FFmpeg
vf_sr.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2018 Sergey Lavrushkin
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  * Filter implementing image super-resolution using deep convolutional networks.
24  * https://arxiv.org/abs/1501.00092
25  * https://arxiv.org/abs/1609.05158
26  */
27 
28 #include "avfilter.h"
29 #include "formats.h"
30 #include "internal.h"
31 #include "libavutil/opt.h"
32 #include "libavutil/pixdesc.h"
33 #include "libavformat/avio.h"
34 #include "libswscale/swscale.h"
35 #include "dnn_interface.h"
36 
37 typedef struct SRContext {
38  const AVClass *class;
39 
49 } SRContext;
50 
51 #define OFFSET(x) offsetof(SRContext, x)
52 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM | AV_OPT_FLAG_VIDEO_PARAM
53 static const AVOption sr_options[] = {
54  { "dnn_backend", "DNN backend used for model execution", OFFSET(backend_type), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, FLAGS, "backend" },
55  { "native", "native backend flag", 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, 0, 0, FLAGS, "backend" },
56 #if (CONFIG_LIBTENSORFLOW == 1)
57  { "tensorflow", "tensorflow backend flag", 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, 0, 0, FLAGS, "backend" },
58 #endif
59  { "scale_factor", "scale factor for SRCNN model", OFFSET(scale_factor), AV_OPT_TYPE_INT, { .i64 = 2 }, 2, 4, FLAGS },
60  { "model", "path to model file specifying network architecture and its parameters", OFFSET(model_filename), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
61  { NULL }
62 };
63 
65 
67 {
68  SRContext *sr_context = context->priv;
69 
70  sr_context->dnn_module = ff_get_dnn_module(sr_context->backend_type);
71  if (!sr_context->dnn_module){
72  av_log(context, AV_LOG_ERROR, "could not create DNN module for requested backend\n");
73  return AVERROR(ENOMEM);
74  }
75 
76  if (!sr_context->model_filename){
77  av_log(context, AV_LOG_ERROR, "model file for network was not specified\n");
78  return AVERROR(EIO);
79  }
80  if (!sr_context->dnn_module->load_model) {
81  av_log(context, AV_LOG_ERROR, "load_model for network was not specified\n");
82  return AVERROR(EIO);
83  }
84  sr_context->model = (sr_context->dnn_module->load_model)(sr_context->model_filename, NULL);
85  if (!sr_context->model){
86  av_log(context, AV_LOG_ERROR, "could not load DNN model\n");
87  return AVERROR(EIO);
88  }
89 
90  sr_context->input.dt = DNN_FLOAT;
91  sr_context->sws_contexts[0] = NULL;
92  sr_context->sws_contexts[1] = NULL;
93  sr_context->sws_contexts[2] = NULL;
94 
95  return 0;
96 }
97 
99 {
103  AVFilterFormats *formats_list;
104 
105  formats_list = ff_make_format_list(pixel_formats);
106  if (!formats_list){
107  av_log(context, AV_LOG_ERROR, "could not create formats list\n");
108  return AVERROR(ENOMEM);
109  }
110 
111  return ff_set_common_formats(context, formats_list);
112 }
113 
115 {
116  AVFilterContext *context = inlink->dst;
117  SRContext *sr_context = context->priv;
118  AVFilterLink *outlink = context->outputs[0];
120  int sws_src_h, sws_src_w, sws_dst_h, sws_dst_w;
121  const char *model_output_name = "y";
122 
123  sr_context->input.width = inlink->w * sr_context->scale_factor;
124  sr_context->input.height = inlink->h * sr_context->scale_factor;
125  sr_context->input.channels = 1;
126 
127  result = (sr_context->model->set_input)(sr_context->model->model, &sr_context->input, "x");
128  if (result != DNN_SUCCESS){
129  av_log(context, AV_LOG_ERROR, "could not set input and output for the model\n");
130  return AVERROR(EIO);
131  }
132 
133  result = (sr_context->dnn_module->execute_model)(sr_context->model, &sr_context->output, &model_output_name, 1);
134  if (result != DNN_SUCCESS){
135  av_log(context, AV_LOG_ERROR, "failed to execute loaded model\n");
136  return AVERROR(EIO);
137  }
138 
139  if (sr_context->input.height != sr_context->output.height || sr_context->input.width != sr_context->output.width){
140  sr_context->input.width = inlink->w;
141  sr_context->input.height = inlink->h;
142  result = (sr_context->model->set_input)(sr_context->model->model, &sr_context->input, "x");
143  if (result != DNN_SUCCESS){
144  av_log(context, AV_LOG_ERROR, "could not set input and output for the model\n");
145  return AVERROR(EIO);
146  }
147  result = (sr_context->dnn_module->execute_model)(sr_context->model, &sr_context->output, &model_output_name, 1);
148  if (result != DNN_SUCCESS){
149  av_log(context, AV_LOG_ERROR, "failed to execute loaded model\n");
150  return AVERROR(EIO);
151  }
152  sr_context->scale_factor = 0;
153  }
154  outlink->h = sr_context->output.height;
155  outlink->w = sr_context->output.width;
156  sr_context->sws_contexts[1] = sws_getContext(sr_context->input.width, sr_context->input.height, AV_PIX_FMT_GRAY8,
157  sr_context->input.width, sr_context->input.height, AV_PIX_FMT_GRAYF32,
158  0, NULL, NULL, NULL);
159  sr_context->sws_input_linesize = sr_context->input.width << 2;
160  sr_context->sws_contexts[2] = sws_getContext(sr_context->output.width, sr_context->output.height, AV_PIX_FMT_GRAYF32,
161  sr_context->output.width, sr_context->output.height, AV_PIX_FMT_GRAY8,
162  0, NULL, NULL, NULL);
163  sr_context->sws_output_linesize = sr_context->output.width << 2;
164  if (!sr_context->sws_contexts[1] || !sr_context->sws_contexts[2]){
165  av_log(context, AV_LOG_ERROR, "could not create SwsContext for conversions\n");
166  return AVERROR(ENOMEM);
167  }
168  if (sr_context->scale_factor){
169  sr_context->sws_contexts[0] = sws_getContext(inlink->w, inlink->h, inlink->format,
170  outlink->w, outlink->h, outlink->format,
172  if (!sr_context->sws_contexts[0]){
173  av_log(context, AV_LOG_ERROR, "could not create SwsContext for scaling\n");
174  return AVERROR(ENOMEM);
175  }
176  sr_context->sws_slice_h = inlink->h;
177  } else {
178  if (inlink->format != AV_PIX_FMT_GRAY8){
180  sws_src_h = AV_CEIL_RSHIFT(sr_context->input.height, desc->log2_chroma_h);
181  sws_src_w = AV_CEIL_RSHIFT(sr_context->input.width, desc->log2_chroma_w);
182  sws_dst_h = AV_CEIL_RSHIFT(sr_context->output.height, desc->log2_chroma_h);
183  sws_dst_w = AV_CEIL_RSHIFT(sr_context->output.width, desc->log2_chroma_w);
184 
185  sr_context->sws_contexts[0] = sws_getContext(sws_src_w, sws_src_h, AV_PIX_FMT_GRAY8,
186  sws_dst_w, sws_dst_h, AV_PIX_FMT_GRAY8,
188  if (!sr_context->sws_contexts[0]){
189  av_log(context, AV_LOG_ERROR, "could not create SwsContext for scaling\n");
190  return AVERROR(ENOMEM);
191  }
192  sr_context->sws_slice_h = sws_src_h;
193  }
194  }
195 
196  return 0;
197 }
198 
200 {
201  AVFilterContext *context = inlink->dst;
202  SRContext *sr_context = context->priv;
203  AVFilterLink *outlink = context->outputs[0];
204  AVFrame *out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
205  DNNReturnType dnn_result;
206  const char *model_output_name = "y";
207 
208  if (!out){
209  av_log(context, AV_LOG_ERROR, "could not allocate memory for output frame\n");
210  av_frame_free(&in);
211  return AVERROR(ENOMEM);
212  }
213  av_frame_copy_props(out, in);
214  out->height = sr_context->output.height;
215  out->width = sr_context->output.width;
216  if (sr_context->scale_factor){
217  sws_scale(sr_context->sws_contexts[0], (const uint8_t **)in->data, in->linesize,
218  0, sr_context->sws_slice_h, out->data, out->linesize);
219 
220  sws_scale(sr_context->sws_contexts[1], (const uint8_t **)out->data, out->linesize,
221  0, out->height, (uint8_t * const*)(&sr_context->input.data),
222  (const int [4]){sr_context->sws_input_linesize, 0, 0, 0});
223  } else {
224  if (sr_context->sws_contexts[0]){
225  sws_scale(sr_context->sws_contexts[0], (const uint8_t **)(in->data + 1), in->linesize + 1,
226  0, sr_context->sws_slice_h, out->data + 1, out->linesize + 1);
227  sws_scale(sr_context->sws_contexts[0], (const uint8_t **)(in->data + 2), in->linesize + 2,
228  0, sr_context->sws_slice_h, out->data + 2, out->linesize + 2);
229  }
230 
231  sws_scale(sr_context->sws_contexts[1], (const uint8_t **)in->data, in->linesize,
232  0, in->height, (uint8_t * const*)(&sr_context->input.data),
233  (const int [4]){sr_context->sws_input_linesize, 0, 0, 0});
234  }
235  av_frame_free(&in);
236 
237  dnn_result = (sr_context->dnn_module->execute_model)(sr_context->model, &sr_context->output, &model_output_name, 1);
238  if (dnn_result != DNN_SUCCESS){
239  av_log(context, AV_LOG_ERROR, "failed to execute loaded model\n");
240  return AVERROR(EIO);
241  }
242 
243  sws_scale(sr_context->sws_contexts[2], (const uint8_t *[4]){(const uint8_t *)sr_context->output.data, 0, 0, 0},
244  (const int[4]){sr_context->sws_output_linesize, 0, 0, 0},
245  0, out->height, (uint8_t * const*)out->data, out->linesize);
246 
247  return ff_filter_frame(outlink, out);
248 }
249 
251 {
252  int i;
253  SRContext *sr_context = context->priv;
254 
255  if (sr_context->dnn_module){
256  (sr_context->dnn_module->free_model)(&sr_context->model);
257  av_freep(&sr_context->dnn_module);
258  }
259 
260  for (i = 0; i < 3; ++i){
261  sws_freeContext(sr_context->sws_contexts[i]);
262  }
263 }
264 
265 static const AVFilterPad sr_inputs[] = {
266  {
267  .name = "default",
268  .type = AVMEDIA_TYPE_VIDEO,
269  .config_props = config_props,
270  .filter_frame = filter_frame,
271  },
272  { NULL }
273 };
274 
275 static const AVFilterPad sr_outputs[] = {
276  {
277  .name = "default",
278  .type = AVMEDIA_TYPE_VIDEO,
279  },
280  { NULL }
281 };
282 
284  .name = "sr",
285  .description = NULL_IF_CONFIG_SMALL("Apply DNN-based image super resolution to the input."),
286  .priv_size = sizeof(SRContext),
287  .init = init,
288  .uninit = uninit,
290  .inputs = sr_inputs,
291  .outputs = sr_outputs,
292  .priv_class = &sr_class,
293 };
void * model
Definition: dnn_interface.h:45
#define NULL
Definition: coverity.c:32
Buffered I/O operations.
const AVPixFmtDescriptor * av_pix_fmt_desc_get(enum AVPixelFormat pix_fmt)
Definition: pixdesc.c:2573
This structure describes decoded (raw) audio or video data.
Definition: frame.h:308
#define SWS_BICUBIC
Definition: swscale.h:60
int sws_output_linesize
Definition: vf_sr.c:48
AVOption.
Definition: opt.h:248
DNNReturnType(* execute_model)(const DNNModel *model, DNNData *outputs, const char **output_names, uint32_t nb_output)
Definition: dnn_interface.h:61
void(* free_model)(DNNModel **model)
Definition: dnn_interface.h:63
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
Definition: vf_sr.c:199
const char * desc
Definition: libsvtav1.c:79
planar YUV 4:4:4, 24bpp, (1 Cr & Cb sample per 1x1 Y samples)
Definition: pixfmt.h:71
Main libavfilter public API header.
AVFilter ff_vf_sr
Definition: vf_sr.c:283
DNNData input
Definition: vf_sr.c:44
int channels
Definition: dnn_interface.h:40
static const AVFilterPad sr_inputs[]
Definition: vf_sr.c:265
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
int sws_input_linesize
Definition: vf_sr.c:48
uint8_t log2_chroma_w
Amount to shift the luma width right to find the chroma width.
Definition: pixdesc.h:92
AVFilterFormats * ff_make_format_list(const int *fmts)
Create a list of supported formats.
Definition: formats.c:287
const char * name
Pad name.
Definition: internal.h:60
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
Definition: avfilter.c:1091
uint8_t
#define av_cold
Definition: attributes.h:88
AVOptions.
DNNModel * model
Definition: vf_sr.c:43
DNNModule * dnn_module
Definition: vf_sr.c:42
struct SwsContext * sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat, int dstW, int dstH, enum AVPixelFormat dstFormat, int flags, SwsFilter *srcFilter, SwsFilter *dstFilter, const double *param)
Allocate and return an SwsContext.
Definition: utils.c:1912
DNNBackendType
Definition: dnn_interface.h:33
external API header
static const AVOption sr_options[]
Definition: vf_sr.c:53
#define av_log(a,...)
A filter pad used for either input or output.
Definition: internal.h:54
int width
Definition: frame.h:366
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:194
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:588
uint8_t log2_chroma_h
Amount to shift the luma height right to find the chroma height.
Definition: pixdesc.h:101
int height
Definition: dnn_interface.h:40
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
Definition: frame.c:203
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:153
void * data
Definition: dnn_interface.h:38
void * priv
private data for use by the filter
Definition: avfilter.h:354
static int config_props(AVFilterLink *inlink)
Definition: vf_sr.c:114
DNNData output
Definition: vf_sr.c:45
DNNBackendType backend_type
Definition: vf_sr.c:41
planar YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples)
Definition: pixfmt.h:70
void sws_freeContext(struct SwsContext *swsContext)
Free the swscaler context swsContext.
Definition: utils.c:2332
static int query_formats(AVFilterContext *context)
Definition: vf_sr.c:98
DNN inference engine interface.
DNNReturnType
Definition: dnn_interface.h:31
static av_cold void uninit(AVFilterContext *context)
Definition: vf_sr.c:250
static const AVFilterPad outputs[]
Definition: af_acontrast.c:203
#define AV_PIX_FMT_GRAYF32
Definition: pixfmt.h:431
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
Definition: frame.h:339
Descriptor that unambiguously describes how the bits of a pixel are stored in the up to 4 data planes...
Definition: pixdesc.h:81
int attribute_align_arg sws_scale(struct SwsContext *c, const uint8_t *const srcSlice[], const int srcStride[], int srcSliceY, int srcSliceH, uint8_t *const dst[], const int dstStride[])
swscale wrapper, so we don&#39;t need to export the SwsContext.
Definition: swscale.c:744
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
uint8_t pi<< 24) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi-0x80)*(1.0f/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi-0x80)*(1.0/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S16, int16_t,(*(const int16_t *) pi >> 8)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S16, int16_t,*(const int16_t *) pi *(1.0f/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S16, int16_t,*(const int16_t *) pi *(1.0/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S32, int32_t,(*(const int32_t *) pi >> 24)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S32, int32_t,*(const int32_t *) pi *(1.0f/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S32, int32_t,*(const int32_t *) pi *(1.0/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_FLT, float, av_clip_uint8(lrintf(*(const float *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, float, av_clip_int16(lrintf(*(const float *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, float, av_clipl_int32(llrintf(*(const float *) pi *(1U<< 31)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_DBL, double, av_clip_uint8(lrint(*(const double *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, double, av_clip_int16(lrint(*(const double *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, double, av_clipl_int32(llrint(*(const double *) pi *(1U<< 31))))#define SET_CONV_FUNC_GROUP(ofmt, ifmt) static void set_generic_function(AudioConvert *ac){}void ff_audio_convert_free(AudioConvert **ac){if(!*ac) return;ff_dither_free(&(*ac) ->dc);av_freep(ac);}AudioConvert *ff_audio_convert_alloc(AVAudioResampleContext *avr, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt, int channels, int sample_rate, int apply_map){AudioConvert *ac;int in_planar, out_planar;ac=av_mallocz(sizeof(*ac));if(!ac) return NULL;ac->avr=avr;ac->out_fmt=out_fmt;ac->in_fmt=in_fmt;ac->channels=channels;ac->apply_map=apply_map;if(avr->dither_method!=AV_RESAMPLE_DITHER_NONE &&av_get_packed_sample_fmt(out_fmt)==AV_SAMPLE_FMT_S16 &&av_get_bytes_per_sample(in_fmt) > 2){ac->dc=ff_dither_alloc(avr, out_fmt, in_fmt, channels, sample_rate, apply_map);if(!ac->dc){av_free(ac);return NULL;}return ac;}in_planar=ff_sample_fmt_is_planar(in_fmt, channels);out_planar=ff_sample_fmt_is_planar(out_fmt, channels);if(in_planar==out_planar){ac->func_type=CONV_FUNC_TYPE_FLAT;ac->planes=in_planar?ac->channels:1;}else if(in_planar) ac->func_type=CONV_FUNC_TYPE_INTERLEAVE;else ac->func_type=CONV_FUNC_TYPE_DEINTERLEAVE;set_generic_function(ac);if(ARCH_AARCH64) ff_audio_convert_init_aarch64(ac);if(ARCH_ARM) ff_audio_convert_init_arm(ac);if(ARCH_X86) ff_audio_convert_init_x86(ac);return ac;}int ff_audio_convert(AudioConvert *ac, AudioData *out, AudioData *in){int use_generic=1;int len=in->nb_samples;int p;if(ac->dc){av_log(ac->avr, AV_LOG_TRACE,"%d samples - audio_convert: %s to %s (dithered)\n", len, av_get_sample_fmt_name(ac->in_fmt), av_get_sample_fmt_name(ac->out_fmt));return ff_convert_dither(ac-> in
planar YUV 4:1:0, 9bpp, (1 Cr & Cb sample per 4x4 Y samples)
Definition: pixfmt.h:72
Describe the class of an AVClass context structure.
Definition: log.h:67
Filter definition.
Definition: avfilter.h:145
#define FLAGS
Definition: vf_sr.c:52
const char * name
Filter name.
Definition: avfilter.h:149
AVFilterLink ** outputs
array of pointers to output links
Definition: avfilter.h:351
DNNModule * ff_get_dnn_module(DNNBackendType backend_type)
Definition: dnn_interface.c:32
int scale_factor
Definition: vf_sr.c:46
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:322
static av_cold int init(AVFilterContext *context)
Definition: vf_sr.c:66
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
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
Definition: pixfmt.h:66
Y , 8bpp.
Definition: pixfmt.h:74
planar YUV 4:1:1, 12bpp, (1 Cr & Cb sample per 4x1 Y samples)
Definition: pixfmt.h:73
DNNReturnType(* set_input)(void *model, DNNData *input, const char *input_name)
Definition: dnn_interface.h:53
struct SwsContext * sws_contexts[3]
Definition: vf_sr.c:47
#define OFFSET(x)
Definition: vf_sr.c:51
A list of supported formats for one end of a filter link.
Definition: formats.h:65
static const AVFilterPad sr_outputs[]
Definition: vf_sr.c:275
An instance of a filter.
Definition: avfilter.h:339
and forward the result(frame or status change) to the corresponding input.If nothing is possible
int height
Definition: frame.h:366
char * model_filename
Definition: vf_sr.c:40
FILE * out
Definition: movenc.c:54
int sws_slice_h
Definition: vf_sr.c:48
#define av_freep(p)
DNNModel *(* load_model)(const char *model_filename, const char *options)
Definition: dnn_interface.h:59
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
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default minimum maximum flags name is the option keep it simple and lowercase description are in without and describe what they for example set the foo of the bar offset is the offset of the field in your local context
AVFILTER_DEFINE_CLASS(sr)
DNNDataType dt
Definition: dnn_interface.h:39
AVPixelFormat
Pixel format.
Definition: pixfmt.h:64
int av_frame_copy_props(AVFrame *dst, const AVFrame *src)
Copy only "metadata" fields from src to dst.
Definition: frame.c:671
int i
Definition: input.c:407
#define AV_CEIL_RSHIFT(a, b)
Definition: common.h:58