x, y;
    AVFrame *in = td->in, *out = td->out;
    const uint32_t *r2y = td->rgbtoyuv;
    const int height = in->height;
    const int width  = in->width;
    const int slice_start = (height *  jobnr   ) / nb_jobs;
    const int slice_end   = (height * (jobnr+1)) / nb_jobs;
    const int dst_linesize = out->linesize[0];
    const int src_linesize =  in->linesize[0];
    uint8_t       *dst = out->data[0] + slice_start * dst_linesize * n;
    const uint8_t *src =  in->data[0] + slice_start * src_linesize;

    const int dst32_linesize = dst_linesize >> 2;
    const int src32_linesize = src_linesize >> 2;

    for (y = slice_start; y < slice_end; y++) {
        const uint32_t *src32 = (const uint32_t *)src;
        uint32_t       *dst32 = (uint32_t *)dst;
        const int prevline = y > 0          ? -src32_linesize : 0;
        const int nextline = y < height - 1 ?  src32_linesize : 0;

        for (x = 0; x < width; x++) {
            const int prevcol = x > 0        ? -1 : 0;
            const int nextcol = x < width -1 ?  1 : 0;
            const uint32_t w[3*3] = {
                src32[prevcol + prevline], src32[prevline], src32[prevline + nextcol],
                src32[prevcol           ], src32[       0], src32[           nextcol],
                src32[prevcol + nextline], src32[nextline], src32[nextline + nextcol]
            };
            const uint32_t yuv1 = rgb2yuv(r2y, w[4]);
            const int pattern = (w[4] != w[0] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[0]))) : 0)
                              | (w[4] != w[1] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[1]))) : 0) << 1
                              | (w[4] != w[2] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[2]))) : 0) << 2
                              | (w[4] != w[3] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[3]))) : 0) << 3
                              | (w[4] != w[5] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[5]))) : 0) << 4
                              | (w[4] != w[6] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[6]))) : 0) << 5
                              | (w[4] != w[7] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[7]))) : 0) << 6
                              | (w[4] != w[8] ? (yuv_diff(yuv1, rgb2yuv(r2y, w[8]))) : 0) << 7;

            if (n == 2) {
                dst32[dst32_linesize*0 + 0] = hq2x_interp_1x1(r2y, pattern, w, 0,1,2,3,4,5,6,7,8);  // 00
                dst32[dst32_linesize*0 + 1] = hq2x_interp_1x1(r2y, pattern, w, 2,1,0,5,4,3,8,7,6);  // 01 (vert mirrored)
                dst32[dst32_linesize*1 + 0] = hq2x_interp_1x1(r2y, pattern, w, 6,7,8,3,4,5,0,1,2);  // 10 (horiz mirrored)
                dst32[dst32_linesize*1 + 1] = hq2x_interp_1x1(r2y, pattern, w, 8,7,6,5,4,3,2,1,0);  // 11 (center mirrored)
            } else if (n == 3) {
                hq3x_interp_2x1(dst32,                        dst32_linesize, r2y, pattern, w, 0,1, 0,1,2,3,4,5,6,7,8, 0);  // 00 01
                hq3x_interp_2x1(dst32 + 1,                    dst32_linesize, r2y, pattern, w, 1,3, 2,5,8,1,4,7,0,3,6, 1);  // 02 12 (rotated to the right)
                hq3x_interp_2x1(dst32 + 1*dst32_linesize,     dst32_linesize, r2y, pattern, w, 2,0, 6,3,0,7,4,1,8,5,2, 1);  // 20 10 (rotated to the left)
                hq3x_interp_2x1(dst32 + 1*dst32_linesize + 1, dst32_linesize, r2y, pattern, w, 3,2, 8,7,6,5,4,3,2,1,0, 0);  // 22 21 (center mirrored)
                dst32[dst32_linesize + 1] = w[4];                                                                           // 11
            } else if (n == 4) {
                hq4x_interp_2x2(dst32,                        dst32_linesize, r2y, pattern, w, 0,1,2,3, 0,1,2,3,4,5,6,7,8); // 00 01 10 11
                hq4x_interp_2x2(dst32 + 2,                    dst32_linesize, r2y, pattern, w, 1,0,3,2, 2,1,0,5,4,3,8,7,6); // 02 03 12 13 (vert mirrored)
                hq4x_interp_2x2(dst32 + 2*dst32_linesize,     dst32_linesize, r2y, pattern, w, 2,3,0,1, 6,7,8,3,4,5,0,1,2); // 20 21 30 31 (horiz mirrored)
                hq4x_interp_2x2(dst32 + 2*dst32_linesize + 2, dst32_linesize, r2y, pattern, w, 3,2,1,0, 8,7,6,5,4,3,2,1,0); // 22 23 32 33 (center mirrored)
            } else {
                av_assert0(0);
            }

            src32 += 1;
            dst32 += n;
        }

        src += src_linesize;
        dst += dst_linesize * n;
    }
}

#define HQX_FUNC(size) \
static int hq##size##x(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs) \
{ \
    hqx_filter(arg, jobnr, nb_jobs, size); \
    return 0; \
}

HQX_FUNC(2)
HQX_FUNC(3)
HQX_FUNC(4)

static int config_output(AVFilterLink *outlink)
{
    AVFilterContext *ctx = outlink->src;
    HQXContext *hqx = ctx->priv;
    AVFilterLink *inlink = ctx->inputs[0];

    outlink->w = inlink->w * hqx->n;
    outlink->h = inlink->h * hqx->n;
    av_log(inlink->dst, AV_LOG_VERBOSE, "fmt:%s size:%dx%d -> size:%dx%d\n",
           av_get_pix_fmt_name(inlink->format),
           inlink->w, inlink->h, outlink->w, outlink->h);
    return 0;
}

static int filter_frame(AVFilterLink *inlink, AVFrame *in)
{
    AVFilterContext *ctx = inlink->dst;
    AVFilterLink *outlink = ctx->outputs[0];
    HQXContext *hqx = ctx->priv;
    ThreadData td;
    AVFrame *out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
    if (!out) {
        av_frame_free(&in);
        return AVERROR(ENOMEM);
    }
    av_frame_copy_props(out, in);
    out->width  = outlink->w;
    out->height = outlink->h;

    td.in = in;
    td.out = out;
    td.rgbtoyuv = hqx->rgbtoyuv;
    ff_filter_execute(ctx, hqx->func, &td, NULL,
                      FFMIN(inlink->h, ff_filter_get_nb_threads(ctx)));

    av_frame_free(&in);
    return ff_filter_frame(outlink, out);
}

static av_cold int init(AVFilterContext *ctx)
{
    HQXContext *hqx = ctx->priv;
    static const hqxfunc_t hqxfuncs[] = {hq2x, hq3x, hq4x};

    uint32_t c;
    int bg, rg, g;

    for (bg=-255; bg<256; bg++) {
        for (rg=-255; rg<256; rg++) {
            const uint32_t u = (uint32_t)((-169*rg + 500*bg)/1000) + 128;
            const uint32_t v = (uint32_t)(( 500*rg -  81*bg)/1000) + 128;
            int startg = FFMAX3(-bg, -rg, 0);
            int endg = FFMIN3(255-bg, 255-rg, 255);
            uint32_t y = (uint32_t)(( 299*rg + 1000*startg + 114*bg)/1000);
            c = bg + rg * (1 << 16) + 0x010101 * startg;
            for (g = startg; g <= endg; g++) {
                hqx->rgbtoyuv[c] = ((y++) << 16) + (u << 8) + v;
                c+= 0x010101;
            }
        }
    }

    hqx->func = hqxfuncs[hqx->n - 2];
    return 0;
}

static const AVFilterPad hqx_inputs[] = {
    {
        .name         = "default",
        .type         = AVMEDIA_TYPE_VIDEO,
        .filter_frame = filter_frame,
    },
};

static const AVFilterPad hqx_outputs[] = {
    {
        .name         = "default",
        .type         = AVMEDIA_TYPE_VIDEO,
        .config_props = config_output,
    },
};

const AVFilter ff_vf_hqx = {
    .name          = "hqx",
    .description   = NULL_IF_CONFIG_SMALL("Scale the input by 2, 3 or 4 using the hq*x magnification algorithm."),
    .priv_size     = sizeof(HQXContext),
    .init          = init,
    FILTER_INPUTS(hqx_inputs),
    FILTER_OUTPUTS(hqx_outputs),
    FILTER_SINGLE_PIXFMT(AV_PIX_FMT_RGB32),
    .priv_class    = &hqx_class,
    .flags         = AVFILTER_FLAG_SLICE_THREADS,
};