24#include "common/args.h"
25#include "common/tools_common.h"
26#include "common/video_writer.h"
27#include "examples/encoder_util.h"
28#include "aom_ports/aom_timer.h"
30#define OPTION_BUFFER_SIZE 1024
33 const char *output_filename;
34 char options[OPTION_BUFFER_SIZE];
35 struct AvxInputContext input_ctx;
52static const arg_def_t outputfile =
53 ARG_DEF(
"o",
"output", 1,
"Output filename");
54static const arg_def_t frames_arg =
55 ARG_DEF(
"f",
"frames", 1,
"Number of frames to encode");
56static const arg_def_t threads_arg =
57 ARG_DEF(
"th",
"threads", 1,
"Number of threads to use");
58static const arg_def_t width_arg = ARG_DEF(
"w",
"width", 1,
"Source width");
59static const arg_def_t height_arg = ARG_DEF(
"h",
"height", 1,
"Source height");
60static const arg_def_t timebase_arg =
61 ARG_DEF(
"t",
"timebase", 1,
"Timebase (num/den)");
62static const arg_def_t bitrate_arg = ARG_DEF(
63 "b",
"target-bitrate", 1,
"Encoding bitrate, in kilobits per second");
64static const arg_def_t spatial_layers_arg =
65 ARG_DEF(
"sl",
"spatial-layers", 1,
"Number of spatial SVC layers");
66static const arg_def_t temporal_layers_arg =
67 ARG_DEF(
"tl",
"temporal-layers", 1,
"Number of temporal SVC layers");
68static const arg_def_t layering_mode_arg =
69 ARG_DEF(
"lm",
"layering-mode", 1,
"Temporal layering scheme.");
70static const arg_def_t kf_dist_arg =
71 ARG_DEF(
"k",
"kf-dist", 1,
"Number of frames between keyframes");
72static const arg_def_t scale_factors_arg =
73 ARG_DEF(
"r",
"scale-factors", 1,
"Scale factors (lowest to highest layer)");
74static const arg_def_t min_q_arg =
75 ARG_DEF(NULL,
"min-q", 1,
"Minimum quantizer");
76static const arg_def_t max_q_arg =
77 ARG_DEF(NULL,
"max-q", 1,
"Maximum quantizer");
78static const arg_def_t speed_arg =
79 ARG_DEF(
"sp",
"speed", 1,
"Speed configuration");
80static const arg_def_t aqmode_arg =
81 ARG_DEF(
"aq",
"aqmode", 1,
"AQ mode off/on");
82static const arg_def_t bitrates_arg =
83 ARG_DEF(
"bl",
"bitrates", 1,
84 "Bitrates[spatial_layer * num_temporal_layer + temporal_layer]");
85static const arg_def_t dropframe_thresh_arg =
86 ARG_DEF(NULL,
"drop-frame", 1,
"Temporal resampling threshold (buf %)");
87static const arg_def_t error_resilient_arg =
88 ARG_DEF(NULL,
"error-resilient", 1,
"Error resilient flag");
89static const arg_def_t output_obu_arg =
90 ARG_DEF(NULL,
"output-obu", 1,
91 "Write OBUs when set to 1. Otherwise write IVF files.");
92static const arg_def_t test_decode_arg =
93 ARG_DEF(NULL,
"test-decode", 1,
94 "Attempt to test decoding the output when set to 1. Default is 1.");
95static const struct arg_enum_list tune_content_enum[] = {
96 {
"default", AOM_CONTENT_DEFAULT },
97 {
"screen", AOM_CONTENT_SCREEN },
98 {
"film", AOM_CONTENT_FILM },
101static const arg_def_t tune_content_arg = ARG_DEF_ENUM(
102 NULL,
"tune-content", 1,
"Tune content type", tune_content_enum);
104#if CONFIG_AV1_HIGHBITDEPTH
105static const struct arg_enum_list bitdepth_enum[] = {
109static const arg_def_t bitdepth_arg = ARG_DEF_ENUM(
110 "d",
"bit-depth", 1,
"Bit depth for codec 8, 10 or 12. ", bitdepth_enum);
113static const arg_def_t *svc_args[] = { &frames_arg,
124 &temporal_layers_arg,
128#if CONFIG_AV1_HIGHBITDEPTH
133 &dropframe_thresh_arg,
134 &error_resilient_arg,
140#define zero(Dest) memset(&(Dest), 0, sizeof(Dest))
142static const char *exec_name;
144void usage_exit(
void) {
145 fprintf(stderr,
"Usage: %s <options> input_filename -o output_filename\n",
147 fprintf(stderr,
"Options:\n");
148 arg_show_usage(stderr, svc_args);
152static int file_is_y4m(
const char detect[4]) {
153 return memcmp(detect,
"YUV4", 4) == 0;
156static int fourcc_is_ivf(
const char detect[4]) {
157 if (memcmp(detect,
"DKIF", 4) == 0) {
163static const int option_max_values[ALL_OPTION_TYPES] = { 63, INT_MAX, INT_MAX,
166static const int option_min_values[ALL_OPTION_TYPES] = { 0, 0, 1, 0 };
168static void open_input_file(
struct AvxInputContext *input,
171 input->file = strcmp(input->filename,
"-") ? fopen(input->filename,
"rb")
172 : set_binary_mode(stdin);
174 if (!input->file) fatal(
"Failed to open input file");
176 if (!fseeko(input->file, 0, SEEK_END)) {
180 input->length = ftello(input->file);
185 input->pixel_aspect_ratio.numerator = 1;
186 input->pixel_aspect_ratio.denominator = 1;
191 input->detect.buf_read = fread(input->detect.buf, 1, 4, input->file);
192 input->detect.position = 0;
194 if (input->detect.buf_read == 4 && file_is_y4m(input->detect.buf)) {
195 if (y4m_input_open(&input->y4m, input->file, input->detect.buf, 4, csp,
196 input->only_i420) >= 0) {
197 input->file_type = FILE_TYPE_Y4M;
198 input->width = input->y4m.pic_w;
199 input->height = input->y4m.pic_h;
200 input->pixel_aspect_ratio.numerator = input->y4m.par_n;
201 input->pixel_aspect_ratio.denominator = input->y4m.par_d;
202 input->framerate.numerator = input->y4m.fps_n;
203 input->framerate.denominator = input->y4m.fps_d;
204 input->fmt = input->y4m.aom_fmt;
205 input->bit_depth = input->y4m.bit_depth;
207 fatal(
"Unsupported Y4M stream.");
209 }
else if (input->detect.buf_read == 4 && fourcc_is_ivf(input->detect.buf)) {
210 fatal(
"IVF is not supported as input.");
212 input->file_type = FILE_TYPE_RAW;
216static aom_codec_err_t extract_option(LAYER_OPTION_TYPE type,
char *input,
217 int *value0,
int *value1) {
218 if (type == SCALE_FACTOR) {
219 *value0 = (int)strtol(input, &input, 10);
221 *value1 = (int)strtol(input, &input, 10);
223 if (*value0 < option_min_values[SCALE_FACTOR] ||
224 *value1 < option_min_values[SCALE_FACTOR] ||
225 *value0 > option_max_values[SCALE_FACTOR] ||
226 *value1 > option_max_values[SCALE_FACTOR] ||
230 *value0 = atoi(input);
231 if (*value0 < option_min_values[type] || *value0 > option_max_values[type])
239 int *option0,
int *option1) {
243 const char *delim =
",";
251 if (input == NULL || option0 == NULL ||
252 (option1 == NULL && type == SCALE_FACTOR))
255 input_string = malloc(strlen(input));
256 if (!input_string) die(
"Failed to allocate input string.");
257 memcpy(input_string, input, strlen(input));
259 token = strtok(input_string, delim);
260 for (i = 0; i < num_layers; ++i) {
262 res = extract_option(type, token, option0 + i, option1 + i);
264 token = strtok(NULL, delim);
276static void parse_command_line(
int argc,
const char **argv_,
284 char string_options[1024] = { 0 };
289 app_input->layering_mode = 0;
290 app_input->output_obu = 0;
291 app_input->decode = 1;
296 argv = argv_dup(argc - 1, argv_ + 1);
298 fprintf(stderr,
"Error allocating argument list\n");
301 for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step) {
304 if (arg_match(&arg, &outputfile, argi)) {
305 app_input->output_filename = arg.val;
306 }
else if (arg_match(&arg, &width_arg, argi)) {
307 enc_cfg->
g_w = arg_parse_uint(&arg);
308 }
else if (arg_match(&arg, &height_arg, argi)) {
309 enc_cfg->
g_h = arg_parse_uint(&arg);
310 }
else if (arg_match(&arg, &timebase_arg, argi)) {
311 enc_cfg->
g_timebase = arg_parse_rational(&arg);
312 }
else if (arg_match(&arg, &bitrate_arg, argi)) {
314 }
else if (arg_match(&arg, &spatial_layers_arg, argi)) {
316 }
else if (arg_match(&arg, &temporal_layers_arg, argi)) {
318 }
else if (arg_match(&arg, &speed_arg, argi)) {
319 app_input->speed = arg_parse_uint(&arg);
320 if (app_input->speed > 10) {
321 aom_tools_warn(
"Mapping speed %d to speed 10.\n", app_input->speed);
323 }
else if (arg_match(&arg, &aqmode_arg, argi)) {
324 app_input->aq_mode = arg_parse_uint(&arg);
325 }
else if (arg_match(&arg, &threads_arg, argi)) {
326 enc_cfg->
g_threads = arg_parse_uint(&arg);
327 }
else if (arg_match(&arg, &layering_mode_arg, argi)) {
328 app_input->layering_mode = arg_parse_int(&arg);
329 }
else if (arg_match(&arg, &kf_dist_arg, argi)) {
332 }
else if (arg_match(&arg, &scale_factors_arg, argi)) {
333 parse_layer_options_from_string(svc_params, SCALE_FACTOR, arg.val,
336 }
else if (arg_match(&arg, &min_q_arg, argi)) {
338 }
else if (arg_match(&arg, &max_q_arg, argi)) {
340#if CONFIG_AV1_HIGHBITDEPTH
341 }
else if (arg_match(&arg, &bitdepth_arg, argi)) {
342 enc_cfg->
g_bit_depth = arg_parse_enum_or_int(&arg);
357 die(
"Error: Invalid bit depth selected (%d)\n", enc_cfg->
g_bit_depth);
361 }
else if (arg_match(&arg, &dropframe_thresh_arg, argi)) {
363 }
else if (arg_match(&arg, &error_resilient_arg, argi)) {
366 die(
"Invalid value for error resilient (0, 1): %d.",
368 }
else if (arg_match(&arg, &output_obu_arg, argi)) {
369 app_input->output_obu = arg_parse_uint(&arg);
370 if (app_input->output_obu != 0 && app_input->output_obu != 1)
371 die(
"Invalid value for obu output flag (0, 1): %d.",
372 app_input->output_obu);
373 }
else if (arg_match(&arg, &test_decode_arg, argi)) {
374 app_input->decode = arg_parse_uint(&arg);
375 if (app_input->decode != 0 && app_input->decode != 1)
376 die(
"Invalid value for test decode flag (0, 1): %d.",
378 }
else if (arg_match(&arg, &tune_content_arg, argi)) {
379 app_input->tune_content = arg_parse_enum_or_int(&arg);
380 printf(
"tune content %d\n", app_input->tune_content);
387 for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step) {
389 if (arg_match(&arg, &bitrates_arg, argi)) {
390 parse_layer_options_from_string(svc_params, BITRATE, arg.val,
398 if (strlen(string_options) > 0)
399 strncpy(app_input->options, string_options, OPTION_BUFFER_SIZE);
402 for (argi = argv; *argi; ++argi)
403 if (argi[0][0] ==
'-' && strlen(argi[0]) > 1)
404 die(
"Error: Unrecognized option %s\n", *argi);
406 if (argv[0] == NULL) {
410 app_input->input_ctx.filename = argv[0];
413 open_input_file(&app_input->input_ctx, 0);
414 if (app_input->input_ctx.file_type == FILE_TYPE_Y4M) {
415 enc_cfg->
g_w = app_input->input_ctx.width;
416 enc_cfg->
g_h = app_input->input_ctx.height;
419 if (enc_cfg->
g_w < 16 || enc_cfg->
g_w % 2 || enc_cfg->
g_h < 16 ||
421 die(
"Invalid resolution: %d x %d\n", enc_cfg->
g_w, enc_cfg->
g_h);
426 "width %u, height: %u\n"
427 "num: %d, den: %d, bitrate: %u\n"
435static unsigned int mode_to_num_temporal_layers[11] = { 1, 2, 3, 3, 2, 1,
437static unsigned int mode_to_num_spatial_layers[11] = { 1, 1, 1, 1, 1, 2,
441struct RateControlMetrics {
458 double avg_st_encoding_bitrate;
460 double variance_st_encoding_bitrate;
479static int read_frame(
struct AvxInputContext *input_ctx,
aom_image_t *img) {
480 FILE *f = input_ctx->file;
481 y4m_input *y4m = &input_ctx->y4m;
484 if (input_ctx->file_type == FILE_TYPE_Y4M) {
485 if (y4m_input_fetch_frame(y4m, f, img) < 1)
return 0;
487 shortread = read_yuv_frame(input_ctx, img);
493static void close_input_file(
struct AvxInputContext *input) {
495 if (input->file_type == FILE_TYPE_Y4M) y4m_input_close(&input->y4m);
504static void set_rate_control_metrics(
struct RateControlMetrics *rc,
506 unsigned int ss_number_layers,
507 unsigned int ts_number_layers) {
509 ts_rate_decimator[0] = 1;
510 if (ts_number_layers == 2) {
511 ts_rate_decimator[0] = 2;
512 ts_rate_decimator[1] = 1;
514 if (ts_number_layers == 3) {
515 ts_rate_decimator[0] = 4;
516 ts_rate_decimator[1] = 2;
517 ts_rate_decimator[2] = 1;
521 for (
unsigned int sl = 0; sl < ss_number_layers; ++sl) {
522 unsigned int i = sl * ts_number_layers;
523 rc->layer_framerate[0] = framerate / ts_rate_decimator[0];
525 1000.0 * rc->layer_target_bitrate[i] / rc->layer_framerate[0];
526 for (
unsigned int tl = 0; tl < ts_number_layers; ++tl) {
527 i = sl * ts_number_layers + tl;
529 rc->layer_framerate[tl] = framerate / ts_rate_decimator[tl];
532 (rc->layer_target_bitrate[i] - rc->layer_target_bitrate[i - 1]) /
533 (rc->layer_framerate[tl] - rc->layer_framerate[tl - 1]);
535 rc->layer_input_frames[tl] = 0;
536 rc->layer_enc_frames[tl] = 0;
537 rc->layer_encoding_bitrate[i] = 0.0;
538 rc->layer_avg_frame_size[i] = 0.0;
539 rc->layer_avg_rate_mismatch[i] = 0.0;
542 rc->window_count = 0;
543 rc->window_size = 15;
544 rc->avg_st_encoding_bitrate = 0.0;
545 rc->variance_st_encoding_bitrate = 0.0;
548static void printout_rate_control_summary(
struct RateControlMetrics *rc,
550 unsigned int ss_number_layers,
551 unsigned int ts_number_layers) {
552 int tot_num_frames = 0;
553 double perc_fluctuation = 0.0;
554 printf(
"Total number of processed frames: %d\n\n", frame_cnt - 1);
555 printf(
"Rate control layer stats for %u layer(s):\n\n", ts_number_layers);
556 for (
unsigned int sl = 0; sl < ss_number_layers; ++sl) {
558 for (
unsigned int tl = 0; tl < ts_number_layers; ++tl) {
559 unsigned int i = sl * ts_number_layers + tl;
560 const int num_dropped =
561 tl > 0 ? rc->layer_input_frames[tl] - rc->layer_enc_frames[tl]
562 : rc->layer_input_frames[tl] - rc->layer_enc_frames[tl] - 1;
563 tot_num_frames += rc->layer_input_frames[tl];
564 rc->layer_encoding_bitrate[i] = 0.001 * rc->layer_framerate[tl] *
565 rc->layer_encoding_bitrate[i] /
567 rc->layer_avg_frame_size[i] =
568 rc->layer_avg_frame_size[i] / rc->layer_enc_frames[tl];
569 rc->layer_avg_rate_mismatch[i] =
570 100.0 * rc->layer_avg_rate_mismatch[i] / rc->layer_enc_frames[tl];
571 printf(
"For layer#: %u %u \n", sl, tl);
572 printf(
"Bitrate (target vs actual): %d %f\n", rc->layer_target_bitrate[i],
573 rc->layer_encoding_bitrate[i]);
574 printf(
"Average frame size (target vs actual): %f %f\n", rc->layer_pfb[i],
575 rc->layer_avg_frame_size[i]);
576 printf(
"Average rate_mismatch: %f\n", rc->layer_avg_rate_mismatch[i]);
578 "Number of input frames, encoded (non-key) frames, "
579 "and perc dropped frames: %d %d %f\n",
580 rc->layer_input_frames[tl], rc->layer_enc_frames[tl],
581 100.0 * num_dropped / rc->layer_input_frames[tl]);
585 rc->avg_st_encoding_bitrate = rc->avg_st_encoding_bitrate / rc->window_count;
586 rc->variance_st_encoding_bitrate =
587 rc->variance_st_encoding_bitrate / rc->window_count -
588 (rc->avg_st_encoding_bitrate * rc->avg_st_encoding_bitrate);
589 perc_fluctuation = 100.0 * sqrt(rc->variance_st_encoding_bitrate) /
590 rc->avg_st_encoding_bitrate;
591 printf(
"Short-time stats, for window of %d frames:\n", rc->window_size);
592 printf(
"Average, rms-variance, and percent-fluct: %f %f %f\n",
593 rc->avg_st_encoding_bitrate, sqrt(rc->variance_st_encoding_bitrate),
595 if (frame_cnt - 1 != tot_num_frames)
596 die(
"Error: Number of input frames not equal to output!\n");
600static void set_layer_pattern(
604 int spatial_layer_id,
int is_key_frame,
int ksvc_mode,
int speed) {
607 int use_rps_example = 0;
609 int enable_longterm_temporal_ref = 1;
610 int shift = (layering_mode == 8) ? 2 : 0;
611 *use_svc_control = 1;
614 int base_count = superframe_cnt >> 2;
621 for (i = 0; i < INTER_REFS_PER_FRAME; i++) ref_frame_config->
ref_idx[i] = i;
622 for (i = 0; i < INTER_REFS_PER_FRAME; i++) ref_frame_config->
reference[i] = 0;
623 for (i = 0; i < REF_FRAMES; i++) ref_frame_config->
refresh[i] = 0;
630 switch (layering_mode) {
632 if (use_rps_example == 0) {
636 ref_frame_config->
refresh[0] = 1;
637 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
645 int last_idx_refresh = 0;
654 if (superframe_cnt > 1) last_idx = (superframe_cnt - 1) % sh;
656 last_idx_refresh = superframe_cnt % sh;
658 if (superframe_cnt > lag_gld) gld_idx = (superframe_cnt - lag_gld) % sh;
660 if (superframe_cnt > lag_alt)
661 alt_ref_idx = (superframe_cnt - lag_alt) % sh;
664 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
665 ref_frame_config->
ref_idx[i] = last_idx;
667 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = last_idx;
668 ref_frame_config->
ref_idx[SVC_LAST2_FRAME] = last_idx_refresh;
669 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = gld_idx;
670 ref_frame_config->
ref_idx[SVC_ALTREF_FRAME] = alt_ref_idx;
672 ref_frame_config->
refresh[last_idx_refresh] = 1;
674 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
675 ref_frame_config->
reference[SVC_ALTREF_FRAME] = 1;
676 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
678 if (superframe_cnt >= 200 && superframe_cnt < 250) {
679 ref_frame_config->
reference[SVC_LAST_FRAME] = 0;
680 ref_frame_config->
reference[SVC_ALTREF_FRAME] = 1;
681 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 0;
684 if (superframe_cnt >= 400 && superframe_cnt < 450) {
685 ref_frame_config->
reference[SVC_LAST_FRAME] = 0;
686 ref_frame_config->
reference[SVC_ALTREF_FRAME] = 0;
687 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
695 if (superframe_cnt % 2 == 0) {
698 ref_frame_config->
refresh[0] = 1;
699 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
703 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
711 if (superframe_cnt % 4 == 0) {
715 ref_frame_config->
refresh[0] = 1;
716 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
717 }
else if ((superframe_cnt - 1) % 4 == 0) {
720 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
721 }
else if ((superframe_cnt - 2) % 4 == 0) {
724 ref_frame_config->
refresh[1] = 1;
725 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
726 }
else if ((superframe_cnt - 3) % 4 == 0) {
731 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
732 ref_frame_config->
ref_idx[SVC_LAST2_FRAME] = 0;
733 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
744 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 3;
747 if (base_count > 0) {
748 lag_index = 5 + (base_count % 3);
749 if (superframe_cnt % 4 != 0) lag_index = 5 + ((base_count + 1) % 3);
752 ref_frame_config->
ref_idx[SVC_ALTREF_FRAME] = lag_index;
753 if (superframe_cnt % 4 == 0) {
757 ref_frame_config->
refresh[0] = 1;
758 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
760 if (base_count % 10 == 0) ref_frame_config->
refresh[3] = 1;
762 ref_frame_config->
refresh[lag_index] = 1;
763 }
else if ((superframe_cnt - 1) % 4 == 0) {
766 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
767 }
else if ((superframe_cnt - 2) % 4 == 0) {
770 ref_frame_config->
refresh[1] = 1;
771 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
772 }
else if ((superframe_cnt - 3) % 4 == 0) {
777 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
778 ref_frame_config->
ref_idx[SVC_LAST2_FRAME] = 0;
779 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
782 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
783 ref_frame_config->
reference[SVC_ALTREF_FRAME] = 1;
793 if (superframe_cnt % 4 == 0) {
797 ref_frame_config->
refresh[0] = 1;
798 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
799 }
else if ((superframe_cnt - 1) % 4 == 0) {
802 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
803 }
else if ((superframe_cnt - 2) % 4 == 0) {
806 ref_frame_config->
refresh[3] = 1;
807 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
808 }
else if ((superframe_cnt - 3) % 4 == 0) {
811 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
819 ref_frame_config->
refresh[0] = 1;
820 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
824 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
825 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 0;
826 ref_frame_config->
refresh[1] = 1;
827 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
828 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
840 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
841 ref_frame_config->
ref_idx[i] = 0;
842 ref_frame_config->
refresh[0] = 1;
843 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
848 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
849 ref_frame_config->
ref_idx[i] = 0;
850 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
851 ref_frame_config->
refresh[1] = 1;
852 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
853 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
858 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
859 ref_frame_config->
ref_idx[i] = 1;
860 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 2;
861 ref_frame_config->
refresh[2] = 1;
862 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
863 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
866 if (enable_longterm_temporal_ref) {
867 ref_frame_config->
ref_idx[SVC_ALTREF_FRAME] = REF_FRAMES - 1;
868 ref_frame_config->
reference[SVC_ALTREF_FRAME] = 1;
869 if (base_count % 10 == 0)
870 ref_frame_config->
refresh[REF_FRAMES - 1] = 1;
876 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
877 if (superframe_cnt % 4 == 0) {
883 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
884 ref_frame_config->
ref_idx[i] = 0;
885 ref_frame_config->
refresh[0] = 1;
888 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
889 ref_frame_config->
ref_idx[i] = 0;
890 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
891 ref_frame_config->
refresh[1] = 1;
893 }
else if ((superframe_cnt - 1) % 4 == 0) {
897 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
898 ref_frame_config->
ref_idx[i] = 0;
899 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 3;
900 ref_frame_config->
refresh[3] = 1;
905 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
906 ref_frame_config->
ref_idx[i] = 3;
907 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
909 }
else if ((superframe_cnt - 2) % 4 == 0) {
916 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
917 ref_frame_config->
ref_idx[i] = 0;
918 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 5 - shift;
919 ref_frame_config->
refresh[5 - shift] = 1;
924 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
925 ref_frame_config->
ref_idx[i] = 5 - shift;
926 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
927 ref_frame_config->
ref_idx[SVC_LAST3_FRAME] = 6 - shift;
928 ref_frame_config->
refresh[6 - shift] = 1;
930 }
else if ((superframe_cnt - 3) % 4 == 0) {
937 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
938 ref_frame_config->
ref_idx[i] = 0;
939 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 5 - shift;
940 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 3;
941 ref_frame_config->
refresh[3] = 1;
945 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
946 ref_frame_config->
ref_idx[i] = 0;
947 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 6 - shift;
948 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 3;
965 ref_frame_config->
reference[SVC_LAST_FRAME] = 1;
966 if (superframe_cnt % 4 == 0) {
972 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
973 ref_frame_config->
ref_idx[i] = 0;
974 ref_frame_config->
refresh[0] = 1;
979 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
980 ref_frame_config->
ref_idx[i] = 0;
981 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
982 ref_frame_config->
refresh[1] = 1;
987 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
988 ref_frame_config->
ref_idx[i] = 1;
989 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 2;
990 ref_frame_config->
refresh[2] = 1;
992 }
else if ((superframe_cnt - 1) % 4 == 0) {
999 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1000 ref_frame_config->
ref_idx[i] = 0;
1001 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 3;
1002 ref_frame_config->
refresh[3] = 1;
1007 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1008 ref_frame_config->
ref_idx[i] = 3;
1009 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
1010 ref_frame_config->
ref_idx[SVC_LAST2_FRAME] = 4;
1011 ref_frame_config->
refresh[4] = 1;
1016 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1017 ref_frame_config->
ref_idx[i] = 4;
1018 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 2;
1020 }
else if ((superframe_cnt - 2) % 4 == 0) {
1027 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1028 ref_frame_config->
ref_idx[i] = 0;
1029 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 5 - shift;
1030 ref_frame_config->
refresh[5 - shift] = 1;
1035 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1036 ref_frame_config->
ref_idx[i] = 5 - shift;
1037 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 1;
1038 ref_frame_config->
ref_idx[SVC_LAST3_FRAME] = 6 - shift;
1039 ref_frame_config->
refresh[6 - shift] = 1;
1044 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1045 ref_frame_config->
ref_idx[i] = 6 - shift;
1046 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 2;
1047 ref_frame_config->
ref_idx[SVC_LAST3_FRAME] = 7 - shift;
1048 ref_frame_config->
refresh[7 - shift] = 1;
1050 }
else if ((superframe_cnt - 3) % 4 == 0) {
1057 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1058 ref_frame_config->
ref_idx[i] = 0;
1059 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 5 - shift;
1060 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 3;
1061 ref_frame_config->
refresh[3] = 1;
1065 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1066 ref_frame_config->
ref_idx[i] = 0;
1067 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 6 - shift;
1068 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 3;
1069 ref_frame_config->
ref_idx[SVC_LAST2_FRAME] = 4;
1070 ref_frame_config->
refresh[4] = 1;
1074 for (i = 0; i < INTER_REFS_PER_FRAME; i++)
1075 ref_frame_config->
ref_idx[i] = 0;
1076 ref_frame_config->
ref_idx[SVC_LAST_FRAME] = 7 - shift;
1077 ref_frame_config->
ref_idx[SVC_GOLDEN_FRAME] = 4;
1082 ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 1;
1086 if (!is_key_frame) ref_frame_config->
reference[SVC_GOLDEN_FRAME] = 0;
1091 ref_frame_config->
reference[SVC_LAST_FRAME] = 0;
1099 layering_mode == 8) {
1100 ref_frame_config->
ref_idx[SVC_ALTREF_FRAME] = REF_FRAMES - 1;
1101 if (!is_key_frame) ref_frame_config->
reference[SVC_ALTREF_FRAME] = 1;
1103 ref_frame_config->
refresh[REF_FRAMES - 1] = 1;
1106 default: assert(0); die(
"Error: Unsupported temporal layering mode!\n");
1110#if CONFIG_AV1_DECODER
1112 const int frames_out,
int *mismatch_seen) {
1115 if (*mismatch_seen)
return;
1121#if CONFIG_AV1_HIGHBITDEPTH
1127 enc_img.
d_w, enc_img.
d_h, 16);
1128 aom_img_truncate_16_to_8(&enc_hbd_img, &enc_img);
1129 enc_img = enc_hbd_img;
1134 dec_img.
d_w, dec_img.
d_h, 16);
1135 aom_img_truncate_16_to_8(&dec_hbd_img, &dec_img);
1136 dec_img = dec_hbd_img;
1141 if (!aom_compare_img(&enc_img, &dec_img)) {
1142 int y[4], u[4], v[4];
1143#if CONFIG_AV1_HIGHBITDEPTH
1145 aom_find_mismatch_high(&enc_img, &dec_img, y, u, v);
1147 aom_find_mismatch(&enc_img, &dec_img, y, u, v);
1150 aom_find_mismatch(&enc_img, &dec_img, y, u, v);
1154 "Encode/decode mismatch on frame %d at"
1155 " Y[%d, %d] {%d/%d},"
1156 " U[%d, %d] {%d/%d},"
1157 " V[%d, %d] {%d/%d}",
1158 frames_out, y[0], y[1], y[2], y[3], u[0], u[1], u[2], u[3], v[0], v[1],
1160 *mismatch_seen = frames_out;
1168int main(
int argc,
const char **argv) {
1172 AvxVideoWriter *total_layer_file = NULL;
1173 FILE *total_layer_obu_file = NULL;
1182 int frame_duration = 1;
1188#if CONFIG_INTERNAL_STATS
1189 FILE *stats_file = fopen(
"opsnr.stt",
"a");
1190 if (stats_file == NULL) {
1191 die(
"Cannot open opsnr.stt\n");
1194#if CONFIG_AV1_DECODER
1195 int mismatch_seen = 0;
1199 struct RateControlMetrics rc;
1200 int64_t cx_time = 0;
1203 double sum_bitrate = 0.0;
1204 double sum_bitrate2 = 0.0;
1205 double framerate = 30.0;
1206 int use_svc_control = 1;
1207 int set_err_resil_frame = 0;
1208 zero(rc.layer_target_bitrate);
1210 memset(&app_input, 0,
sizeof(AppInput));
1211 memset(&svc_params, 0,
sizeof(svc_params));
1215 const int test_dynamic_scaling_single_layer = 0;
1218 app_input.input_ctx.framerate.numerator = 30;
1219 app_input.input_ctx.framerate.denominator = 1;
1220 app_input.input_ctx.only_i420 = 1;
1221 app_input.input_ctx.bit_depth = 0;
1222 app_input.speed = 7;
1223 exec_name = argv[0];
1247 parse_command_line(argc, argv, &app_input, &svc_params, &cfg);
1252 unsigned int width = cfg.
g_w;
1253 unsigned int height = cfg.
g_h;
1255 if (app_input.layering_mode >= 0) {
1256 if (ts_number_layers !=
1257 mode_to_num_temporal_layers[app_input.layering_mode] ||
1259 mode_to_num_spatial_layers[app_input.layering_mode]) {
1260 die(
"Number of layers doesn't match layering mode.");
1265 if (app_input.input_ctx.file_type != FILE_TYPE_Y4M) {
1267 die(
"Failed to allocate image (%dx%d)", width, height);
1276 unsigned int total_rate = 0;
1277 for (i = 0; i < ss_number_layers; i++) {
1283 die(
"Incorrect total target bitrate");
1287 if (ts_number_layers == 2) {
1290 }
else if (ts_number_layers == 3) {
1296 if (app_input.input_ctx.file_type == FILE_TYPE_Y4M) {
1298 cfg.
g_w = app_input.input_ctx.width;
1299 cfg.
g_h = app_input.input_ctx.height;
1301 cfg.
g_timebase.
num = app_input.input_ctx.framerate.denominator;
1302 cfg.
g_timebase.
den = app_input.input_ctx.framerate.numerator;
1305 set_rate_control_metrics(&rc, framerate, ss_number_layers, ts_number_layers);
1308 info.codec_fourcc = get_fourcc_by_aom_encoder(encoder);
1309 info.frame_width = cfg.
g_w;
1310 info.frame_height = cfg.
g_h;
1314 for (
unsigned int sl = 0; sl < ss_number_layers; ++sl) {
1315 for (
unsigned tl = 0; tl < ts_number_layers; ++tl) {
1316 i = sl * ts_number_layers + tl;
1317 char file_name[PATH_MAX];
1318 snprintf(file_name,
sizeof(file_name),
"%s_%u.av1",
1319 app_input.output_filename, i);
1320 if (app_input.output_obu) {
1321 obu_files[i] = fopen(file_name,
"wb");
1322 if (!obu_files[i]) die(
"Failed to open %s for writing", file_name);
1324 outfile[i] = aom_video_writer_open(file_name, kContainerIVF, &info);
1325 if (!outfile[i]) die(
"Failed to open %s for writing", file_name);
1329 if (app_input.output_obu) {
1330 total_layer_obu_file = fopen(app_input.output_filename,
"wb");
1331 if (!total_layer_obu_file)
1332 die(
"Failed to open %s for writing", app_input.output_filename);
1335 aom_video_writer_open(app_input.output_filename, kContainerIVF, &info);
1336 if (!total_layer_file)
1337 die(
"Failed to open %s for writing", app_input.output_filename);
1343 die(
"Failed to initialize encoder");
1345#if CONFIG_AV1_DECODER
1346 if (app_input.decode) {
1348 die(
"Failed to initialize decoder");
1382 if (app_input.tune_content == AOM_CONTENT_SCREEN) {
1391 for (i = 0; i < ss_number_layers * ts_number_layers; ++i) {
1395 for (i = 0; i < ss_number_layers; ++i) {
1399 if (ss_number_layers == 2) {
1402 }
else if (ss_number_layers == 3) {
1415 const int max_intra_size_pct = 300;
1417 max_intra_size_pct);
1420 for (
unsigned int lx = 0; lx < ts_number_layers * ss_number_layers; lx++) {
1421 cx_time_layer[lx] = 0;
1422 frame_cnt_layer[lx] = 0;
1426 while (frame_avail || got_data) {
1427 struct aom_usec_timer timer;
1428 frame_avail = read_frame(&(app_input.input_ctx), &raw);
1430 for (
unsigned int slx = 0; slx < ss_number_layers; slx++) {
1435 int is_key_frame = (frame_cnt % cfg.
kf_max_dist) == 0;
1437 if (app_input.layering_mode >= 0) {
1440 set_layer_pattern(app_input.layering_mode, frame_cnt, &layer_id,
1441 &ref_frame_config, &ref_frame_comp_pred,
1442 &use_svc_control, slx, is_key_frame,
1443 (app_input.layering_mode == 10), app_input.speed);
1445 if (use_svc_control) {
1449 &ref_frame_comp_pred);
1457 if (ts_number_layers == 2) {
1459 }
else if (ts_number_layers == 3) {
1460 if (frame_cnt % 2 != 0)
1462 else if ((frame_cnt > 1) && ((frame_cnt - 2) % 4 == 0))
1468 if (set_err_resil_frame) {
1471 int err_resil_mode =
1478 if (frame_avail && slx == 0) ++rc.layer_input_frames[layer];
1480 if (test_dynamic_scaling_single_layer) {
1483 int frame_2x2 = 200;
1484 int frame_4x4 = 400;
1485 int frame_2x2up = 600;
1486 int frame_orig = 800;
1487 if (frame_cnt >= frame_2x2 && frame_cnt < frame_4x4) {
1491 }
else if (frame_cnt >= frame_4x4 && frame_cnt < frame_2x2up) {
1495 }
else if (frame_cnt >= frame_2x2up && frame_cnt < frame_orig) {
1499 }
else if (frame_cnt >= frame_orig) {
1504 if (frame_cnt == frame_2x2 || frame_cnt == frame_4x4 ||
1505 frame_cnt == frame_2x2up || frame_cnt == frame_orig) {
1511 for (i = 0; i < REF_FRAMES; i++) ref_frame_config.
refresh[i] = 1;
1512 if (use_svc_control) {
1516 &ref_frame_comp_pred);
1522 aom_usec_timer_start(&timer);
1524 die_codec(&codec,
"Failed to encode frame");
1525 aom_usec_timer_mark(&timer);
1526 cx_time += aom_usec_timer_elapsed(&timer);
1527 cx_time_layer[layer] += aom_usec_timer_elapsed(&timer);
1528 frame_cnt_layer[layer] += 1;
1533 switch (pkt->
kind) {
1536 sl < ss_number_layers; ++sl) {
1538 tl < ts_number_layers; ++tl) {
1539 unsigned int j = sl * ts_number_layers + tl;
1540 if (app_input.output_obu) {
1544 aom_video_writer_write_frame(outfile[j], pkt->
data.
frame.
buf,
1548 rc.layer_encoding_bitrate[j] += 8.0 * pkt->
data.
frame.
sz;
1552 if (app_input.output_obu) {
1554 total_layer_obu_file);
1556 aom_video_writer_write_frame(total_layer_file,
1564 rc.layer_avg_frame_size[j] += 8.0 * pkt->
data.
frame.
sz;
1565 rc.layer_avg_rate_mismatch[j] +=
1566 fabs(8.0 * pkt->
data.
frame.
sz - rc.layer_pfb[j]) /
1575 if (frame_cnt > rc.window_size && slx == ss_number_layers - 1) {
1576 sum_bitrate += 0.001 * 8.0 * pkt->
data.
frame.
sz * framerate;
1577 rc.window_size = (rc.window_size <= 0) ? 1 : rc.window_size;
1578 if (frame_cnt % rc.window_size == 0) {
1579 rc.window_count += 1;
1580 rc.avg_st_encoding_bitrate += sum_bitrate / rc.window_size;
1581 rc.variance_st_encoding_bitrate +=
1582 (sum_bitrate / rc.window_size) *
1583 (sum_bitrate / rc.window_size);
1588 if (frame_cnt > rc.window_size + rc.window_size / 2 &&
1589 slx == ss_number_layers - 1) {
1590 sum_bitrate2 += 0.001 * 8.0 * pkt->
data.
frame.
sz * framerate;
1591 if (frame_cnt > 2 * rc.window_size &&
1592 frame_cnt % rc.window_size == 0) {
1593 rc.window_count += 1;
1594 rc.avg_st_encoding_bitrate += sum_bitrate2 / rc.window_size;
1595 rc.variance_st_encoding_bitrate +=
1596 (sum_bitrate2 / rc.window_size) *
1597 (sum_bitrate2 / rc.window_size);
1602#if CONFIG_AV1_DECODER
1603 if (app_input.decode) {
1606 die_codec(&decoder,
"Failed to decode frame.");
1614#if CONFIG_AV1_DECODER
1615 if (app_input.decode) {
1618 if ((ss_number_layers > 1 || ts_number_layers > 1) &&
1621 test_decode(&codec, &decoder, frame_cnt, &mismatch_seen);
1627 pts += frame_duration;
1630 close_input_file(&(app_input.input_ctx));
1631 printout_rate_control_summary(&rc, frame_cnt, ss_number_layers,
1635 for (
unsigned int slx = 0; slx < ss_number_layers; slx++)
1636 for (
unsigned int tlx = 0; tlx < ts_number_layers; tlx++) {
1637 int lx = slx * ts_number_layers + tlx;
1638 printf(
"Per layer encoding time/FPS stats for encoder: %d %d %d %f %f \n",
1639 slx, tlx, frame_cnt_layer[lx],
1640 (
float)cx_time_layer[lx] / (
double)(frame_cnt_layer[lx] * 1000),
1641 1000000 * (
double)frame_cnt_layer[lx] / (
double)cx_time_layer[lx]);
1645 printf(
"Frame cnt and encoding time/FPS stats for encoding: %d %f %f\n",
1646 frame_cnt, 1000 * (
float)cx_time / (
double)(frame_cnt * 1000000),
1647 1000000 * (
double)frame_cnt / (
double)cx_time);
1651#if CONFIG_INTERNAL_STATS
1652 if (mismatch_seen) {
1653 fprintf(stats_file,
"First mismatch occurred in frame %d\n", mismatch_seen);
1655 fprintf(stats_file,
"No mismatch detected in recon buffers\n");
1661 for (i = 0; i < ss_number_layers * ts_number_layers; ++i)
1662 aom_video_writer_close(outfile[i]);
1663 aom_video_writer_close(total_layer_file);
1665 if (app_input.input_ctx.file_type != FILE_TYPE_Y4M) {
1668 return EXIT_SUCCESS;
Describes the encoder algorithm interface to applications.
enum aom_chroma_sample_position aom_chroma_sample_position_t
List of chroma sample positions.
#define AOM_IMG_FMT_HIGHBITDEPTH
Definition: aom_image.h:38
aom_image_t * aom_img_alloc(aom_image_t *img, aom_img_fmt_t fmt, unsigned int d_w, unsigned int d_h, unsigned int align)
Open a descriptor, allocating storage for the underlying image.
@ AOM_IMG_FMT_I420
Definition: aom_image.h:45
void aom_img_free(aom_image_t *img)
Close an image descriptor.
Provides definitions for using AOM or AV1 encoder algorithm within the aom Codec Interface.
Declares top-level encoder structures and functions.
#define AOM_MAX_LAYERS
Definition: aomcx.h:1601
#define AOM_MAX_TS_LAYERS
Definition: aomcx.h:1603
aom_codec_iface_t * aom_codec_av1_cx(void)
The interface to the AV1 encoder.
@ AV1E_SET_ROW_MT
Codec control function to enable the row based multi-threading of the encoder, unsigned int parameter...
Definition: aomcx.h:361
@ AV1E_SET_ENABLE_SMOOTH_INTRA
Codec control function to turn on / off smooth intra modes usage, int parameter.
Definition: aomcx.h:1070
@ AV1E_SET_ENABLE_TPL_MODEL
Codec control function to enable RDO modulated by frame temporal dependency, unsigned int parameter.
Definition: aomcx.h:408
@ AV1E_SET_AQ_MODE
Codec control function to set adaptive quantization mode, unsigned int parameter.
Definition: aomcx.h:468
@ AV1E_SET_SVC_LAYER_ID
Codec control function to set the layer id, aom_svc_layer_id_t* parameter.
Definition: aomcx.h:1276
@ AV1E_SET_SVC_REF_FRAME_CONFIG
Codec control function to set reference frame config: the ref_idx and the refresh flags for each buff...
Definition: aomcx.h:1287
@ AV1E_SET_TUNE_CONTENT
Codec control function to set content type, aom_tune_content parameter.
Definition: aomcx.h:497
@ AV1E_SET_CDF_UPDATE_MODE
Codec control function to set CDF update mode, unsigned int parameter.
Definition: aomcx.h:506
@ AV1E_SET_ENABLE_ANGLE_DELTA
Codec control function to turn on/off intra angle delta, int parameter.
Definition: aomcx.h:1117
@ AV1E_SET_MV_COST_UPD_FREQ
Control to set frequency of the cost updates for motion vectors, unsigned int parameter.
Definition: aomcx.h:1254
@ AV1E_SET_INTRA_DEFAULT_TX_ONLY
Control to use default tx type only for intra modes, int parameter.
Definition: aomcx.h:1203
@ AV1E_SET_SVC_REF_FRAME_COMP_PRED
Codec control function to set reference frame compound prediction. aom_svc_ref_frame_comp_pred_t* par...
Definition: aomcx.h:1392
@ AV1E_SET_ENABLE_INTRABC
Codec control function to turn on/off intra block copy mode, int parameter.
Definition: aomcx.h:1113
@ AV1E_SET_ENABLE_WARPED_MOTION
Codec control function to turn on / off warped motion usage at sequence level, int parameter.
Definition: aomcx.h:1038
@ AV1E_SET_COEFF_COST_UPD_FREQ
Control to set frequency of the cost updates for coefficients, unsigned int parameter.
Definition: aomcx.h:1234
@ AV1E_SET_ENABLE_CDEF
Codec control function to encode with CDEF, unsigned int parameter.
Definition: aomcx.h:670
@ AV1E_SET_DV_COST_UPD_FREQ
Control to set frequency of the cost updates for intrabc motion vectors, unsigned int parameter.
Definition: aomcx.h:1358
@ AV1E_SET_SVC_PARAMS
Codec control function to set SVC parameters, aom_svc_params_t* parameter.
Definition: aomcx.h:1281
@ AV1E_SET_ENABLE_FILTER_INTRA
Codec control function to turn on / off filter intra usage at sequence level, int parameter.
Definition: aomcx.h:1059
@ AV1E_SET_ENABLE_PALETTE
Codec control function to turn on/off palette mode, int parameter.
Definition: aomcx.h:1109
@ AV1E_SET_ENABLE_CFL_INTRA
Codec control function to turn on / off CFL uv intra mode usage, int parameter.
Definition: aomcx.h:1088
@ AOME_SET_MAX_INTRA_BITRATE_PCT
Codec control function to set max data rate for intra frames, unsigned int parameter.
Definition: aomcx.h:306
@ AV1E_SET_ERROR_RESILIENT_MODE
Codec control function to enable error_resilient_mode, int parameter.
Definition: aomcx.h:442
@ AV1E_SET_ENABLE_OBMC
Codec control function to predict with OBMC mode, unsigned int parameter.
Definition: aomcx.h:697
@ AV1E_SET_LOOPFILTER_CONTROL
Codec control to control loop filter.
Definition: aomcx.h:1407
@ AOME_SET_SCALEMODE
Codec control function to set encoder scaling mode for the next frame to be coded,...
Definition: aomcx.h:197
@ AV1E_SET_TILE_COLUMNS
Codec control function to set number of tile columns. unsigned int parameter.
Definition: aomcx.h:380
@ AV1E_SET_ENABLE_ORDER_HINT
Codec control function to turn on / off frame order hint (int parameter). Affects: joint compound mod...
Definition: aomcx.h:865
@ AV1E_SET_DELTAQ_MODE
Codec control function to set the delta q mode, unsigned int parameter.
Definition: aomcx.h:1131
@ AV1E_SET_ENABLE_GLOBAL_MOTION
Codec control function to turn on / off global motion usage for a sequence, int parameter.
Definition: aomcx.h:1028
@ AOME_SET_CPUUSED
Codec control function to set encoder internal speed settings, int parameter.
Definition: aomcx.h:220
@ AV1E_SET_GF_CBR_BOOST_PCT
Boost percentage for Golden Frame in CBR mode, unsigned int parameter.
Definition: aomcx.h:339
@ AV1E_SET_MODE_COST_UPD_FREQ
Control to set frequency of the cost updates for mode, unsigned int parameter.
Definition: aomcx.h:1244
@ AV1_GET_NEW_FRAME_IMAGE
Codec control function to get a pointer to the new frame.
Definition: aom.h:70
const char * aom_codec_iface_name(aom_codec_iface_t *iface)
Return the name for a given interface.
aom_codec_err_t aom_codec_control(aom_codec_ctx_t *ctx, int ctrl_id,...)
Algorithm Control.
const struct aom_codec_iface aom_codec_iface_t
Codec interface structure.
Definition: aom_codec.h:254
aom_codec_err_t aom_codec_destroy(aom_codec_ctx_t *ctx)
Destroy a codec instance.
const char * aom_codec_err_to_string(aom_codec_err_t err)
Convert error number to printable string.
aom_codec_err_t
Algorithm return codes.
Definition: aom_codec.h:155
#define AOM_CODEC_CONTROL_TYPECHECKED(ctx, id, data)
aom_codec_control wrapper macro (adds type-checking, less flexible)
Definition: aom_codec.h:521
const void * aom_codec_iter_t
Iterator.
Definition: aom_codec.h:288
#define AOM_FRAME_IS_KEY
Definition: aom_codec.h:271
@ AOM_BITS_12
Definition: aom_codec.h:321
@ AOM_BITS_8
Definition: aom_codec.h:319
@ AOM_BITS_10
Definition: aom_codec.h:320
@ AOM_CODEC_INVALID_PARAM
An application-supplied parameter is not valid.
Definition: aom_codec.h:200
@ AOM_CODEC_MEM_ERROR
Memory operation failed.
Definition: aom_codec.h:163
@ AOM_CODEC_OK
Operation completed without error.
Definition: aom_codec.h:157
aom_codec_err_t aom_codec_decode(aom_codec_ctx_t *ctx, const uint8_t *data, size_t data_sz, void *user_priv)
Decode data.
#define aom_codec_dec_init(ctx, iface, cfg, flags)
Convenience macro for aom_codec_dec_init_ver()
Definition: aom_decoder.h:129
const aom_codec_cx_pkt_t * aom_codec_get_cx_data(aom_codec_ctx_t *ctx, aom_codec_iter_t *iter)
Encoded data iterator.
aom_codec_err_t aom_codec_encode(aom_codec_ctx_t *ctx, const aom_image_t *img, aom_codec_pts_t pts, unsigned long duration, aom_enc_frame_flags_t flags)
Encode a frame.
#define aom_codec_enc_init(ctx, iface, cfg, flags)
Convenience macro for aom_codec_enc_init_ver()
Definition: aom_encoder.h:935
aom_codec_err_t aom_codec_enc_config_default(aom_codec_iface_t *iface, aom_codec_enc_cfg_t *cfg, unsigned int usage)
Get the default configuration for a usage.
#define AOM_USAGE_REALTIME
usage parameter analogous to AV1 REALTIME mode.
Definition: aom_encoder.h:1008
@ AOM_CBR
Definition: aom_encoder.h:185
@ AOM_KF_AUTO
Definition: aom_encoder.h:200
@ AOM_CODEC_CX_FRAME_PKT
Definition: aom_encoder.h:108
Codec context structure.
Definition: aom_codec.h:298
aom_codec_err_t err
Definition: aom_codec.h:301
Encoder output packet.
Definition: aom_encoder.h:120
size_t sz
Definition: aom_encoder.h:125
enum aom_codec_cx_pkt_kind kind
Definition: aom_encoder.h:121
union aom_codec_cx_pkt::@1 data
struct aom_codec_cx_pkt::@1::@2 frame
aom_codec_frame_flags_t flags
Definition: aom_encoder.h:130
void * buf
Definition: aom_encoder.h:124
Encoder configuration structure.
Definition: aom_encoder.h:385
unsigned int g_input_bit_depth
Bit-depth of the input frames.
Definition: aom_encoder.h:473
unsigned int rc_dropframe_thresh
Temporal resampling configuration, if supported by the codec.
Definition: aom_encoder.h:538
struct aom_rational g_timebase
Stream timebase units.
Definition: aom_encoder.h:487
unsigned int g_usage
Algorithm specific "usage" value.
Definition: aom_encoder.h:397
unsigned int rc_buf_sz
Decoder Buffer Size.
Definition: aom_encoder.h:702
unsigned int g_h
Height of the frame.
Definition: aom_encoder.h:433
enum aom_kf_mode kf_mode
Keyframe placement mode.
Definition: aom_encoder.h:765
enum aom_rc_mode rc_end_usage
Rate control algorithm to use.
Definition: aom_encoder.h:621
unsigned int g_threads
Maximum number of threads to use.
Definition: aom_encoder.h:405
unsigned int kf_min_dist
Keyframe minimum interval.
Definition: aom_encoder.h:774
unsigned int g_lag_in_frames
Allow lagged encoding.
Definition: aom_encoder.h:516
unsigned int rc_buf_initial_sz
Decoder Buffer Initial Size.
Definition: aom_encoder.h:711
unsigned int g_profile
Bitstream profile to use.
Definition: aom_encoder.h:415
aom_bit_depth_t g_bit_depth
Bit-depth of the codec.
Definition: aom_encoder.h:465
unsigned int g_w
Width of the frame.
Definition: aom_encoder.h:424
unsigned int rc_undershoot_pct
Rate control adaptation undershoot control.
Definition: aom_encoder.h:678
unsigned int kf_max_dist
Keyframe maximum interval.
Definition: aom_encoder.h:783
aom_codec_er_flags_t g_error_resilient
Enable error resilient modes.
Definition: aom_encoder.h:495
unsigned int rc_max_quantizer
Maximum (Worst Quality) Quantizer.
Definition: aom_encoder.h:665
unsigned int rc_buf_optimal_sz
Decoder Buffer Optimal Size.
Definition: aom_encoder.h:720
unsigned int rc_min_quantizer
Minimum (Best Quality) Quantizer.
Definition: aom_encoder.h:655
unsigned int rc_target_bitrate
Target data rate.
Definition: aom_encoder.h:641
unsigned int rc_resize_mode
Mode for spatial resampling, if supported by the codec.
Definition: aom_encoder.h:547
unsigned int rc_overshoot_pct
Rate control adaptation overshoot control.
Definition: aom_encoder.h:687
Image Descriptor.
Definition: aom_image.h:180
aom_img_fmt_t fmt
Definition: aom_image.h:181
unsigned int d_w
Definition: aom_image.h:195
unsigned int d_h
Definition: aom_image.h:196
int num
Definition: aom_encoder.h:163
int den
Definition: aom_encoder.h:164
aom image scaling mode
Definition: aomcx.h:1548
int temporal_layer_id
Definition: aomcx.h:1608
int spatial_layer_id
Definition: aomcx.h:1607
int max_quantizers[32]
Definition: aomcx.h:1615
int number_spatial_layers
Definition: aomcx.h:1613
int layer_target_bitrate[32]
Definition: aomcx.h:1620
int framerate_factor[8]
Definition: aomcx.h:1622
int min_quantizers[32]
Definition: aomcx.h:1616
int scaling_factor_den[4]
Definition: aomcx.h:1618
int number_temporal_layers
Definition: aomcx.h:1614
int scaling_factor_num[4]
Definition: aomcx.h:1617
int use_comp_pred[3]
Definition: aomcx.h:1639
int reference[7]
Definition: aomcx.h:1629
int refresh[8]
Definition: aomcx.h:1632
int ref_idx[7]
Definition: aomcx.h:1631