FFmpeg
vf_elbg.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2013 Stefano Sabatini
3  *
4  * This file is part of FFmpeg.
5  *
6  * FFmpeg is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU Lesser General Public
8  * License as published by the Free Software Foundation; either
9  * version 2.1 of the License, or (at your option) any later version.
10  *
11  * FFmpeg is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14  * Lesser General Public License for more details.
15  *
16  * You should have received a copy of the GNU Lesser General Public
17  * License along with FFmpeg; if not, write to the Free Software
18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19  */
20 
21 /**
22  * @file
23  * video quantizer filter based on ELBG
24  */
25 
26 #include "libavcodec/elbg.h"
27 #include "libavutil/opt.h"
28 #include "libavutil/pixdesc.h"
29 #include "libavutil/random_seed.h"
30 
31 #include "avfilter.h"
32 #include "drawutils.h"
33 #include "internal.h"
34 #include "video.h"
35 
36 typedef struct ELBGContext {
37  const AVClass *class;
39  unsigned int lfg_seed;
41  int *codeword;
44  int *codebook;
48  int pal8;
49 } ELBGContext;
50 
51 #define OFFSET(x) offsetof(ELBGContext, x)
52 #define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
53 
54 static const AVOption elbg_options[] = {
55  { "codebook_length", "set codebook length", OFFSET(codebook_length), AV_OPT_TYPE_INT, { .i64 = 256 }, 1, INT_MAX, FLAGS },
56  { "l", "set codebook length", OFFSET(codebook_length), AV_OPT_TYPE_INT, { .i64 = 256 }, 1, INT_MAX, FLAGS },
57  { "nb_steps", "set max number of steps used to compute the mapping", OFFSET(max_steps_nb), AV_OPT_TYPE_INT, { .i64 = 1 }, 1, INT_MAX, FLAGS },
58  { "n", "set max number of steps used to compute the mapping", OFFSET(max_steps_nb), AV_OPT_TYPE_INT, { .i64 = 1 }, 1, INT_MAX, FLAGS },
59  { "seed", "set the random seed", OFFSET(lfg_seed), AV_OPT_TYPE_INT, {.i64 = -1}, -1, UINT32_MAX, FLAGS },
60  { "s", "set the random seed", OFFSET(lfg_seed), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, UINT32_MAX, FLAGS },
61  { "pal8", "set the pal8 output", OFFSET(pal8), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, FLAGS },
62  { NULL }
63 };
64 
66 
68 {
69  ELBGContext *elbg = ctx->priv;
70 
71  if (elbg->pal8 && elbg->codebook_length > 256) {
72  av_log(ctx, AV_LOG_ERROR, "pal8 output allows max 256 codebook length.\n");
73  return AVERROR(EINVAL);
74  }
75 
76  if (elbg->lfg_seed == -1)
77  elbg->lfg_seed = av_get_random_seed();
78 
79  av_lfg_init(&elbg->lfg, elbg->lfg_seed);
80  return 0;
81 }
82 
84 {
85  ELBGContext *elbg = ctx->priv;
86  int ret;
87 
88  static const enum AVPixelFormat pix_fmts[] = {
92  };
93  if (!elbg->pal8) {
94  AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts);
95  if (!fmts_list)
96  return AVERROR(ENOMEM);
97  return ff_set_common_formats(ctx, fmts_list);
98  } else {
99  static const enum AVPixelFormat pal8_fmt[] = {
101  AV_PIX_FMT_NONE
102  };
103  if ((ret = ff_formats_ref(ff_make_format_list(pix_fmts), &ctx->inputs[0]->out_formats)) < 0 ||
104  (ret = ff_formats_ref(ff_make_format_list(pal8_fmt), &ctx->outputs[0]->in_formats)) < 0)
105  return ret;
106  }
107  return 0;
108 }
109 
110 #define NB_COMPONENTS 3
111 
113 {
114  AVFilterContext *ctx = inlink->dst;
115  ELBGContext *elbg = ctx->priv;
116 
117  elbg->pix_desc = av_pix_fmt_desc_get(inlink->format);
118  elbg->codeword_length = inlink->w * inlink->h;
119  elbg->codeword = av_realloc_f(elbg->codeword, elbg->codeword_length,
120  NB_COMPONENTS * sizeof(*elbg->codeword));
121  if (!elbg->codeword)
122  return AVERROR(ENOMEM);
123 
126  sizeof(*elbg->codeword_closest_codebook_idxs));
128  return AVERROR(ENOMEM);
129 
130  elbg->codebook = av_realloc_f(elbg->codebook, elbg->codebook_length,
131  NB_COMPONENTS * sizeof(*elbg->codebook));
132  if (!elbg->codebook)
133  return AVERROR(ENOMEM);
134 
135  ff_fill_rgba_map(elbg->rgba_map, inlink->format);
136 
137  return 0;
138 }
139 
140 #define R 0
141 #define G 1
142 #define B 2
143 
145 {
146  ELBGContext *elbg = inlink->dst->priv;
147  int i, j, k;
148  uint8_t *p, *p0;
149 
150  const uint8_t r_idx = elbg->rgba_map[R];
151  const uint8_t g_idx = elbg->rgba_map[G];
152  const uint8_t b_idx = elbg->rgba_map[B];
153 
154  /* build the codeword */
155  p0 = frame->data[0];
156  k = 0;
157  for (i = 0; i < inlink->h; i++) {
158  p = p0;
159  for (j = 0; j < inlink->w; j++) {
160  elbg->codeword[k++] = p[r_idx];
161  elbg->codeword[k++] = p[g_idx];
162  elbg->codeword[k++] = p[b_idx];
163  p += elbg->pix_desc->nb_components;
164  }
165  p0 += frame->linesize[0];
166  }
167 
168  /* compute the codebook */
170  elbg->codebook, elbg->codebook_length, elbg->max_steps_nb,
171  elbg->codeword_closest_codebook_idxs, &elbg->lfg);
173  elbg->codebook, elbg->codebook_length, elbg->max_steps_nb,
174  elbg->codeword_closest_codebook_idxs, &elbg->lfg);
175 
176  if (elbg->pal8) {
177  AVFilterLink *outlink = inlink->dst->outputs[0];
178  AVFrame *out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
179  uint32_t *pal;
180 
181  if (!out) {
182  av_frame_free(&frame);
183  return AVERROR(ENOMEM);
184  }
185  out->pts = frame->pts;
186  av_frame_free(&frame);
187  pal = (uint32_t *)out->data[1];
188  p0 = (uint8_t *)out->data[0];
189 
190  for (i = 0; i < elbg->codebook_length; i++) {
191  pal[i] = 0xFFU << 24 |
192  (elbg->codebook[i*3 ] << 16) |
193  (elbg->codebook[i*3+1] << 8) |
194  elbg->codebook[i*3+2];
195  }
196 
197  k = 0;
198  for (i = 0; i < inlink->h; i++) {
199  p = p0;
200  for (j = 0; j < inlink->w; j++, p++) {
201  p[0] = elbg->codeword_closest_codebook_idxs[k++];
202  }
203  p0 += out->linesize[0];
204  }
205 
206  return ff_filter_frame(outlink, out);
207  }
208 
209  /* fill the output with the codebook values */
210  p0 = frame->data[0];
211 
212  k = 0;
213  for (i = 0; i < inlink->h; i++) {
214  p = p0;
215  for (j = 0; j < inlink->w; j++) {
216  int cb_idx = NB_COMPONENTS * elbg->codeword_closest_codebook_idxs[k++];
217  p[r_idx] = elbg->codebook[cb_idx];
218  p[g_idx] = elbg->codebook[cb_idx+1];
219  p[b_idx] = elbg->codebook[cb_idx+2];
220  p += elbg->pix_desc->nb_components;
221  }
222  p0 += frame->linesize[0];
223  }
224 
225  return ff_filter_frame(inlink->dst->outputs[0], frame);
226 }
227 
229 {
230  ELBGContext *elbg = ctx->priv;
231 
232  av_freep(&elbg->codebook);
233  av_freep(&elbg->codeword);
235 }
236 
237 static const AVFilterPad elbg_inputs[] = {
238  {
239  .name = "default",
240  .type = AVMEDIA_TYPE_VIDEO,
241  .config_props = config_input,
242  .filter_frame = filter_frame,
243  .needs_writable = 1,
244  },
245  { NULL }
246 };
247 
248 static const AVFilterPad elbg_outputs[] = {
249  {
250  .name = "default",
251  .type = AVMEDIA_TYPE_VIDEO,
252  },
253  { NULL }
254 };
255 
257  .name = "elbg",
258  .description = NULL_IF_CONFIG_SMALL("Apply posterize effect, using the ELBG algorithm."),
259  .priv_size = sizeof(ELBGContext),
260  .priv_class = &elbg_class,
262  .init = init,
263  .uninit = uninit,
264  .inputs = elbg_inputs,
265  .outputs = elbg_outputs,
266 };
Definition: lfg.h:27
#define NULL
Definition: coverity.c:32
static av_cold void uninit(AVFilterContext *ctx)
Definition: vf_elbg.c:228
AVFilter ff_vf_elbg
Definition: vf_elbg.c:256
const AVPixFmtDescriptor * av_pix_fmt_desc_get(enum AVPixelFormat pix_fmt)
Definition: pixdesc.c:2522
This structure describes decoded (raw) audio or video data.
Definition: frame.h:295
#define av_realloc_f(p, o, n)
AVOption.
Definition: opt.h:246
#define OFFSET(x)
Definition: vf_elbg.c:51
Main libavfilter public API header.
packed RGB 8:8:8, 24bpp, RGBRGB...
Definition: pixfmt.h:68
static int config_input(AVFilterLink *inlink)
Definition: vf_elbg.c:112
#define B
Definition: vf_elbg.c:142
int * codebook
Definition: vf_elbg.c:44
AVFrame * ff_get_video_buffer(AVFilterLink *link, int w, int h)
Request a picture buffer with a specific set of permissions.
Definition: video.c:99
AVFilterFormats * ff_make_format_list(const int *fmts)
Create a list of supported formats.
Definition: formats.c:283
static const AVFilterPad elbg_inputs[]
Definition: vf_elbg.c:237
const char * name
Pad name.
Definition: internal.h:60
AVFilterLink ** inputs
array of pointers to input links
Definition: avfilter.h:346
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
Definition: avfilter.c:1093
uint8_t
#define av_cold
Definition: attributes.h:82
8 bits with AV_PIX_FMT_RGB32 palette
Definition: pixfmt.h:77
AVOptions.
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:388
unsigned int lfg_seed
Definition: vf_elbg.c:39
int avpriv_do_elbg(int *points, int dim, int numpoints, int *codebook, int numCB, int max_steps, int *closest_cb, AVLFG *rand_state)
Implementation of the Enhanced LBG Algorithm Based on the paper "Neural Networks 14:1219-1237" that c...
Definition: elbg.c:371
packed ABGR 8:8:8:8, 32bpp, ABGRABGR...
Definition: pixfmt.h:94
int pal8
Definition: vf_elbg.c:48
#define av_log(a,...)
AVLFG lfg
Definition: vf_elbg.c:38
A filter pad used for either input or output.
Definition: internal.h:54
#define U(x)
Definition: vp56_arith.h:37
int max_steps_nb
Definition: vf_elbg.c:40
#define i(width, name, range_min, range_max)
Definition: cbs_h2645.c:259
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:176
int ff_set_common_formats(AVFilterContext *ctx, AVFilterFormats *formats)
A helper for query_formats() which sets all links to the same list of formats.
Definition: formats.c:569
static int query_formats(AVFilterContext *ctx)
Definition: vf_elbg.c:83
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
Definition: frame.c:202
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:186
packed BGRA 8:8:8:8, 32bpp, BGRABGRA...
Definition: pixfmt.h:95
void * priv
private data for use by the filter
Definition: avfilter.h:353
#define FLAGS
Definition: vf_elbg.c:52
packed ARGB 8:8:8:8, 32bpp, ARGBARGB...
Definition: pixfmt.h:92
packed RGBA 8:8:8:8, 32bpp, RGBARGBA...
Definition: pixfmt.h:93
static const AVOption elbg_options[]
Definition: vf_elbg.c:54
int * codeword
Definition: vf_elbg.c:41
uint8_t nb_components
The number of components each pixel has, (1-4)
Definition: pixdesc.h:83
int ff_formats_ref(AVFilterFormats *f, AVFilterFormats **ref)
Add *ref as a new reference to formats.
Definition: formats.c:439
AVFormatContext * ctx
Definition: movenc.c:48
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return or at least make progress towards producing a frame
packed RGB 8:8:8, 24bpp, BGRBGR...
Definition: pixfmt.h:69
static const AVFilterPad outputs[]
Definition: af_acontrast.c:203
int ff_fill_rgba_map(uint8_t *rgba_map, enum AVPixelFormat pix_fmt)
Definition: drawutils.c:35
static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
Definition: vf_elbg.c:144
static const AVFilterPad elbg_outputs[]
Definition: vf_elbg.c:248
static av_cold int init(AVFilterContext *ctx)
Definition: vf_elbg.c:67
uint8_t rgba_map[4]
Definition: vf_elbg.c:47
int avpriv_init_elbg(int *points, int dim, int numpoints, int *codebook, int numCB, int max_steps, int *closest_cb, AVLFG *rand_state)
Initialize the **codebook vector for the elbg algorithm.
Definition: elbg.c:337
misc drawing utilities
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
Definition: frame.h:326
Descriptor that unambiguously describes how the bits of a pixel are stored in the up to 4 data planes...
Definition: pixdesc.h:81
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several inputs
Describe the class of an AVClass context structure.
Definition: log.h:67
const AVPixFmtDescriptor * pix_desc
Definition: vf_elbg.c:46
Filter definition.
Definition: avfilter.h:144
const char * name
Filter name.
Definition: avfilter.h:148
av_cold void av_lfg_init(AVLFG *c, unsigned int seed)
Definition: lfg.c:32
AVFilterLink ** outputs
array of pointers to output links
Definition: avfilter.h:350
static enum AVPixelFormat pix_fmts[]
Definition: libkvazaar.c:275
#define G
Definition: vf_elbg.c:141
#define NB_COMPONENTS
Definition: vf_elbg.c:110
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:309
The exact code depends on how similar the blocks are and how related they are to the and needs to apply these operations to the correct inlink or outlink if there are several Macros are available to factor that when no extra processing is inlink
int codeword_length
Definition: vf_elbg.c:42
int codebook_length
Definition: vf_elbg.c:45
AVFILTER_DEFINE_CLASS(elbg)
A list of supported formats for one end of a filter link.
Definition: formats.h:64
An instance of a filter.
Definition: avfilter.h:338
FILE * out
Definition: movenc.c:54
#define av_freep(p)
uint32_t av_get_random_seed(void)
Get a seed to use in conjunction with random functions.
Definition: random_seed.c:120
internal API functions
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later.That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another.Frame references ownership and permissions
#define R
Definition: vf_elbg.c:140
AVPixelFormat
Pixel format.
Definition: pixfmt.h:64
int * codeword_closest_codebook_idxs
Definition: vf_elbg.c:43