lavfi: port mp=fspp to a native libavfilter filter

Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

LICENSE.md

@@ -31,6 +31,7 @@ Specifically, the GPL parts of FFmpeg are:
     - vf_cropdetect.c
     - vf_decimate.c
     - vf_delogo.c
+    - vf_fspp.c
     - vf_geq.c
     - vf_histeq.c
     - vf_hqdn3d.c

configure

@@ -2582,6 +2582,7 @@ ebur128_filter_deps="gpl"
 flite_filter_deps="libflite"
 frei0r_filter_deps="frei0r dlopen"
 frei0r_src_filter_deps="frei0r dlopen"
+fspp_filter_deps="gpl"
 geq_filter_deps="gpl"
 histeq_filter_deps="gpl"
 hqdn3d_filter_deps="gpl"

doc/filters.texi

@@ -5066,6 +5066,37 @@ frei0r=perspective:0.2/0.2|0.8/0.2
 For more information, see
 @url{http://frei0r.dyne.org}
 
+@section fspp
+
+Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
+
+It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
+processing filter, one of them is performed once per block, not per pixel.
+This allows for much higher speed.
+
+The filter accepts the following options:
+
+@table @option
+@item quality
+Set quality. This option defines the number of levels for averaging. It accepts
+an integer in the range 4-5. Default value is @code{4}.
+
+@item qp
+Force a constant quantization parameter. It accepts an integer in range 0-63.
+If not set, the filter will use the QP from the video stream (if available).
+
+@item strength
+Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
+more details but also more artifacts, while higher values make the image smoother
+but also blurrier. Default value is @code{0} − PSNR optimal.
+
+@item use_bframe_qp
+Enable the use of the QP from the B-Frames if set to @code{1}. Using this
+option may cause flicker since the B-Frames have often larger QP. Default is
+@code{0} (not enabled).
+
+@end table
+
 @section geq
 
 The filter accepts the following options:
@@ -8361,6 +8392,7 @@ stereo3d=abl:sbsr
 @end example
 @end itemize
 
+@anchor{spp}
 @section spp
 
 Apply a simple postprocessing filter that compresses and decompresses the image

libavfilter/Makefile

@@ -126,6 +126,7 @@ OBJS-$(CONFIG_FRAMESTEP_FILTER)              += vf_framestep.o
 OBJS-$(CONFIG_FPS_FILTER)                    += vf_fps.o
 OBJS-$(CONFIG_FRAMEPACK_FILTER)              += vf_framepack.o
 OBJS-$(CONFIG_FREI0R_FILTER)                 += vf_frei0r.o
+OBJS-$(CONFIG_FSPP_FILTER)                   += vf_fspp.o
 OBJS-$(CONFIG_GEQ_FILTER)                    += vf_geq.o
 OBJS-$(CONFIG_GRADFUN_FILTER)                += vf_gradfun.o
 OBJS-$(CONFIG_HALDCLUT_FILTER)               += vf_lut3d.o dualinput.o framesync.o

libavfilter/allfilters.c

@@ -142,6 +142,7 @@ void avfilter_register_all(void)
     REGISTER_FILTER(FRAMEPACK,      framepack,      vf);
     REGISTER_FILTER(FRAMESTEP,      framestep,      vf);
     REGISTER_FILTER(FREI0R,         frei0r,         vf);
+    REGISTER_FILTER(FSPP,           fspp,           vf);
     REGISTER_FILTER(GEQ,            geq,            vf);
     REGISTER_FILTER(GRADFUN,        gradfun,        vf);
     REGISTER_FILTER(HALDCLUT,       haldclut,       vf);

libavfilter/libmpcodecs/vf_fspp.c

@@ -710,8 +710,8 @@ const vf_info_t ff_vf_info_fspp = {
 #if HAVE_MMX_INLINE
 
 DECLARE_ASM_CONST(8, uint64_t, MM_FIX_0_382683433)=FIX64(0.382683433, 14);
-DECLARE_ALIGNED(8, uint64_t, ff_MM_FIX_0_541196100)=FIX64(0.541196100, 14);
-DECLARE_ALIGNED(8, uint64_t, ff_MM_FIX_0_707106781)=FIX64(0.707106781, 14);
+extern uint64_t ff_MM_FIX_0_707106781;
+extern uint64_t ff_MM_FIX_0_541196100;
 DECLARE_ASM_CONST(8, uint64_t, MM_FIX_1_306562965)=FIX64(1.306562965, 14);
 
 DECLARE_ASM_CONST(8, uint64_t, MM_FIX_1_414213562_A)=FIX64(1.414213562, 14);

libavfilter/version.h

@@ -30,7 +30,7 @@
 #include "libavutil/version.h"
 
 #define LIBAVFILTER_VERSION_MAJOR  5
-#define LIBAVFILTER_VERSION_MINOR  4
+#define LIBAVFILTER_VERSION_MINOR  5
 #define LIBAVFILTER_VERSION_MICRO 100
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \

libavfilter/vf_fspp.h

+/*
+ * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2005 Nikolaj Poroshin <porosh3@psu.ru>
+ * Copyright (c) 2014 Arwa Arif <arwaarif1994@gmail.com>
+ *
+ * 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_FSPP_H
+#define AVFILTER_FSPP_H
+
+#include "avfilter.h"
+
+#define BLOCKSZ 12
+#define MAX_LEVEL 5
+
+#define DCTSIZE 8
+#define DCTSIZE_S "8"
+
+#define FIX(x,s)  ((int) ((x) * (1 << s) + 0.5) & 0xffff)
+#define C64(x)    ((uint64_t)((x) | (x) << 16)) <<32 | (uint64_t)(x) | (uint64_t)(x) << 16
+#define FIX64(x,s)  C64(FIX(x,s))
+
+#define MULTIPLY16H(x,k)   (((x) * (k)) >> 16)
+#define THRESHOLD(r,x,t)                         \
+    if(((unsigned)((x) + t)) > t * 2) r = (x);   \
+    else r = 0;
+#define DESCALE(x,n)  (((x) + (1 << ((n) - 1))) >> n)
+
+typedef int32_t int_simd16_t;
+static const int16_t FIX_0_382683433   = FIX(0.382683433, 14);
+static const int16_t FIX_0_541196100   = FIX(0.541196100, 14);
+static const int16_t FIX_0_707106781   = FIX(0.707106781, 14);
+static const int16_t FIX_1_306562965   = FIX(1.306562965, 14);
+static const int16_t FIX_1_414213562_A = FIX(1.414213562, 14);
+static const int16_t FIX_1_847759065   = FIX(1.847759065, 13);
+static const int16_t FIX_2_613125930   = FIX(-2.613125930, 13);
+static const int16_t FIX_1_414213562   = FIX(1.414213562, 13);
+static const int16_t FIX_1_082392200   = FIX(1.082392200, 13);
+
+typedef struct FSPPContext {
+    AVClass *class;
+    uint64_t threshold_mtx_noq[8 * 2];
+    uint64_t threshold_mtx[8 * 2];        //used in both C & MMX (& later SSE2) versions
+
+    int log2_count;
+    int strength;
+    int hsub;
+    int vsub;
+    int temp_stride;
+    int qp;
+    int qscale_type;
+    int prev_q;
+    uint8_t *src;
+    int16_t *temp;
+    uint8_t *non_b_qp_table;
+    int non_b_qp_alloc_size;
+    int use_bframe_qp;
+
+    void (*store_slice)(uint8_t *dst, int16_t *src,
+                        ptrdiff_t dst_stride, ptrdiff_t src_stride,
+                        ptrdiff_t width, ptrdiff_t height, ptrdiff_t log2_scale);
+
+    void (*store_slice2)(uint8_t *dst, int16_t *src,
+                         ptrdiff_t dst_stride, ptrdiff_t src_stride,
+                         ptrdiff_t width, ptrdiff_t height, ptrdiff_t log2_scale);
+
+    void (*mul_thrmat)(struct FSPPContext *fspp, int q);
+
+    void (*column_fidct)(int16_t *thr_adr, int16_t *data,
+                         int16_t *output, int cnt);
+
+    void (*row_idct)(int16_t *workspace, int16_t *output_adr,
+                     int output_stride, int cnt);
+
+    void (*row_fdct)(int16_t *data, const uint8_t *pixels,
+                     int line_size, int cnt);
+
+} FSPPContext;
+
+void ff_fspp_init_x86(FSPPContext *fspp);
+
+#endif /* AVFILTER_FSPP_H */

libavfilter/x86/Makefile

+OBJS-$(CONFIG_FSPP_FILTER)                   += x86/vf_fspp.o
 OBJS-$(CONFIG_GRADFUN_FILTER)                += x86/vf_gradfun_init.o
 OBJS-$(CONFIG_HQDN3D_FILTER)                 += x86/vf_hqdn3d_init.o
 OBJS-$(CONFIG_IDET_FILTER)                   += x86/vf_idet_init.o