Go to the documentation of this file.
63 #define OFFSET(x) offsetof(LIBVMAFContext, x)
64 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
69 {
"log_fmt",
"Set the format of the log (csv, json, xml, or sub).",
OFFSET(log_fmt),
AV_OPT_TYPE_STRING, {.str=
"xml"}, 0, 1,
FLAGS},
76 {
"n_threads",
"Set number of threads to be used when computing vmaf.",
OFFSET(n_threads),
AV_OPT_TYPE_INT, {.i64=0}, 0, UINT_MAX,
FLAGS},
77 {
"n_subsample",
"Set interval for frame subsampling used when computing vmaf.",
OFFSET(n_subsample),
AV_OPT_TYPE_INT, {.i64=1}, 1, UINT_MAX,
FLAGS},
93 return VMAF_PIX_FMT_YUV420P;
98 return VMAF_PIX_FMT_YUV422P;
103 return VMAF_PIX_FMT_YUV444P;
105 return VMAF_PIX_FMT_UNKNOWN;
116 for (
unsigned i = 0;
i < 3;
i++) {
117 uint8_t *src_data =
src->data[
i];
118 uint8_t *dst_data = dst->data[
i];
119 for (
unsigned j = 0; j < dst->h[
i]; j++) {
120 memcpy(dst_data, src_data,
sizeof(*dst_data) * dst->w[
i]);
121 src_data +=
src->linesize[
i];
122 dst_data += dst->stride[
i];
133 VmafPicture pic_ref, pic_dist;
140 if (
ctx->is_disabled || !
ref)
152 vmaf_picture_unref(&pic_ref);
156 err = vmaf_read_pictures(
s->vmaf, &pic_ref, &pic_dist,
s->frame_cnt++);
169 char *str_copy =
NULL;
170 char *saveptr =
NULL;
178 for (
char *p =
str; *p; p++) {
192 for (
unsigned i = 0;
i < cnt2;
i++) {
206 for (
unsigned i = 0;
i < *cnt;
i++) {
231 "could not parse feature config: %s\n",
s->feature_cfg);
235 for (
unsigned i = 0;
i < dict_cnt;
i++) {
236 char *feature_name =
NULL;
237 VmafFeatureDictionary *feature_opts_dict =
NULL;
242 feature_name = e->
value;
246 err = vmaf_feature_dictionary_set(&feature_opts_dict, e->
key,
250 "could not set feature option: %s.%s=%s\n",
256 err = vmaf_use_feature(
s->vmaf, feature_name, feature_opts_dict);
259 "problem during vmaf_use_feature: %s\n", feature_name);
265 for (
unsigned i = 0;
i < dict_cnt;
i++) {
280 if (!
s->model_cfg)
return 0;
286 "could not parse model config: %s\n",
s->model_cfg);
290 s->model_cnt = dict_cnt;
295 for (
unsigned i = 0;
i < dict_cnt;
i++) {
296 VmafModelConfig model_cfg = { 0 };
304 VMAF_MODEL_FLAG_DISABLE_CLIP : 0;
310 VMAF_MODEL_FLAG_ENABLE_TRANSFORM : 0;
315 model_cfg.name = e->
value;
331 err = vmaf_model_load(&
s->model[
i], &model_cfg,
version);
334 "could not load libvmaf model with version: %s\n",
340 if (path && !
s->model[
i]) {
341 err = vmaf_model_load_from_path(&
s->model[
i], &model_cfg, path);
344 "could not load libvmaf model with path: %s\n",
352 "could not load libvmaf model with config: %s\n",
358 VmafFeatureDictionary *feature_opts_dict =
NULL;
359 char *feature_opt =
NULL;
361 char *feature_name =
av_strtok(e->
key,
".", &feature_opt);
365 err = vmaf_feature_dictionary_set(&feature_opts_dict,
366 feature_opt, e->
value);
369 "could not set feature option: %s.%s=%s\n",
370 feature_name, feature_opt, e->
value);
375 err = vmaf_model_feature_overload(
s->model[
i], feature_name,
379 "could not overload feature: %s\n", feature_name);
386 for (
unsigned i = 0;
i <
s->model_cnt;
i++) {
387 err = vmaf_use_features_from_model(
s->vmaf,
s->model[
i]);
390 "problem during vmaf_use_features_from_model\n");
397 for (
unsigned i = 0;
i < dict_cnt;
i++) {
409 return VMAF_LOG_LEVEL_NONE;
411 return VMAF_LOG_LEVEL_ERROR;
413 return VMAF_LOG_LEVEL_WARNING;
415 return VMAF_LOG_LEVEL_INFO;
417 return VMAF_LOG_LEVEL_DEBUG;
419 return VMAF_LOG_LEVEL_INFO;
426 VmafModel *model =
NULL;
427 VmafModelCollection *model_collection =
NULL;
428 enum VmafModelFlags
flags = VMAF_MODEL_FLAGS_DEFAULT;
431 VmafModelConfig model_cfg = {
436 if (
s->enable_transform ||
s->phone_model)
437 flags |= VMAF_MODEL_FLAG_ENABLE_TRANSFORM;
440 goto extra_metrics_only;
442 if (
s->enable_conf_interval) {
443 err = vmaf_model_collection_load_from_path(&model, &model_collection,
444 &model_cfg,
s->model_path);
447 "problem loading model file: %s\n",
s->model_path);
451 err = vmaf_use_features_from_model_collection(
s->vmaf, model_collection);
454 "problem loading feature extractors from model file: %s\n",
459 err = vmaf_model_load_from_path(&model, &model_cfg,
s->model_path);
462 "problem loading model file: %s\n",
s->model_path);
465 err = vmaf_use_features_from_model(
s->vmaf, model);
468 "problem loading feature extractors from model file: %s\n",
476 VmafFeatureDictionary *
d =
NULL;
477 vmaf_feature_dictionary_set(&
d,
"enable_chroma",
"false");
479 err = vmaf_use_feature(
s->vmaf,
"psnr",
d);
482 "problem loading feature extractor: psnr\n");
488 err = vmaf_use_feature(
s->vmaf,
"float_ssim",
NULL);
491 "problem loading feature extractor: ssim\n");
497 err = vmaf_use_feature(
s->vmaf,
"float_ms_ssim",
NULL);
500 "problem loading feature extractor: ms_ssim\n");
514 VmafConfiguration cfg = {
516 .n_subsample =
s->n_subsample,
517 .n_threads =
s->n_threads,
520 err = vmaf_init(&
s->vmaf, cfg);
553 if (
ctx->inputs[0]->w !=
ctx->inputs[1]->w) {
558 if (
ctx->inputs[0]->h !=
ctx->inputs[1]->h) {
563 if (
ctx->inputs[0]->format !=
ctx->inputs[1]->format) {
572 s->bpc =
desc->comp[0].depth;
587 outlink->
w = mainlink->
w;
588 outlink->
h = mainlink->
h;
608 return VMAF_OUTPUT_FORMAT_XML;
610 return VMAF_OUTPUT_FORMAT_JSON;
612 return VMAF_OUTPUT_FORMAT_CSV;
614 return VMAF_OUTPUT_FORMAT_SUB;
617 return VMAF_OUTPUT_FORMAT_XML;
624 return VMAF_POOL_METHOD_MIN;
626 return VMAF_POOL_METHOD_MEAN;
628 return VMAF_POOL_METHOD_HARMONIC_MEAN;
631 return VMAF_POOL_METHOD_MEAN;
644 err = vmaf_read_pictures(
s->vmaf,
NULL,
NULL, 0);
647 "problem flushing libvmaf context.\n");
650 for (
unsigned i = 0;
i <
s->model_cnt;
i++) {
653 &vmaf_score, 0,
s->frame_cnt - 1);
656 "problem getting pooled vmaf score.\n");
663 if (
s->log_path && !err)
669 for (
unsigned i = 0;
i <
s->model_cnt;
i++) {
671 vmaf_model_destroy(
s->model[
i]);
702 .preinit = libvmaf_framesync_preinit,
707 .priv_class = &libvmaf_class,
int ff_framesync_configure(FFFrameSync *fs)
Configure a frame sync structure.
#define AV_LOG_WARNING
Something somehow does not look correct.
AVPixelFormat
Pixel format.
static double psnr(double d)
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
static const AVFilterPad libvmaf_inputs[]
void ff_framesync_uninit(FFFrameSync *fs)
Free all memory currently allocated.
#define AV_LOG_QUIET
Print no output.
char * av_stristr(const char *s1, const char *s2)
Locate the first case-independent occurrence in the string haystack of the string needle.
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
const AVPixFmtDescriptor * av_pix_fmt_desc_get(enum AVPixelFormat pix_fmt)
#define FILTER_PIXFMTS_ARRAY(array)
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
This structure describes decoded (raw) audio or video data.
static enum VmafOutputFormat log_fmt_map(const char *log_fmt)
#define AV_DICT_IGNORE_SUFFIX
Return first entry in a dictionary whose first part corresponds to the search key,...
@ AV_PIX_FMT_YUV420P16LE
planar YUV 4:2:0, 24bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian
const char * name
Filter name.
A link between two filters.
@ AV_PIX_FMT_YUV444P16LE
planar YUV 4:4:4, 48bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian
@ AV_PIX_FMT_YUV420P12LE
planar YUV 4:2:0,18bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian
A filter pad used for either input or output.
@ AV_PIX_FMT_YUV420P10LE
planar YUV 4:2:0, 15bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian
@ AV_PIX_FMT_YUV444P12LE
planar YUV 4:4:4,36bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
AVDictionaryEntry * av_dict_get(const AVDictionary *m, const char *key, const AVDictionaryEntry *prev, int flags)
Get a dictionary entry with matching key.
const AVFilter ff_vf_libvmaf
static int config_input_ref(AVFilterLink *inlink)
char * av_strtok(char *s, const char *delim, char **saveptr)
Split the string into several tokens which can be accessed by successive calls to av_strtok().
AVRational sample_aspect_ratio
agreed upon sample aspect ratio
static int copy_picture_data(AVFrame *src, VmafPicture *dst, unsigned bpc)
AVRational frame_rate
Frame rate of the stream on the link, or 1/0 if unknown or variable; if left to 0/0,...
#define AV_LOG_DEBUG
Stuff which is only useful for libav* developers.
@ AV_PIX_FMT_YUV420P
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
FRAMESYNC_DEFINE_CLASS(libvmaf, LIBVMAFContext, fs)
#define FILTER_INPUTS(array)
@ AV_PIX_FMT_YUV444P10LE
planar YUV 4:4:4, 30bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian
static int activate(AVFilterContext *ctx)
int av_log_get_level(void)
Get the current log level.
@ AV_PIX_FMT_YUV422P16LE
planar YUV 4:2:2, 32bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian
Describe the class of an AVClass context structure.
static int do_vmaf(FFFrameSync *fs)
#define fs(width, name, subs,...)
static av_cold void uninit(AVFilterContext *ctx)
static enum AVPixelFormat pix_fmts[]
static enum VmafPoolingMethod pool_method_map(const char *pool_method)
static int config_output(AVFilterLink *outlink)
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification.
int ff_framesync_init_dualinput(FFFrameSync *fs, AVFilterContext *parent)
Initialize a frame sync structure for dualinput.
@ AV_PIX_FMT_YUV422P10LE
planar YUV 4:2:2, 20bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian
AVFilterContext * src
source filter
void av_dict_free(AVDictionary **pm)
Free all the memory allocated for an AVDictionary struct and all keys and values.
#define AV_LOG_INFO
Standard information.
static AVDictionary ** delimited_dict_parse(char *str, unsigned *cnt)
#define i(width, name, range_min, range_max)
int w
agreed upon image width
const char * name
Pad name.
void * av_calloc(size_t nmemb, size_t size)
static const AVFilterPad libvmaf_outputs[]
static enum VmafPixelFormat pix_fmt_map(enum AVPixelFormat av_pix_fmt)
int h
agreed upon image height
static const AVOption libvmaf_options[]
int av_dict_parse_string(AVDictionary **pm, const char *str, const char *key_val_sep, const char *pairs_sep, int flags)
Parse the key/value pairs list and add the parsed entries to a dictionary.
static int ref[MAX_W *MAX_W]
AVRational time_base
Define the time base used by the PTS of the frames/samples which will pass through this link.
@ AV_PIX_FMT_YUV444P
planar YUV 4:4:4, 24bpp, (1 Cr & Cb sample per 1x1 Y samples)
char * av_strdup(const char *s)
Duplicate a string.
@ AV_PIX_FMT_YUV422P
planar YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples)
Descriptor that unambiguously describes how the bits of a pixel are stored in the up to 4 data planes...
static av_cold int init(AVFilterContext *ctx)
#define FILTER_OUTPUTS(array)
static enum VmafLogLevel log_level_map(int log_level)
#define flags(name, subs,...)
static int parse_models(AVFilterContext *ctx)
int ff_framesync_activate(FFFrameSync *fs)
Examine the frames in the filter's input and try to produce output.
int ff_framesync_dualinput_get(FFFrameSync *fs, AVFrame **f0, AVFrame **f1)
static int parse_features(AVFilterContext *ctx)
@ AV_PIX_FMT_YUV422P12LE
planar YUV 4:2:2,24bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian
static int parse_deprecated_options(AVFilterContext *ctx)
#define AV_OPT_FLAG_DEPRECATED
set if option is deprecated, users should refer to AVOption.help text for more information