view libmpcodecs/vf_gradfun.c @ 30795:1001c606f94c

Make emulated Win32 critical sections thread safe. Earlier, cs->locked was accessed outside the mutex to get around the problem that default pthread mutexes are not recursive (ie., you cannot do a double-lock from the same thread), causing a thread-safety problem, as both detected by Helgrind and showing up in some multithreaded codecs. The ideal solution here would be to simply use recursive pthread mutexes, but there were concerns about reduced debuggability and possibly portability. Thus, instead, rewrite the critical sections to be a simple lock count (with owner) protected by a regular mutex. Whenever a thread wants to enter the critical section and lock_count is not 0, it sleeps on a special event that tells it when the critical section is available.

/*
 * Copyright (C) 2009 Loren Merritt <[e-mail]>
 *
 * This file is part of MPlayer.
 *
 * MPlayer 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.
 *
 * MPlayer 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 MPlayer; if not, write to the Free Software Foundation, Inc.,
 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */

/*
 * Debanding algorithm (from gradfun2db by prunedtree):
 * Boxblur.
 * Foreach pixel, if it's within threshold of the blurred value, make it closer.
 * So now we have a smoothed and higher bitdepth version of all the shallow
 * gradients, while leaving detailed areas untouched.
 * Dither it back to 8bit.
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <inttypes.h>

#include "config.h"
#include "cpudetect.h"
#include "img_format.h"
#include "mp_image.h"
#include "vf.h"
#include "libvo/fastmemcpy.h"
#include "libavutil/avutil.h"
#include "libavutil/x86_cpu.h"

struct vf_priv_s {
    int thresh;
    int radius;
    uint16_t *buf;
    void (*filter_line)(uint8_t *dst, uint8_t *src, uint16_t *dc,
                        int width, int thresh, const uint16_t *dithers);
    void (*blur_line)(uint16_t *dc, uint16_t *buf, uint16_t *buf1,
                      uint8_t *src, int sstride, int width);
};

static const uint16_t __attribute__((aligned(16))) pw_7f[8] = {127,127,127,127,127,127,127,127};
static const uint16_t __attribute__((aligned(16))) pw_ff[8] = {255,255,255,255,255,255,255,255};
static const uint16_t __attribute__((aligned(16))) dither[8][8] = {
    {  0, 96, 24,120,  6,102, 30,126 },
    { 64, 32, 88, 56, 70, 38, 94, 62 },
    { 16,112,  8,104, 22,118, 14,110 },
    { 80, 48, 72, 40, 86, 54, 78, 46 },
    {  4,100, 28,124,  2, 98, 26,122 },
    { 68, 36, 92, 60, 66, 34, 90, 58 },
    { 20,116, 12,108, 18,114, 10,106 },
    { 84, 52, 76, 44, 82, 50, 74, 42 },
};

static void filter_line_c(uint8_t *dst, uint8_t *src, uint16_t *dc,
                          int width, int thresh, const uint16_t *dithers)
{
    int x;
    for (x=0; x<width; x++, dc+=x&1) {
        int pix = src[x]<<7;
        int delta = dc[0] - pix;
        int m = abs(delta) * thresh >> 16;
        m = FFMAX(0, 127-m);
        m = m*m*delta >> 14;
        pix += m + dithers[x&7];
        dst[x] = av_clip_uint8(pix>>7);
    }
}

static void blur_line_c(uint16_t *dc, uint16_t *buf, uint16_t *buf1,
                        uint8_t *src, int sstride, int width)
{
    int x, v, old;
    for (x=0; x<width; x++) {
        v = buf1[x] + src[2*x] + src[2*x+1] + src[2*x+sstride] + src[2*x+1+sstride];
        old = buf[x];
        buf[x] = v;
        dc[x] = v - old;
    }
}

#if HAVE_SSSE3
static void filter_line_mmx2(uint8_t *dst, uint8_t *src, uint16_t *dc,
                             int width, int thresh, const uint16_t *dithers)
{
    intptr_t x;
    if (width&3) {
        x = width&~3;
        filter_line_c(dst+x, src+x, dc+x/2, width-x, thresh, dithers);
        width = x;
    }
    x = -width;
    asm volatile(
        "movd          %4, %%mm5 \n"
        "pxor       %%mm7, %%mm7 \n"
        "pshufw $0, %%mm5, %%mm5 \n"
        "movq          %6, %%mm6 \n"
        "movq          %5, %%mm4 \n"
        "1: \n"
        "movd     (%2,%0), %%mm0 \n"
        "movd     (%3,%0), %%mm1 \n"
        "punpcklbw  %%mm7, %%mm0 \n"
        "punpcklwd  %%mm1, %%mm1 \n"
        "psllw         $7, %%mm0 \n"
        "pxor       %%mm2, %%mm2 \n"
        "psubw      %%mm0, %%mm1 \n" // delta = dc - pix
        "psubw      %%mm1, %%mm2 \n"
        "pmaxsw     %%mm1, %%mm2 \n"
        "pmulhuw    %%mm5, %%mm2 \n" // m = abs(delta) * thresh >> 16
        "psubw      %%mm6, %%mm2 \n"
        "pminsw     %%mm7, %%mm2 \n" // m = -max(0, 127-m)
        "pmullw     %%mm2, %%mm2 \n"
        "paddw      %%mm4, %%mm0 \n" // pix += dither
        "pmulhw     %%mm2, %%mm1 \n"
        "psllw         $2, %%mm1 \n" // m = m*m*delta >> 14
        "paddw      %%mm1, %%mm0 \n" // pix += m
        "psraw         $7, %%mm0 \n"
        "packuswb   %%mm0, %%mm0 \n"
        "movd       %%mm0, (%1,%0) \n" // dst = clip(pix>>7)
        "add           $4, %0 \n"
        "jl 1b \n"
        "emms \n"
        :"+r"(x)
        :"r"(dst+width), "r"(src+width), "r"(dc+width/2),
         "rm"(thresh), "m"(*dithers), "m"(*pw_7f)
        :"memory"
    );
}

static void filter_line_ssse3(uint8_t *dst, uint8_t *src, uint16_t *dc,
                              int width, int thresh, const uint16_t *dithers)
{
    intptr_t x;
    if (width&7) {
        // could be 10% faster if I somehow eliminated this
        x = width&~7;
        filter_line_c(dst+x, src+x, dc+x/2, width-x, thresh, dithers);
        width = x;
    }
    x = -width;
    asm volatile(
        "movd           %4, %%xmm5 \n"
        "pxor       %%xmm7, %%xmm7 \n"
        "pshuflw $0,%%xmm5, %%xmm5 \n"
        "movdqa         %6, %%xmm6 \n"
        "punpcklqdq %%xmm5, %%xmm5 \n"
        "movdqa         %5, %%xmm4 \n"
        "1: \n"
        "movq      (%2,%0), %%xmm0 \n"
        "movq      (%3,%0), %%xmm1 \n"
        "punpcklbw  %%xmm7, %%xmm0 \n"
        "punpcklwd  %%xmm1, %%xmm1 \n"
        "psllw          $7, %%xmm0 \n"
        "psubw      %%xmm0, %%xmm1 \n" // delta = dc - pix
        "pabsw      %%xmm1, %%xmm2 \n"
        "pmulhuw    %%xmm5, %%xmm2 \n" // m = abs(delta) * thresh >> 16
        "psubw      %%xmm6, %%xmm2 \n"
        "pminsw     %%xmm7, %%xmm2 \n" // m = -max(0, 127-m)
        "pmullw     %%xmm2, %%xmm2 \n"
        "psllw          $1, %%xmm2 \n"
        "paddw      %%xmm4, %%xmm0 \n" // pix += dither
        "pmulhrsw   %%xmm2, %%xmm1 \n" // m = m*m*delta >> 14
        "paddw      %%xmm1, %%xmm0 \n" // pix += m
        "psraw          $7, %%xmm0 \n"
        "packuswb   %%xmm0, %%xmm0 \n"
        "movq       %%xmm0, (%1,%0) \n" // dst = clip(pix>>7)
        "add            $8, %0 \n"
        "jl 1b \n"
        :"+&r"(x)
        :"r"(dst+width), "r"(src+width), "r"(dc+width/2),
         "rm"(thresh), "m"(*dithers), "m"(*pw_7f)
        :"memory"
    );
}

#define BLURV(load)\
    intptr_t x = -2*width;\
    asm volatile(\
        "movdqa %6, %%xmm7 \n"\
        "1: \n"\
        load"   (%4,%0), %%xmm0 \n"\
        load"   (%5,%0), %%xmm1 \n"\
        "movdqa  %%xmm0, %%xmm2 \n"\
        "movdqa  %%xmm1, %%xmm3 \n"\
        "psrlw       $8, %%xmm0 \n"\
        "psrlw       $8, %%xmm1 \n"\
        "pand    %%xmm7, %%xmm2 \n"\
        "pand    %%xmm7, %%xmm3 \n"\
        "paddw   %%xmm1, %%xmm0 \n"\
        "paddw   %%xmm3, %%xmm2 \n"\
        "paddw   %%xmm2, %%xmm0 \n"\
        "paddw  (%2,%0), %%xmm0 \n"\
        "movdqa (%1,%0), %%xmm1 \n"\
        "movdqa  %%xmm0, (%1,%0) \n"\
        "psubw   %%xmm1, %%xmm0 \n"\
        "movdqa  %%xmm0, (%3,%0) \n"\
        "add        $16, %0 \n"\
        "jl 1b \n"\
        :"+&r"(x)\
        :"r"(buf+width),\
         "r"(buf1+width),\
         "r"(dc+width),\
         "r"(src+width*2),\
         "r"(src+width*2+sstride),\
         "m"(*pw_ff)\
        :"memory"\
    );

#if HAVE_6REGS
static void blur_line_sse2(uint16_t *dc, uint16_t *buf, uint16_t *buf1,
                           uint8_t *src, int sstride, int width)
{
    if (((intptr_t)src|sstride)&15) {
        BLURV("movdqu");
    } else {
        BLURV("movdqa");
    }
}
#endif // HAVE_6REGS
#endif // HAVE_SSSE3

static void filter(struct vf_priv_s *ctx, uint8_t *dst, uint8_t *src,
                   int width, int height, int dstride, int sstride, int r)
{
    int bstride = ((width+15)&~15)/2;
    int y;
    uint32_t dc_factor = (1<<21)/(r*r);
    uint16_t *dc = ctx->buf+16;
    uint16_t *buf = ctx->buf+bstride+32;
    int thresh = ctx->thresh;

    memset(dc, 0, (bstride+16)*sizeof(*buf));
    for (y=0; y<r; y++)
        ctx->blur_line(dc, buf+y*bstride, buf+(y-1)*bstride, src+2*y*sstride, sstride, width/2);
    for (;;) {
        if (y < height-r) {
            int mod = ((y+r)/2)%r;
            uint16_t *buf0 = buf+mod*bstride;
            uint16_t *buf1 = buf+(mod?mod-1:r-1)*bstride;
            int x, v;
            ctx->blur_line(dc, buf0, buf1, src+(y+r)*sstride, sstride, width/2);
            for (x=v=0; x<r; x++)
                v += dc[x];
            for (; x<width/2; x++) {
                v += dc[x] - dc[x-r];
                dc[x-r] = v * dc_factor >> 16;
            }
            for (; x<(width+r+1)/2; x++)
                dc[x-r] = v * dc_factor >> 16;
            for (x=-r/2; x<0; x++)
                dc[x] = dc[0];
        }
        if (y == r) {
            for (y=0; y<r; y++)
                ctx->filter_line(dst+y*dstride, src+y*sstride, dc-r/2, width, thresh, dither[y&7]);
        }
        ctx->filter_line(dst+y*dstride, src+y*sstride, dc-r/2, width, thresh, dither[y&7]);
        if (++y >= height) break;
        ctx->filter_line(dst+y*dstride, src+y*sstride, dc-r/2, width, thresh, dither[y&7]);
        if (++y >= height) break;
    }
}

static void get_image(struct vf_instance *vf, mp_image_t *mpi)
{
    if (mpi->flags&MP_IMGFLAG_PRESERVE) return; // don't change
    // ok, we can do pp in-place:
    vf->dmpi = vf_get_image(vf->next, mpi->imgfmt,
                            mpi->type, mpi->flags, mpi->width, mpi->height);
    mpi->planes[0] = vf->dmpi->planes[0];
    mpi->stride[0] = vf->dmpi->stride[0];
    mpi->width = vf->dmpi->width;
    if (mpi->flags&MP_IMGFLAG_PLANAR){
        mpi->planes[1] = vf->dmpi->planes[1];
        mpi->planes[2] = vf->dmpi->planes[2];
        mpi->stride[1] = vf->dmpi->stride[1];
        mpi->stride[2] = vf->dmpi->stride[2];
    }
    mpi->flags |= MP_IMGFLAG_DIRECT;
}

static int put_image(struct vf_instance *vf, mp_image_t *mpi, double pts)
{
    mp_image_t *dmpi = vf->dmpi;
    int p;

    if (!(mpi->flags&MP_IMGFLAG_DIRECT)) {
        // no DR, so get a new image. hope we'll get DR buffer:
        dmpi = vf_get_image(vf->next,mpi->imgfmt, MP_IMGTYPE_TEMP,
                            MP_IMGFLAG_ACCEPT_STRIDE|MP_IMGFLAG_PREFER_ALIGNED_STRIDE,
                            mpi->w, mpi->h);
    }
    vf_clone_mpi_attributes(dmpi, mpi);

    for (p=0; p<mpi->num_planes; p++) {
        int w = mpi->w;
        int h = mpi->h;
        int r = vf->priv->radius;
        if (p) {
            w >>= mpi->chroma_x_shift;
            h >>= mpi->chroma_y_shift;
            r = ((r>>mpi->chroma_x_shift) + (r>>mpi->chroma_y_shift)) / 2;
            r = av_clip((r+1)&~1,4,32);
        }
        if (FFMIN(w,h) > 2*r)
            filter(vf->priv, dmpi->planes[p], mpi->planes[p], w, h,
                             dmpi->stride[p], mpi->stride[p], r);
        else if (dmpi->planes[p] != mpi->planes[p])
            memcpy_pic(dmpi->planes[p], mpi->planes[p], w, h,
                       dmpi->stride[p], mpi->stride[p]);
    }

    return vf_next_put_image(vf, dmpi, pts);
}

static int query_format(struct vf_instance *vf, unsigned int fmt)
{
    switch (fmt){
    case IMGFMT_YVU9:
    case IMGFMT_IF09:
    case IMGFMT_YV12:
    case IMGFMT_I420:
    case IMGFMT_IYUV:
    case IMGFMT_CLPL:
    case IMGFMT_Y800:
    case IMGFMT_Y8:
    case IMGFMT_NV12:
    case IMGFMT_NV21:
    case IMGFMT_444P:
    case IMGFMT_422P:
    case IMGFMT_411P:
    case IMGFMT_HM12:
        return vf_next_query_format(vf,fmt);
    }
    return 0;
}

static int config(struct vf_instance *vf,
                  int width, int height, int d_width, int d_height,
                  unsigned int flags, unsigned int outfmt)
{
    free(vf->priv->buf);
    vf->priv->buf = av_mallocz((((width+15)&~15)*(vf->priv->radius+1)/2+32)*sizeof(uint16_t));
    return vf_next_config(vf,width,height,d_width,d_height,flags,outfmt);
}

static void uninit(struct vf_instance *vf)
{
    if (!vf->priv) return;
    av_free(vf->priv->buf);
    free(vf->priv);
    vf->priv = NULL;
}

static int vf_open(vf_instance_t *vf, char *args)
{
    float thresh = 1.2;
    int radius = 16;

    vf->get_image=get_image;
    vf->put_image=put_image;
    vf->query_format=query_format;
    vf->config=config;
    vf->uninit=uninit;
    vf->priv=malloc(sizeof(struct vf_priv_s));
    memset(vf->priv, 0, sizeof(struct vf_priv_s));

    if (args) sscanf(args, "%f:%d", &thresh, &radius);
    vf->priv->thresh = (1<<15)/av_clipf(thresh,0.51,255);
    vf->priv->radius = av_clip((radius+1)&~1,4,32);

    vf->priv->blur_line = blur_line_c;
    vf->priv->filter_line = filter_line_c;
#if HAVE_SSSE3
#if HAVE_6REGS
    if (gCpuCaps.hasSSE2)
        vf->priv->blur_line = blur_line_sse2;
#endif
    if (gCpuCaps.hasMMX2)
        vf->priv->filter_line = filter_line_mmx2;
    if (gCpuCaps.hasSSSE3)
        vf->priv->filter_line = filter_line_ssse3;
#endif

    return 1;
}

const vf_info_t vf_info_gradfun = {
    "gradient deband",
    "gradfun",
    "Loren Merritt",
    "",
    vf_open,
    NULL
};