avfilter: port pullup filter from libmpcodecs

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Changelog

@@ -24,6 +24,7 @@ version <next>
 - changed DTS stream id in lavf mpeg ps muxer from 0x8a to 0x88, to be
   more consistent with other muxers.
 - adelay filter
+- pullup filter ported from libmpcodecs
 
 
 version 2.0:

LICENSE

@@ -41,6 +41,7 @@ Specifically, the GPL parts of FFmpeg are
     - vf_perspective.c
     - vf_phase.c
     - vf_pp.c
+    - vf_pullup.c
     - vf_sab.c
     - vf_smartblur.c
     - vf_spp.c

configure

@@ -2221,6 +2221,7 @@ owdenoise_filter_deps="gpl"
 pan_filter_deps="swresample"
 phase_filter_deps="gpl"
 pp_filter_deps="gpl postproc"
+pullup_filter_deps="gpl"
 removelogo_filter_deps="avcodec avformat swscale"
 sab_filter_deps="gpl swscale"
 scale_filter_deps="swscale"

doc/filters.texi

@@ -6184,6 +6184,59 @@ On this example the input file being processed is compared with the
 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
 is stored in @file{stats.log}.
 
+@section pullup
+
+Pulldown reversal (inverse telecine) filter, capable of handling mixed
+hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
+content.
+
+The pullup filter is designed to take advantage of future context in making
+its decisions. This filter is stateless in the sense that it does not lock
+onto a pattern to follow, but it instead looks forward to the following
+fields in order to identify matches and rebuild progressive frames.
+
+The filter accepts the following options:
+
+@table @option
+@item jl
+@item jr
+@item jt
+@item jb
+These options set the amount of "junk" to ignore at the left, right, top, and
+bottom of the image, respectively. Left and right are in units of 8 pixels,
+while top and bottom are in units of 2 lines.
+The default is 8 pixels on each side.
+
+@item sb
+Set the strict breaks. Setting this option to 1 will reduce the chances of
+filter generating an occasional mismatched frame, but it may also cause an
+excessive number of frames to be dropped during high motion sequences.
+Conversely, setting it to -1 will make filter match fields more easily.
+This may help processing of video where there is slight blurring between
+the fields, but may also cause there to be interlaced frames in the output.
+Default value is @code{0}.
+
+@item mp
+Set the metric plane to use. It accepts the following values:
+@table @samp
+@item l
+Use luma plane.
+
+@item u
+Use chroma blue plane.
+
+@item v
+Use chroma red plane.
+@end table
+
+This option may be set to use chroma plane instead of the default luma plane
+for doing filter's computations. This may improve accuracy on very clean
+source material, but more likely will decrease accuracy, especially if there
+is chroma noise (rainbow effect) or any grayscale video.
+The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
+load and make pullup usable in realtime on slow machines.
+@end table
+
 @section removelogo
 
 Suppress a TV station logo, using an image file to determine which

libavfilter/Makefile

@@ -173,6 +173,7 @@ OBJS-$(CONFIG_PHASE_FILTER)                  += vf_phase.o
 OBJS-$(CONFIG_PIXDESCTEST_FILTER)            += vf_pixdesctest.o
 OBJS-$(CONFIG_PP_FILTER)                     += vf_pp.o
 OBJS-$(CONFIG_PSNR_FILTER)                   += vf_psnr.o dualinput.o
+OBJS-$(CONFIG_PULLUP_FILTER)                 += vf_pullup.o
 OBJS-$(CONFIG_REMOVELOGO_FILTER)             += bbox.o lswsutils.o lavfutils.o vf_removelogo.o
 OBJS-$(CONFIG_ROTATE_FILTER)                 += vf_rotate.o
 OBJS-$(CONFIG_SEPARATEFIELDS_FILTER)         += vf_separatefields.o

libavfilter/allfilters.c

@@ -168,6 +168,7 @@ void avfilter_register_all(void)
     REGISTER_FILTER(PIXDESCTEST,    pixdesctest,    vf);
     REGISTER_FILTER(PP,             pp,             vf);
     REGISTER_FILTER(PSNR,           psnr,           vf);
+    REGISTER_FILTER(PULLUP,         pullup,         vf);
     REGISTER_FILTER(REMOVELOGO,     removelogo,     vf);
     REGISTER_FILTER(ROTATE,         rotate,         vf);
     REGISTER_FILTER(SAB,            sab,            vf);

libavfilter/version.h

@@ -30,7 +30,7 @@
 #include "libavutil/avutil.h"
 
 #define LIBAVFILTER_VERSION_MAJOR  3
-#define LIBAVFILTER_VERSION_MINOR  85
+#define LIBAVFILTER_VERSION_MINOR  86
 #define LIBAVFILTER_VERSION_MICRO 100
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \

libavfilter/vf_pullup.c

+/*
+ * Copyright (c) 2003 Rich Felker
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg 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.
+ *
+ * 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU 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
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/imgutils.h"
+#include "libavutil/opt.h"
+#include "libavutil/pixdesc.h"
+#include "avfilter.h"
+#include "formats.h"
+#include "internal.h"
+#include "video.h"
+#include "vf_pullup.h"
+
+#define F_HAVE_BREAKS   1
+#define F_HAVE_AFFINITY 2
+
+#define BREAK_LEFT  1
+#define BREAK_RIGHT 2
+
+#define OFFSET(x) offsetof(PullupContext, x)
+#define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+
+static const AVOption pullup_options[] = {
+    { "jl", "set left junk size",  OFFSET(junk_left),  AV_OPT_TYPE_INT, {.i64=1}, 0, INT_MAX, FLAGS },
+    { "jr", "set right junk size", OFFSET(junk_right), AV_OPT_TYPE_INT, {.i64=1}, 0, INT_MAX, FLAGS },
+    { "jt", "set top junk size",   OFFSET(junk_top),   AV_OPT_TYPE_INT, {.i64=4}, 1, INT_MAX, FLAGS },
+    { "jd", "set down junk size",  OFFSET(junk_down),  AV_OPT_TYPE_INT, {.i64=4}, 1, INT_MAX, FLAGS },
+    { "sb", "set strict breaks", OFFSET(strict_breaks), AV_OPT_TYPE_INT, {.i64=0},-1, 1, FLAGS },
+    { "mp", "set metric plane",  OFFSET(metric_plane),  AV_OPT_TYPE_INT, {.i64=0}, 0, 2, FLAGS, "mp" },
+    { "y", "luma",        0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, FLAGS, "mp" },
+    { "u", "chroma blue", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, FLAGS, "mp" },
+    { "v", "chroma red",  0, AV_OPT_TYPE_CONST, {.i64=2}, 0, 0, FLAGS, "mp" },
+    { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(pullup);
+
+static int query_formats(AVFilterContext *ctx)
+{
+    static const enum AVPixelFormat pix_fmts[] = {
+        AV_PIX_FMT_YUVJ444P, AV_PIX_FMT_YUVJ440P,
+        AV_PIX_FMT_YUVJ422P, AV_PIX_FMT_YUVJ420P,
+        AV_PIX_FMT_YUV444P,  AV_PIX_FMT_YUV440P,
+        AV_PIX_FMT_YUV422P,  AV_PIX_FMT_YUV420P,
+        AV_PIX_FMT_YUV411P,  AV_PIX_FMT_YUV410P,
+        AV_PIX_FMT_YUVJ411P, AV_PIX_FMT_GRAY8,
+        AV_PIX_FMT_NONE
+    };
+    ff_set_common_formats(ctx, ff_make_format_list(pix_fmts));
+    return 0;
+}
+
+#define ABS(a) (((a) ^ ((a) >> 31)) - ((a) >> 31))
+
+static int diff_c(const uint8_t *a, const uint8_t *b, int s)
+{
+    int i, j, diff = 0;
+
+    for (i = 0; i < 4; i++) {
+        for (j = 0; j < 8; j++)
+            diff += ABS(a[j] - b[j]);
+        a += s;
+        b += s;
+    }
+
+    return diff;
+}
+
+static int comb_c(const uint8_t *a, const uint8_t *b, int s)
+{
+    int i, j, comb = 0;
+
+    for (i = 0; i < 4; i++) {
+        for (j = 0; j < 8; j++)
+            comb += ABS((a[j] << 1) - b[j - s] - b[j    ]) +
+                    ABS((b[j] << 1) - a[j    ] - a[j + s]);
+        a += s;
+        b += s;
+    }
+
+    return comb;
+}
+
+static int var_c(const uint8_t *a, const uint8_t *b, int s)
+{
+    int i, j, var = 0;
+
+    for (i = 0; i < 3; i++) {
+        for (j = 0; j < 8; j++)
+            var += ABS(a[j] - a[j + s]);
+        a += s;
+    }
+
+    return 4 * var; /* match comb scaling */
+}
+
+static int alloc_metrics(PullupContext *s, PullupField *f)
+{
+    f->diffs = av_calloc(FFALIGN(s->metric_length, 16), sizeof(*f->diffs));
+    f->combs = av_calloc(FFALIGN(s->metric_length, 16), sizeof(*f->combs));
+    f->vars  = av_calloc(FFALIGN(s->metric_length, 16), sizeof(*f->vars));
+
+    if (!f->diffs || !f->combs || !f->vars) {
+        av_freep(&f->diffs);
+        av_freep(&f->combs);
+        av_freep(&f->vars);
+        return AVERROR(ENOMEM);
+    }
+    return 0;
+}
+
+static PullupField *make_field_queue(PullupContext *s, int len)
+{
+    PullupField *head, *f;
+
+    f = head = av_mallocz(sizeof(*head));
+    if (!f)
+        return NULL;
+
+    if (alloc_metrics(s, f) < 0) {
+        av_free(f);
+        return NULL;
+    }
+
+    for (; len > 0; len--) {
+        f->next = av_mallocz(sizeof(*f->next));
+        if (!f->next)
+            return NULL;
+
+        f->next->prev = f;
+        f = f->next;
+        if (alloc_metrics(s, f) < 0)
+            return NULL;
+    }
+
+    f->next = head;
+    head->prev = f;
+
+    return head;
+}
+
+static int config_input(AVFilterLink *inlink)
+{
+    AVFilterContext *ctx = inlink->dst;
+    PullupContext *s = ctx->priv;
+    const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
+    const int mp = s->metric_plane;
+
+    s->planeheight[1] = s->planeheight[2] = FF_CEIL_RSHIFT(inlink->h, desc->log2_chroma_h);
+    s->planeheight[0] = s->planeheight[3] = inlink->h;
+    s->planewidth[1]  = s->planewidth[2]  = FF_CEIL_RSHIFT(inlink->w, desc->log2_chroma_w);
+    s->planewidth[0]  = s->planewidth[3]  = inlink->w;
+    s->nb_planes = av_pix_fmt_count_planes(inlink->format);
+
+    s->metric_w      = (s->planewidth[mp]  - ((s->junk_left + s->junk_right)  << 3)) >> 3;
+    s->metric_h      = (s->planeheight[mp] - ((s->junk_top  + s->junk_bottom) << 1)) >> 3;
+    s->metric_offset = (s->junk_left << 3) + (s->junk_top << 1) * s->planewidth[mp];
+    s->metric_length = s->metric_w * s->metric_h;
+
+    av_log(ctx, AV_LOG_DEBUG, "w: %d h: %d\n", s->metric_w, s->metric_h);
+    av_log(ctx, AV_LOG_DEBUG, "offset: %d length: %d\n", s->metric_offset, s->metric_length);
+
+    s->head = make_field_queue(s, 8);
+    if (!s->head)
+        return AVERROR(ENOMEM);
+
+    s->diff = diff_c;
+    s->comb = comb_c;
+    s->var  = var_c;
+
+    if (ARCH_X86)
+        ff_pullup_init_x86(s);
+    return 0;
+}
+
+static int config_output(AVFilterLink *outlink)
+{
+    outlink->flags |= FF_LINK_FLAG_REQUEST_LOOP;
+    return 0;
+}
+
+static PullupBuffer *pullup_lock_buffer(PullupBuffer *b, int parity)
+{
+    if (!b)
+        return NULL;
+
+    if ((parity + 1) & 1)
+        b->lock[0]++;
+    if ((parity + 1) & 2)
+        b->lock[1]++;
+
+    return b;
+}
+
+static void pullup_release_buffer(PullupBuffer *b, int parity)
+{
+    if (!b)
+        return;
+
+    if ((parity + 1) & 1)
+        b->lock[0]--;
+    if ((parity + 1) & 2)
+        b->lock[1]--;
+}
+
+static int alloc_buffer(PullupContext *s, PullupBuffer *b)
+{
+    int i;
+
+    if (b->planes[0])
+        return 0;
+    for (i = 0; i < s->nb_planes; i++) {
+        b->planes[i] = av_malloc(s->planeheight[i] * s->planewidth[i]);
+    }
+
+    return 0;
+}
+
+static PullupBuffer *pullup_get_buffer(PullupContext *s, int parity)
+{
+    int i;
+
+    /* Try first to get the sister buffer for the previous field */
+    if (parity < 2 && s->last && parity != s->last->parity
+        && !s->last->buffer->lock[parity]) {
+        alloc_buffer(s, s->last->buffer);
+        return pullup_lock_buffer(s->last->buffer, parity);
+    }
+
+    /* Prefer a buffer with both fields open */
+    for (i = 0; i < FF_ARRAY_ELEMS(s->buffers); i++) {
+        if (s->buffers[i].lock[0])
+            continue;
+        if (s->buffers[i].lock[1])
+            continue;
+        alloc_buffer(s, &s->buffers[i]);
+        return pullup_lock_buffer(&s->buffers[i], parity);
+    }
+
+    if (parity == 2)
+        return 0;
+
+    /* Search for any half-free buffer */
+    for (i = 0; i < FF_ARRAY_ELEMS(s->buffers); i++) {
+        if (((parity + 1) & 1) && s->buffers[i].lock[0])
+            continue;
+        if (((parity + 1) & 2) && s->buffers[i].lock[1])
+            continue;
+        alloc_buffer(s, &s->buffers[i]);
+        return pullup_lock_buffer(&s->buffers[i], parity);
+    }
+
+    return NULL;
+}
+
+static int queue_length(PullupField *begin, PullupField *end)
+{
+    PullupField *f;
+    int count = 1;
+
+    if (!begin || !end)
+        return 0;
+
+    for (f = begin; f != end; f = f->next)
+        count++;
+
+    return count;
+}
+
+static int find_first_break(PullupField *f, int max)
+{
+    int i;
+
+    for (i = 0; i < max; i++) {
+        if (f->breaks & BREAK_RIGHT || f->next->breaks & BREAK_LEFT)
+            return i + 1;
+        f = f->next;
+    }
+
+    return 0;
+}
+
+static void compute_breaks(PullupContext *s, PullupField *f0)
+{
+    PullupField *f1 = f0->next;
+    PullupField *f2 = f1->next;
+    PullupField *f3 = f2->next;
+    int i, l, max_l = 0, max_r = 0;
+
+    if (f0->flags & F_HAVE_BREAKS)
+        return;
+
+    f0->flags |= F_HAVE_BREAKS;
+
+    /* Special case when fields are 100% identical */
+    if (f0->buffer == f2->buffer && f1->buffer != f3->buffer) {
+        f2->breaks |= BREAK_RIGHT;
+        return;
+    }
+
+    if (f0->buffer != f2->buffer && f1->buffer == f3->buffer) {
+        f1->breaks |= BREAK_LEFT;
+        return;
+    }
+
+    for (i = 0; i < s->metric_length; i++) {
+        l = f2->diffs[i] - f3->diffs[i];
+
+        if ( l > max_l)
+            max_l =  l;
+        if (-l > max_r)
+            max_r = -l;
+    }
+
+    /* Don't get tripped up when differences are mostly quant error */
+    if (max_l + max_r < 128)
+        return;
+    if (max_l > 4 * max_r)
+        f1->breaks |= BREAK_LEFT;
+    if (max_r > 4 * max_l)
+        f2->breaks |= BREAK_RIGHT;
+}
+
+static void compute_affinity(PullupContext *s, PullupField *f)
+{
+    int i, max_l = 0, max_r = 0, l;
+
+    if (f->flags & F_HAVE_AFFINITY)
+        return;
+
+    f->flags |= F_HAVE_AFFINITY;
+
+    if (f->buffer == f->next->next->buffer) {
+        f->affinity             =  1;
+        f->next->affinity       =  0;
+        f->next->next->affinity = -1;
+        f->next->flags         |= F_HAVE_AFFINITY;
+        f->next->next->flags   |= F_HAVE_AFFINITY;
+        return;
+    }
+
+    for (i = 0; i < s->metric_length; i++) {
+        int v  = f->vars[i];
+        int lv = f->prev->vars[i];
+        int rv = f->next->vars[i];
+        int lc = f->combs[i] - (v + lv) + ABS(v - lv);
+        int rc = f->next->combs[i] - (v + rv) + ABS(v - rv);
+
+        lc = FFMAX(lc, 0);
+        rc = FFMAX(rc, 0);
+        l  = lc - rc;
+
+        if ( l > max_l)
+            max_l =  l;
+        if (-l > max_r)
+            max_r = -l;
+    }
+
+    if (max_l + max_r < 64)
+        return;
+
+    if (max_r > 6 * max_l)
+        f->affinity = -1;
+    else if (max_l > 6 * max_r)
+        f->affinity =  1;
+}
+
+static int decide_frame_length(PullupContext *s)
+{
+    PullupField *f0 = s->first;
+    PullupField *f1 = f0->next;
+    PullupField *f2 = f1->next;
+    PullupField *f;
+    int i, l, n;
+
+    if (queue_length(s->first, s->last) < 4)
+        return 0;
+
+    f = s->first;
+    n = queue_length(f, s->last);
+    for (i = 0; i < n - 1; i++) {
+        if (i < n - 3)
+            compute_breaks(s, f);
+
+        compute_affinity(s, f);
+
+        f = f->next;
+    }
+
+    if (f0->affinity == -1)
+        return 1;
+
+    l = find_first_break(f0, 3);
+
+    if (l == 1 && s->strict_breaks < 0)
+        l = 0;
+
+    switch (l) {
+    case 1:
+        return 1 + (s->strict_breaks < 1 && f0->affinity == 1 && f1->affinity == -1);
+    case 2:
+        /* FIXME: strictly speaking, f0->prev is no longer valid... :) */
+        if (s->strict_pairs
+            && (f0->prev->breaks & BREAK_RIGHT) && (f2->breaks & BREAK_LEFT)
+            && (f0->affinity != 1 || f1->affinity != -1) )
+            return 1;
+        return 1 + (f1->affinity != 1);
+    case 3:
+        return 2 + (f2->affinity != 1);
+    default:
+        /* 9 possibilities covered before switch */
+        if (f1->affinity == 1)
+            return 1; /* covers 6 */
+        else if (f1->affinity == -1)
+            return 2; /* covers 6 */
+        else if (f2->affinity == -1) { /* covers 2 */
+            return (f0->affinity == 1) ? 3 : 1;
+        } else {
+            return 2; /* the remaining 6 */
+        }
+    }
+}
+
+static PullupFrame *pullup_get_frame(PullupContext *s)
+{
+    PullupFrame *fr = &s->frame;
+    int i, n = decide_frame_length(s);
+    int aff = s->first->next->affinity;
+
+    av_assert1(n < FF_ARRAY_ELEMS(fr->ifields));
+    if (!n || fr->lock)
+        return NULL;
+
+    fr->lock++;
+    fr->length = n;
+    fr->parity = s->first->parity;
+    fr->buffer = 0;
+
+    for (i = 0; i < n; i++) {
+        /* We cheat and steal the buffer without release+relock */
+        fr->ifields[i]   = s->first->buffer;
+        s->first->buffer = 0;
+        s->first         = s->first->next;
+    }
+
+    if (n == 1) {
+        fr->ofields[fr->parity    ] = fr->ifields[0];
+        fr->ofields[fr->parity ^ 1] = 0;
+    } else if (n == 2) {
+        fr->ofields[fr->parity    ] = fr->ifields[0];
+        fr->ofields[fr->parity ^ 1] = fr->ifields[1];
+    } else if (n == 3) {
+        if (!aff)
+            aff = (fr->ifields[0] == fr->ifields[1]) ? -1 : 1;
+        fr->ofields[fr->parity    ] = fr->ifields[1 + aff];
+        fr->ofields[fr->parity ^ 1] = fr->ifields[1      ];
+    }
+
+    pullup_lock_buffer(fr->ofields[0], 0);
+    pullup_lock_buffer(fr->ofields[1], 1);
+
+    if (fr->ofields[0] == fr->ofields[1]) {
+        fr->buffer = fr->ofields[0];
+        pullup_lock_buffer(fr->buffer, 2);
+        return fr;
+    }
+
+    return fr;
+}
+
+static void pullup_release_frame(PullupFrame *f)
+{
+    int i;
+
+    for (i = 0; i < f->length; i++)
+        pullup_release_buffer(f->ifields[i], f->parity ^ (i & 1));
+
+    pullup_release_buffer(f->ofields[0], 0);
+    pullup_release_buffer(f->ofields[1], 1);
+
+    if (f->buffer)
+        pullup_release_buffer(f->buffer, 2);
+    f->lock--;
+}
+
+static void compute_metric(PullupContext *s, int *dest,
+                           PullupField *fa, int pa, PullupField *fb, int pb,
+                           int (*func)(const uint8_t *, const uint8_t *, int))
+{
+    int mp = s->metric_plane;
+    int xstep = 8;
+    int ystep = s->planewidth[mp] << 3;
+    int stride = s->planewidth[mp] << 1; /* field stride */
+    int w = s->metric_w * xstep;
+    uint8_t *a, *b;
+    int x, y;
+
+    if (!fa->buffer || !fb->buffer)
+        return;
+
+    /* Shortcut for duplicate fields (e.g. from RFF flag) */
+    if (fa->buffer == fb->buffer && pa == pb) {
+        memset(dest, 0, s->metric_length * sizeof(*dest));
+        return;
+    }
+
+    a = fa->buffer->planes[mp] + pa * s->planewidth[mp] + s->metric_offset;
+    b = fb->buffer->planes[mp] + pb * s->planewidth[mp] + s->metric_offset;
+
+    for (y = 0; y < s->metric_h; y++) {
+        for (x = 0; x < w; x += xstep)
+            *dest++ = func(a + x, b + x, stride);
+        a += ystep; b += ystep;
+    }
+}
+
+static int check_field_queue(PullupContext *s)
+{
+    int ret;
+
+    if (s->head->next == s->first) {
+        PullupField *f = av_mallocz(sizeof(*f));
+
+        if (!f)
+            return AVERROR(ENOMEM);
+
+        if ((ret = alloc_metrics(s, f)) < 0) {
+            av_free(f);
+            return ret;
+        }
+
+        f->prev        = s->head;
+        f->next        = s->first;
+        s->head->next  = f;
+        s->first->prev = f;
+    }
+
+    return 0;
+}
+
+static void pullup_submit_field(PullupContext *s, PullupBuffer *b, int parity)
+{
+    PullupField *f;
+
+    /* Grow the circular list if needed */
+    if (check_field_queue(s) < 0)
+        return;
+
+    /* Cannot have two fields of same parity in a row; drop the new one */
+    if (s->last && s->last->parity == parity)
+        return;
+
+    f = s->head;
+    f->parity   = parity;
+    f->buffer   = pullup_lock_buffer(b, parity);
+    f->flags    = 0;
+    f->breaks   = 0;
+    f->affinity = 0;
+
+    compute_metric(s, f->diffs, f, parity, f->prev->prev, parity, s->diff);
+    compute_metric(s, f->combs, parity ? f->prev : f, 0, parity ? f : f->prev, 1, s->comb);
+    compute_metric(s, f->vars, f, parity, f, -1, s->var);
+    emms_c();
+
+    /* Advance the circular list */
+    if (!s->first)
+        s->first = s->head;
+
+    s->last = s->head;
+    s->head = s->head->next;
+}
+
+static void copy_field(PullupContext *s,
+                       PullupBuffer *dst, PullupBuffer *src, int parity)
+{
+    uint8_t *dd, *ss;
+    int i;
+
+    for (i = 0; i < s->nb_planes; i++) {
+        ss = src->planes[i] + parity * s->planewidth[i];
+        dd = dst->planes[i] + parity * s->planewidth[i];
+
+        av_image_copy_plane(dd, s->planewidth[i] << 1,
+                            ss, s->planewidth[i] << 1,
+                            s->planewidth[i], s->planeheight[i] >> 1);
+    }
+}
+
+static void pullup_pack_frame(PullupContext *s, PullupFrame *fr)
+{
+    int i;
+
+    if (fr->buffer)
+        return;
+
+    if (fr->length < 2)
+        return; /* FIXME: deal with this */
+
+    for (i = 0; i < 2; i++) {
+        if (fr->ofields[i]->lock[i^1])
+            continue;
+
+        fr->buffer = fr->ofields[i];
+        pullup_lock_buffer(fr->buffer, 2);
+        copy_field(s, fr->buffer, fr->ofields[i^1], i^1);
+        return;
+    }
+
+    fr->buffer = pullup_get_buffer(s, 2);
+
+    copy_field(s, fr->buffer, fr->ofields[0], 0);
+    copy_field(s, fr->buffer, fr->ofields[1], 1);
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *in)
+{
+    AVFilterContext *ctx = inlink->dst;
+    AVFilterLink *outlink = ctx->outputs[0];
+    PullupContext *s = ctx->priv;
+    PullupBuffer *b;
+    PullupFrame *f;
+    AVFrame *out;
+    int p, ret = 0;
+
+    b = pullup_get_buffer(s, 2);
+    if (!b) {
+        av_log(ctx, AV_LOG_WARNING, "Could not get buffer!\n");
+        f = pullup_get_frame(s);
+        pullup_release_frame(f);
+        goto end;
+    }
+
+    av_image_copy(b->planes, s->planewidth,
+                  (const uint8_t**)in->data, in->linesize,
+                  inlink->format, inlink->w, inlink->h);
+
+    p = !!in->interlaced_frame;
+    pullup_submit_field(s, b, p  );
+    pullup_submit_field(s, b, p^1);
+
+    if (in->repeat_pict)
+        pullup_submit_field(s, b, p);
+
+    pullup_release_buffer(b, 2);
+
+    f = pullup_get_frame(s);
+    if (!f)
+        goto end;
+
+    if (f->length < 2) {
+        pullup_release_frame(f);
+        f = pullup_get_frame(s);
+        if (!f)
+            goto end;
+        if (f->length < 2) {
+            pullup_release_frame(f);
+            if (!in->repeat_pict)
+                goto end;
+            f = pullup_get_frame(s);
+            if (!f)
+                goto end;
+            if (f->length < 2) {
+                pullup_release_frame(f);
+                goto end;
+            }
+        }
+    }
+
+    /* If the frame isn't already exportable... */
+    if (!f->buffer)
+        pullup_pack_frame(s, f);
+
+    out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
+    if (!out) {
+        ret = AVERROR(ENOMEM);
+        goto end;
+    }
+    av_frame_copy_props(out, in);
+
+    av_image_copy(out->data, out->linesize,
+                  (const uint8_t**)f->buffer->planes, s->planewidth,
+                  inlink->format, inlink->w, inlink->h);
+
+    ret = ff_filter_frame(outlink, out);
+    pullup_release_frame(f);
+end:
+    av_frame_free(&in);
+    return ret;
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    PullupContext *s = ctx->priv;
+    PullupField *f;
+    int i;
+
+    f = s->head;
+    do {
+        if (!f)
+            break;
+
+        av_free(f->diffs);
+        av_free(f->combs);
+        av_free(f->vars);
+        f = f->next;
+        av_freep(&f->prev);
+    } while (f != s->last);
+    av_freep(&s->last);
+
+    for (i = 0; i < FF_ARRAY_ELEMS(s->buffers); i++) {
+        av_freep(&s->buffers[i].planes[0]);
+        av_freep(&s->buffers[i].planes[1]);
+        av_freep(&s->buffers[i].planes[2]);
+    }
+}
+
+static const AVFilterPad pullup_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+        .filter_frame = filter_frame,
+        .config_props = config_input,
+    },
+    { NULL }
+};
+
+static const AVFilterPad pullup_outputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+        .config_props = config_output,
+    },
+    { NULL }
+};
+
+AVFilter avfilter_vf_pullup = {
+    .name          = "pullup",
+    .description   = NULL_IF_CONFIG_SMALL("Pullup from field sequence to frames."),
+    .priv_size     = sizeof(PullupContext),
+    .priv_class    = &pullup_class,
+    .uninit        = uninit,
+    .query_formats = query_formats,
+    .inputs        = pullup_inputs,
+    .outputs       = pullup_outputs,
+};

libavfilter/vf_pullup.h

+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg 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.
+ *
+ * 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU 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
+ */
+
+#ifndef AVFILTER_PULLUP_H
+#define AVFILTER_PULLUP_H
+
+#include "avfilter.h"
+
+typedef struct PullupBuffer {
+    int lock[2];
+    uint8_t *planes[4];
+} PullupBuffer;
+
+typedef struct PullupField {
+    int parity;
+    PullupBuffer *buffer;
+    unsigned flags;
+    int breaks;
+    int affinity;
+    int *diffs;
+    int *combs;
+    int *vars;
+    struct PullupField *prev, *next;
+} PullupField;
+
+typedef struct PullupFrame {
+    int lock;
+    int length;
+    int parity;
+    PullupBuffer *ifields[4], *ofields[2];
+    PullupBuffer *buffer;
+} PullupFrame;
+
+typedef struct PullupContext {
+    const AVClass *class;
+    int junk_left, junk_right, junk_top, junk_down, junk_bottom;
+    int metric_plane;
+    int strict_breaks;
+    int strict_pairs;
+    int metric_w, metric_h, metric_length;
+    int metric_offset;
+    int nb_planes;
+    int planewidth[4];
+    int planeheight[4];
+    PullupField *first, *last, *head;
+    PullupBuffer buffers[10];
+    PullupFrame frame;
+
+    int (*diff)(const uint8_t *a, const uint8_t *b, int s);
+    int (*comb)(const uint8_t *a, const uint8_t *b, int s);
+    int (*var )(const uint8_t *a, const uint8_t *b, int s);
+} PullupContext;
+
+void ff_pullup_init_x86(PullupContext *s);
+
+#endif /* AVFILTER_PULLUP_H */

libavfilter/x86/Makefile

 OBJS-$(CONFIG_GRADFUN_FILTER)                += x86/vf_gradfun.o
 OBJS-$(CONFIG_HQDN3D_FILTER)                 += x86/vf_hqdn3d_init.o
+OBJS-$(CONFIG_PULLUP_FILTER)                 += x86/vf_pullup_init.o
 OBJS-$(CONFIG_SPP_FILTER)                    += x86/vf_spp.o
 OBJS-$(CONFIG_VOLUME_FILTER)                 += x86/af_volume_init.o
 OBJS-$(CONFIG_YADIF_FILTER)                  += x86/vf_yadif_init.o
 
 YASM-OBJS-$(CONFIG_HQDN3D_FILTER)            += x86/vf_hqdn3d.o
+YASM-OBJS-$(CONFIG_PULLUP_FILTER)            += x86/vf_pullup.o
 YASM-OBJS-$(CONFIG_VOLUME_FILTER)            += x86/af_volume.o
 YASM-OBJS-$(CONFIG_YADIF_FILTER)             += x86/vf_yadif.o x86/yadif-16.o x86/yadif-10.o

libavfilter/x86/vf_pullup.asm

+;*****************************************************************************
+;* x86-optimized functions for pullup filter
+;*
+;* This file is part of FFmpeg.
+;*
+;* FFmpeg 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.
+;*
+;* 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 General Public License for more details.
+;*
+;* You should have received a copy of the GNU 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.
+;******************************************************************************
+
+%include "libavutil/x86/x86util.asm"
+
+SECTION_TEXT
+
+INIT_MMX mmx
+cglobal pullup_filter_diff, 3, 5, 8, first, second, size
+    mov        r3, 4
+    pxor       m4, m4
+    pxor       m7, m7
+
+.loop:
+    movq       m0, [firstq]
+    movq       m2, [firstq]
+    add        firstq, sizeq
+    movq       m1, [secondq]
+    add        secondq, sizeq
+    psubusb    m2, m1
+    psubusb    m1, m0
+    movq       m0, m2
+    movq       m3, m1
+    punpcklbw  m0, m7
+    punpcklbw  m1, m7
+    punpckhbw  m2, m7
+    punpckhbw  m3, m7
+    paddw      m4, m0
+    paddw      m4, m1
+    paddw      m4, m2
+    paddw      m4, m3
+
+    dec        r3
+    jnz .loop
+
+    movq       m3, m4
+    punpcklwd  m4, m7
+    punpckhwd  m3, m7
+    paddd      m3, m4
+    movd      eax, m3
+    psrlq      m3, 32
+    movd       r4, m3
+    add       eax, r4
+    RET
+
+INIT_MMX mmx
+cglobal pullup_filter_comb, 3, 5, 8, first, second, size
+    mov        r3, 4
+    pxor       m6, m6
+    pxor       m7, m7
+    sub        secondq, sizeq
+
+.loop:
+    movq       m0, [secondq]
+    movq       m1, [secondq]
+    punpcklbw  m0, m7
+    movq       m2, [secondq+sizeq]
+    punpcklbw  m1, m7
+    punpcklbw  m2, m7
+    paddw      m0, m0
+    paddw      m1, m2
+    movq       m2, m0
+    psubusw    m0, m1
+    psubusw    m1, m2
+    paddw      m6, m0
+    paddw      m6, m1
+
+    movq       m0, [firstq]
+    movq       m1, [secondq]
+    punpckhbw  m0, m7
+    movq       m2, [secondq+sizeq]
+    punpckhbw  m1, m7
+    punpckhbw  m2, m7
+    paddw      m0, m0
+    paddw      m1, m2
+    movq       m2, m0
+    psubusw    m0, m1
+    psubusw    m1, m2
+    paddw      m6, m0
+    paddw      m6, m1
+
+    movq       m0, [secondq+sizeq]
+    movq       m1, [firstq]
+    punpcklbw  m0, m7
+    movq       m2, [firstq+sizeq]
+    punpcklbw  m1, m7
+    punpcklbw  m2, m7
+    paddw      m0, m0
+    paddw      m1, m2
+    movq       m2, m0
+    psubusw    m0, m1
+    psubusw    m1, m2
+    paddw      m6, m0
+    paddw      m6, m1
+
+    movq       m0, [secondq+sizeq]
+    movq       m1, [firstq]
+    punpckhbw  m0, m7
+    movq       m2, [firstq+sizeq]
+    punpckhbw  m1, m7
+    punpckhbw  m2, m7
+    paddw      m0, m0
+    paddw      m1, m2
+    movq       m2, m0
+    psubusw    m0, m1
+    psubusw    m1, m2
+    paddw      m6, m0
+    paddw      m6, m1
+
+    add        firstq, sizeq
+    add        secondq, sizeq
+    dec        r3
+    jnz .loop
+
+    movq       m5, m6
+    punpcklwd  m6, m7
+    punpckhwd  m5, m7
+    paddd      m5, m6
+    movd      eax, m5
+    psrlq      m5, 32
+    movd       r4, m5
+    add       eax, r4
+    RET
+
+INIT_MMX mmx
+cglobal pullup_filter_var, 3, 5, 8, first, second, size
+    mov        r3, 3
+    pxor       m4, m4
+    pxor       m7, m7
+
+.loop:
+    movq       m0, [firstq]
+    movq       m2, [firstq]
+    movq       m1, [firstq+sizeq]
+    add        firstq, sizeq
+    psubusb    m2, m1
+    psubusb    m1, m0
+    movq       m0, m2
+    movq       m3, m1
+    punpcklbw  m0, m7
+    punpcklbw  m1, m7
+    punpckhbw  m2, m7
+    punpckhbw  m3, m7
+    paddw      m4, m0
+    paddw      m4, m1
+    paddw      m4, m2
+    paddw      m4, m3
+
+    dec        r3
+    jnz .loop
+
+    movq       m3, m4
+    punpcklwd  m4, m7
+    punpckhwd  m3, m7
+    paddd      m3, m4
+    movd      eax, m3
+    psrlq      m3, 32
+    movd       r4, m3
+    add       eax, r4
+    shl       eax, 2
+    RET

libavfilter/x86/vf_pullup_init.c

+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg 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.
+ *
+ * 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU 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
+*/
+
+#include "libavutil/attributes.h"
+#include "libavutil/cpu.h"
+#include "libavutil/mem.h"
+#include "libavutil/x86/asm.h"
+#include "libavutil/x86/cpu.h"
+#include "libavfilter/vf_pullup.h"
+
+int ff_pullup_filter_diff_mmx(const uint8_t *a, const uint8_t *b, int s);
+int ff_pullup_filter_comb_mmx(const uint8_t *a, const uint8_t *b, int s);
+int ff_pullup_filter_var_mmx (const uint8_t *a, const uint8_t *b, int s);
+
+av_cold void ff_pullup_init_x86(PullupContext *s)
+{
+#if HAVE_YASM
+    int cpu_flags = av_get_cpu_flags();
+
+    if (EXTERNAL_MMX(cpu_flags)) {
+        s->diff = ff_pullup_filter_diff_mmx;
+        s->comb = ff_pullup_filter_comb_mmx;
+        s->var  = ff_pullup_filter_var_mmx;
+    }
+#endif
+}