diff options
Diffstat (limited to 'examples')
-rw-r--r-- | examples/decode_to_md5.c | 40 | ||||
-rw-r--r-- | examples/decode_with_drops.c | 56 | ||||
-rw-r--r-- | examples/postproc.c | 44 | ||||
-rw-r--r-- | examples/simple_decoder.c | 44 | ||||
-rw-r--r-- | examples/simple_encoder.c | 272 | ||||
-rw-r--r-- | examples/twopass_encoder.c | 336 | ||||
-rw-r--r-- | examples/vp8_set_maps.c | 4 | ||||
-rw-r--r-- | examples/vpx_temporal_scalable_patterns.c | 551 |
8 files changed, 885 insertions, 462 deletions
diff --git a/examples/decode_to_md5.c b/examples/decode_to_md5.c index bba218209..077513cc7 100644 --- a/examples/decode_to_md5.c +++ b/examples/decode_to_md5.c @@ -38,9 +38,9 @@ #include "vpx/vp8dx.h" #include "vpx/vpx_decoder.h" -#include "./ivfdec.h" #include "./md5_utils.h" #include "./tools_common.h" +#include "./video_reader.h" #include "./vpx_config.h" static void get_image_md5(const vpx_image_t *img, unsigned char digest[16]) { @@ -79,41 +79,42 @@ void usage_exit() { } int main(int argc, char **argv) { - FILE *infile, *outfile; + int frame_cnt = 0; + FILE *outfile = NULL; vpx_codec_ctx_t codec; - vpx_codec_iface_t *iface; - int flags = 0, frame_cnt = 0; - vpx_video_t *video; + vpx_codec_iface_t *iface = NULL; + VpxVideoReader *reader = NULL; + const VpxVideoInfo *info = NULL; exec_name = argv[0]; if (argc != 3) - die("Invalid number of arguments"); + die("Invalid number of arguments."); - if (!(infile = fopen(argv[1], "rb"))) - die("Failed to open %s for reading", argv[1]); + reader = vpx_video_reader_open(argv[1]); + if (!reader) + die("Failed to open %s for reading.", argv[1]); if (!(outfile = fopen(argv[2], "wb"))) - die("Failed to open %s for writing", argv[2]); + die("Failed to open %s for writing.", argv[2]); - video = vpx_video_open_file(infile); - if (!video) - die("%s is not an IVF file.", argv[1]); + info = vpx_video_reader_get_info(reader); - iface = get_codec_interface(vpx_video_get_fourcc(video)); + iface = get_codec_interface(info->codec_fourcc); if (!iface) - die("Unknown FOURCC code."); + die("Unknown input codec."); printf("Using %s\n", vpx_codec_iface_name(iface)); - if (vpx_codec_dec_init(&codec, iface, NULL, flags)) + if (vpx_codec_dec_init(&codec, iface, NULL, 0)) die_codec(&codec, "Failed to initialize decoder"); - while (vpx_video_read_frame(video)) { + while (vpx_video_reader_read_frame(reader)) { vpx_codec_iter_t iter = NULL; vpx_image_t *img = NULL; size_t frame_size = 0; - const unsigned char *frame = vpx_video_get_frame(video, &frame_size); + const unsigned char *frame = vpx_video_reader_get_frame(reader, + &frame_size); if (vpx_codec_decode(&codec, frame, frame_size, NULL, 0)) die_codec(&codec, "Failed to decode frame"); @@ -129,11 +130,10 @@ int main(int argc, char **argv) { printf("Processed %d frames.\n", frame_cnt); if (vpx_codec_destroy(&codec)) - die_codec(&codec, "Failed to destroy codec"); + die_codec(&codec, "Failed to destroy codec."); - vpx_video_close(video); + vpx_video_reader_close(reader); fclose(outfile); - fclose(infile); return EXIT_SUCCESS; } diff --git a/examples/decode_with_drops.c b/examples/decode_with_drops.c index 12686dedd..e8fc0766b 100644 --- a/examples/decode_with_drops.c +++ b/examples/decode_with_drops.c @@ -56,14 +56,13 @@ #include <stdlib.h> #include <string.h> -#include "./ivfdec.h" - #define VPX_CODEC_DISABLE_COMPAT 1 #include "vpx/vp8dx.h" #include "vpx/vpx_decoder.h" #include "./tools_common.h" +#include "./video_reader.h" #include "./vpx_config.h" static const char *exec_name; @@ -74,52 +73,55 @@ void usage_exit() { } int main(int argc, char **argv) { - FILE *infile, *outfile; + int frame_cnt = 0; + FILE *outfile = NULL; vpx_codec_ctx_t codec; - vpx_codec_iface_t *iface; - int flags = 0, frame_cnt = 0; - vpx_video_t *video; - int n, m, is_range; - char *nptr; + vpx_codec_iface_t *iface = NULL; + VpxVideoReader *reader = NULL; + const VpxVideoInfo *info = NULL; + int n = 0; + int m = 0; + int is_range = 0; + char *nptr = NULL; exec_name = argv[0]; if (argc != 4) - die("Invalid number of arguments"); + die("Invalid number of arguments."); - if (!(infile = fopen(argv[1], "rb"))) - die("Failed to open %s for reading", argv[1]); + reader = vpx_video_reader_open(argv[1]); + if (!reader) + die("Failed to open %s for reading.", argv[1]); if (!(outfile = fopen(argv[2], "wb"))) - die("Failed to open %s for writing", argv[2]); + die("Failed to open %s for writing.", argv[2]); n = strtol(argv[3], &nptr, 0); m = strtol(nptr + 1, NULL, 0); is_range = (*nptr == '-'); if (!n || !m || (*nptr != '-' && *nptr != '/')) - die("Couldn't parse pattern %s\n", argv[3]); + die("Couldn't parse pattern %s.\n", argv[3]); - video = vpx_video_open_file(infile); - if (!video) - die("%s is not a supported input file.", argv[1]); + info = vpx_video_reader_get_info(reader); - iface = get_codec_interface(vpx_video_get_fourcc(video)); + iface = get_codec_interface(info->codec_fourcc); if (!iface) - die("Unknown FOURCC code."); + die("Unknown input codec."); printf("Using %s\n", vpx_codec_iface_name(iface)); - if (vpx_codec_dec_init(&codec, iface, NULL, flags)) - die_codec(&codec, "Failed to initialize decoder"); + if (vpx_codec_dec_init(&codec, iface, NULL, 0)) + die_codec(&codec, "Failed to initialize decoder."); - while (vpx_video_read_frame(video)) { + while (vpx_video_reader_read_frame(reader)) { vpx_codec_iter_t iter = NULL; vpx_image_t *img = NULL; size_t frame_size = 0; int skip; - const unsigned char *frame = vpx_video_get_frame(video, &frame_size); + const unsigned char *frame = vpx_video_reader_get_frame(reader, + &frame_size); if (vpx_codec_decode(&codec, frame, frame_size, NULL, 0)) - die_codec(&codec, "Failed to decode frame"); + die_codec(&codec, "Failed to decode frame."); ++frame_cnt; @@ -140,15 +142,13 @@ int main(int argc, char **argv) { printf("Processed %d frames.\n", frame_cnt); if (vpx_codec_destroy(&codec)) - die_codec(&codec, "Failed to destroy codec"); + die_codec(&codec, "Failed to destroy codec."); printf("Play: ffplay -f rawvideo -pix_fmt yuv420p -s %dx%d %s\n", - vpx_video_get_width(video), vpx_video_get_height(video), argv[2]); - - vpx_video_close(video); + info->frame_width, info->frame_height, argv[2]); + vpx_video_reader_close(reader); fclose(outfile); - fclose(infile); return EXIT_SUCCESS; } diff --git a/examples/postproc.c b/examples/postproc.c index 4ec2d1f1c..7281f1e3d 100644 --- a/examples/postproc.c +++ b/examples/postproc.c @@ -43,14 +43,13 @@ #include <stdlib.h> #include <string.h> -#include "./ivfdec.h" - #define VPX_CODEC_DISABLE_COMPAT 1 #include "vpx/vp8dx.h" #include "vpx/vpx_decoder.h" #include "./tools_common.h" +#include "./video_reader.h" #include "./vpx_config.h" static const char *exec_name; @@ -61,35 +60,34 @@ void usage_exit() { } int main(int argc, char **argv) { - FILE *infile, *outfile; - vpx_codec_ctx_t codec; - vpx_codec_iface_t *iface; int frame_cnt = 0; - vpx_video_t *video; + FILE *outfile = NULL; + vpx_codec_ctx_t codec; vpx_codec_err_t res; + vpx_codec_iface_t *iface = NULL; + VpxVideoReader *reader = NULL; + const VpxVideoInfo *info = NULL; exec_name = argv[0]; if (argc != 3) - die("Invalid number of arguments"); + die("Invalid number of arguments."); - if (!(infile = fopen(argv[1], "rb"))) - die("Failed to open %s for reading", argv[1]); + reader = vpx_video_reader_open(argv[1]); + if (!reader) + die("Failed to open %s for reading.", argv[1]); if (!(outfile = fopen(argv[2], "wb"))) die("Failed to open %s for writing", argv[2]); - video = vpx_video_open_file(infile); - if (!video) - die("%s is not a supported input file.", argv[1]); + info = vpx_video_reader_get_info(reader); - iface = get_codec_interface(vpx_video_get_fourcc(video)); + iface = get_codec_interface(info->codec_fourcc); if (!iface) - die("Unknown FOURCC code."); + die("Unknown input codec."); printf("Using %s\n", vpx_codec_iface_name(iface)); - res = vpx_codec_dec_init(&codec, iface, NULL, VPX_CODEC_USE_POSTPROC); if (res == VPX_CODEC_INCAPABLE) { printf("NOTICE: Postproc not supported.\n"); @@ -97,13 +95,14 @@ int main(int argc, char **argv) { } if (res) - die_codec(&codec, "Failed to initialize decoder"); + die_codec(&codec, "Failed to initialize decoder."); - while (vpx_video_read_frame(video)) { + while (vpx_video_reader_read_frame(reader)) { vpx_codec_iter_t iter = NULL; vpx_image_t *img = NULL; size_t frame_size = 0; - const unsigned char *frame = vpx_video_get_frame(video, &frame_size); + const unsigned char *frame = vpx_video_reader_get_frame(reader, + &frame_size); ++frame_cnt; @@ -111,12 +110,12 @@ int main(int argc, char **argv) { vp8_postproc_cfg_t pp = {0, 0, 0}; if (vpx_codec_control(&codec, VP8_SET_POSTPROC, &pp)) - die_codec(&codec, "Failed to turn off postproc"); + die_codec(&codec, "Failed to turn off postproc."); } else if (frame_cnt % 30 == 16) { vp8_postproc_cfg_t pp = {VP8_DEBLOCK | VP8_DEMACROBLOCK | VP8_MFQE, 4, 0}; if (vpx_codec_control(&codec, VP8_SET_POSTPROC, &pp)) - die_codec(&codec, "Failed to turn on postproc"); + die_codec(&codec, "Failed to turn on postproc."); }; // Decode the frame with 15ms deadline @@ -133,11 +132,10 @@ int main(int argc, char **argv) { die_codec(&codec, "Failed to destroy codec"); printf("Play: ffplay -f rawvideo -pix_fmt yuv420p -s %dx%d %s\n", - vpx_video_get_width(video), vpx_video_get_height(video), argv[2]); + info->frame_width, info->frame_height, argv[2]); - vpx_video_close(video); + vpx_video_reader_close(reader); fclose(outfile); - fclose(infile); return EXIT_SUCCESS; } diff --git a/examples/simple_decoder.c b/examples/simple_decoder.c index 23399f44f..4dc930897 100644 --- a/examples/simple_decoder.c +++ b/examples/simple_decoder.c @@ -86,8 +86,8 @@ #include "vpx/vp8dx.h" #include "vpx/vpx_decoder.h" -#include "./ivfdec.h" #include "./tools_common.h" +#include "./video_reader.h" #include "./vpx_config.h" static const char *exec_name; @@ -98,43 +98,44 @@ void usage_exit() { } int main(int argc, char **argv) { - FILE *infile, *outfile; + int frame_cnt = 0; + FILE *outfile = NULL; vpx_codec_ctx_t codec; - vpx_codec_iface_t *iface; - int flags = 0, frame_cnt = 0; - vpx_video_t *video; + vpx_codec_iface_t *iface = NULL; + VpxVideoReader *reader = NULL; + const VpxVideoInfo *info = NULL; exec_name = argv[0]; if (argc != 3) - die("Invalid number of arguments"); + die("Invalid number of arguments."); - if (!(infile = fopen(argv[1], "rb"))) - die("Failed to open %s for reading", argv[1]); + reader = vpx_video_reader_open(argv[1]); + if (!reader) + die("Failed to open %s for reading.", argv[1]); if (!(outfile = fopen(argv[2], "wb"))) - die("Failed to open %s for writing", argv[2]); + die("Failed to open %s for writing.", argv[2]); - video = vpx_video_open_file(infile); - if (!video) - die("%s is not an IVF file.", argv[1]); + info = vpx_video_reader_get_info(reader); - iface = get_codec_interface(vpx_video_get_fourcc(video)); + iface = get_codec_interface(info->codec_fourcc); if (!iface) - die("Unknown FOURCC code."); + die("Unknown input codec."); printf("Using %s\n", vpx_codec_iface_name(iface)); - if (vpx_codec_dec_init(&codec, iface, NULL, flags)) - die_codec(&codec, "Failed to initialize decoder"); + if (vpx_codec_dec_init(&codec, iface, NULL, 0)) + die_codec(&codec, "Failed to initialize decoder."); - while (vpx_video_read_frame(video)) { + while (vpx_video_reader_read_frame(reader)) { vpx_codec_iter_t iter = NULL; vpx_image_t *img = NULL; size_t frame_size = 0; - const unsigned char *frame = vpx_video_get_frame(video, &frame_size); + const unsigned char *frame = vpx_video_reader_get_frame(reader, + &frame_size); if (vpx_codec_decode(&codec, frame, frame_size, NULL, 0)) - die_codec(&codec, "Failed to decode frame"); + die_codec(&codec, "Failed to decode frame."); while ((img = vpx_codec_get_frame(&codec, &iter)) != NULL) { vpx_img_write(img, outfile); @@ -147,12 +148,11 @@ int main(int argc, char **argv) { die_codec(&codec, "Failed to destroy codec"); printf("Play: ffplay -f rawvideo -pix_fmt yuv420p -s %dx%d %s\n", - vpx_video_get_width(video), vpx_video_get_height(video), argv[2]); + info->frame_width, info->frame_height, argv[2]); - vpx_video_close(video); + vpx_video_reader_close(reader); fclose(outfile); - fclose(infile); return EXIT_SUCCESS; } diff --git a/examples/simple_encoder.c b/examples/simple_encoder.c index e64a962ae..50760549a 100644 --- a/examples/simple_encoder.c +++ b/examples/simple_encoder.c @@ -83,194 +83,114 @@ #include <stdio.h> #include <stdlib.h> -#include <stdarg.h> #include <string.h> + #define VPX_CODEC_DISABLE_COMPAT 1 -#include "vpx/vpx_encoder.h" #include "vpx/vp8cx.h" -#define interface (vpx_codec_vp8_cx()) -#define fourcc 0x30385056 - -#define IVF_FILE_HDR_SZ (32) -#define IVF_FRAME_HDR_SZ (12) - -static void mem_put_le16(char *mem, unsigned int val) { - mem[0] = val; - mem[1] = val>>8; -} - -static void mem_put_le32(char *mem, unsigned int val) { - mem[0] = val; - mem[1] = val>>8; - mem[2] = val>>16; - mem[3] = val>>24; -} - -static void die(const char *fmt, ...) { - va_list ap; - - va_start(ap, fmt); - vprintf(fmt, ap); - if(fmt[strlen(fmt)-1] != '\n') - printf("\n"); - exit(EXIT_FAILURE); -} - -static void die_codec(vpx_codec_ctx_t *ctx, const char *s) { - const char *detail = vpx_codec_error_detail(ctx); - - printf("%s: %s\n", s, vpx_codec_error(ctx)); - if(detail) - printf(" %s\n",detail); - exit(EXIT_FAILURE); -} - -static int read_frame(FILE *f, vpx_image_t *img) { - size_t nbytes, to_read; - int res = 1; - - to_read = img->w*img->h*3/2; - nbytes = fread(img->planes[0], 1, to_read, f); - if(nbytes != to_read) { - res = 0; - if(nbytes > 0) - printf("Warning: Read partial frame. Check your width & height!\n"); - } - return res; -} - -static void write_ivf_file_header(FILE *outfile, - const vpx_codec_enc_cfg_t *cfg, - int frame_cnt) { - char header[32]; - - if(cfg->g_pass != VPX_RC_ONE_PASS && cfg->g_pass != VPX_RC_LAST_PASS) - return; - header[0] = 'D'; - header[1] = 'K'; - header[2] = 'I'; - header[3] = 'F'; - mem_put_le16(header+4, 0); /* version */ - mem_put_le16(header+6, 32); /* headersize */ - mem_put_le32(header+8, fourcc); /* headersize */ - mem_put_le16(header+12, cfg->g_w); /* width */ - mem_put_le16(header+14, cfg->g_h); /* height */ - mem_put_le32(header+16, cfg->g_timebase.den); /* rate */ - mem_put_le32(header+20, cfg->g_timebase.num); /* scale */ - mem_put_le32(header+24, frame_cnt); /* length */ - mem_put_le32(header+28, 0); /* unused */ - - (void) fwrite(header, 1, 32, outfile); -} - +#include "vpx/vpx_encoder.h" -static void write_ivf_frame_header(FILE *outfile, - const vpx_codec_cx_pkt_t *pkt) -{ - char header[12]; - vpx_codec_pts_t pts; +#include "./tools_common.h" +#include "./video_writer.h" - if(pkt->kind != VPX_CODEC_CX_FRAME_PKT) - return; +#define interface (vpx_codec_vp8_cx()) - pts = pkt->data.frame.pts; - mem_put_le32(header, pkt->data.frame.sz); - mem_put_le32(header+4, pts&0xFFFFFFFF); - mem_put_le32(header+8, pts >> 32); +static const char *exec_name; - (void) fwrite(header, 1, 12, outfile); +void usage_exit() { + fprintf(stderr, "Usage: %s <width> <height> <infile> <outfile>\n", exec_name); + exit(EXIT_FAILURE); } int main(int argc, char **argv) { - FILE *infile, *outfile; - vpx_codec_ctx_t codec; - vpx_codec_enc_cfg_t cfg; - int frame_cnt = 0; - vpx_image_t raw; - vpx_codec_err_t res; - long width; - long height; - int frame_avail; - int got_data; - int flags = 0; - - /* Open files */ - if(argc!=5) - die("Usage: %s <width> <height> <infile> <outfile>\n", argv[0]); - width = strtol(argv[1], NULL, 0); - height = strtol(argv[2], NULL, 0); - if(width < 16 || width%2 || height <16 || height%2) - die("Invalid resolution: %ldx%ld", width, height); - if(!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, width, height, 1)) - die("Faile to allocate image", width, height); - if(!(outfile = fopen(argv[4], "wb"))) - die("Failed to open %s for writing", argv[4]); - - printf("Using %s\n",vpx_codec_iface_name(interface)); - - /* Populate encoder configuration */ - res = vpx_codec_enc_config_default(interface, &cfg, 0); - if(res) { - printf("Failed to get config: %s\n", vpx_codec_err_to_string(res)); - return EXIT_FAILURE; + FILE *infile = NULL; + vpx_codec_ctx_t codec; + vpx_codec_enc_cfg_t cfg; + int frame_count = 0; + vpx_image_t raw; + vpx_codec_err_t res; + VpxVideoInfo info = {0}; + VpxVideoWriter *writer = NULL; + const int fps = 30; // TODO(dkovalev) add command line argument + const int bitrate = 200; // kbit/s TODO(dkovalev) add command line argument + + exec_name = argv[0]; + + if (argc != 5) + die("Invalid number of arguments"); + + info.codec_fourcc = VP8_FOURCC; + info.frame_width = strtol(argv[1], NULL, 0); + info.frame_height = strtol(argv[2], NULL, 0); + info.time_base.numerator = 1; + info.time_base.denominator = fps; + + if (info.frame_width <= 0 || + info.frame_height <= 0 || + (info.frame_width % 2) != 0 || + (info.frame_height % 2) != 0) { + die("Invalid frame size: %dx%d", info.frame_width, info.frame_height); + } + + if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, info.frame_width, + info.frame_height, 1)) { + die("Failed to allocate image."); + } + + printf("Using %s\n", vpx_codec_iface_name(interface)); + + res = vpx_codec_enc_config_default(interface, &cfg, 0); + if (res) + die_codec(&codec, "Failed to get default codec config."); + + cfg.g_w = info.frame_width; + cfg.g_h = info.frame_height; + cfg.g_timebase.num = info.time_base.numerator; + cfg.g_timebase.den = info.time_base.denominator; + cfg.rc_target_bitrate = bitrate; + + writer = vpx_video_writer_open(argv[4], kContainerIVF, &info); + if (!writer) + die("Failed to open %s for writing.", argv[4]); + + if (!(infile = fopen(argv[3], "rb"))) + die("Failed to open %s for reading.", argv[3]); + + if (vpx_codec_enc_init(&codec, interface, &cfg, 0)) + die_codec(&codec, "Failed to initialize encoder"); + + while (vpx_img_read(&raw, infile)) { + vpx_codec_iter_t iter = NULL; + const vpx_codec_cx_pkt_t *pkt = NULL; + + ++frame_count; + + res = vpx_codec_encode(&codec, &raw, frame_count, 1, 0, + VPX_DL_GOOD_QUALITY); + if (res != VPX_CODEC_OK) + die_codec(&codec, "Failed to encode frame"); + + while ((pkt = vpx_codec_get_cx_data(&codec, &iter)) != NULL) { + if (pkt->kind == VPX_CODEC_CX_FRAME_PKT) { + const int keyframe = (pkt->data.frame.flags & VPX_FRAME_IS_KEY) != 0; + if (!vpx_video_writer_write_frame(writer, + pkt->data.frame.buf, + pkt->data.frame.sz, + pkt->data.frame.pts)) + die_codec(&codec, "Failed to write compressed frame."); + printf(keyframe ? "K" : "."); + fflush(stdout); + } } + } + printf("\n"); + fclose(infile); + printf("Processed %d frames.\n", frame_count); - /* Update the default configuration with our settings */ - cfg.rc_target_bitrate = width * height * cfg.rc_target_bitrate - / cfg.g_w / cfg.g_h; - cfg.g_w = width; - cfg.g_h = height; - - write_ivf_file_header(outfile, &cfg, 0); - - - /* Open input file for this encoding pass */ - if(!(infile = fopen(argv[3], "rb"))) - die("Failed to open %s for reading", argv[3]); - - /* Initialize codec */ - if(vpx_codec_enc_init(&codec, interface, &cfg, 0)) - die_codec(&codec, "Failed to initialize encoder"); - - frame_avail = 1; - got_data = 0; - while(frame_avail || got_data) { - vpx_codec_iter_t iter = NULL; - const vpx_codec_cx_pkt_t *pkt; - - frame_avail = read_frame(infile, &raw); - if(vpx_codec_encode(&codec, frame_avail? &raw : NULL, frame_cnt, - 1, flags, VPX_DL_REALTIME)) - die_codec(&codec, "Failed to encode frame"); - got_data = 0; - while( (pkt = vpx_codec_get_cx_data(&codec, &iter)) ) { - got_data = 1; - switch(pkt->kind) { - case VPX_CODEC_CX_FRAME_PKT: - write_ivf_frame_header(outfile, pkt); - (void) fwrite(pkt->data.frame.buf, 1, pkt->data.frame.sz, - outfile); - break; - default: - break; - } - printf(pkt->kind == VPX_CODEC_CX_FRAME_PKT - && (pkt->data.frame.flags & VPX_FRAME_IS_KEY)? "K":"."); - fflush(stdout); - } - frame_cnt++; - } - printf("\n"); - fclose(infile); + vpx_img_free(&raw); + if (vpx_codec_destroy(&codec)) + die_codec(&codec, "Failed to destroy codec."); - printf("Processed %d frames.\n",frame_cnt-1); - vpx_img_free(&raw); - if(vpx_codec_destroy(&codec)) - die_codec(&codec, "Failed to destroy codec"); + vpx_video_writer_close(writer); - /* Try to rewrite the file header with the actual frame count */ - if(!fseek(outfile, 0, SEEK_SET)) - write_ivf_file_header(outfile, &cfg, frame_cnt-1); - fclose(outfile); - return EXIT_SUCCESS; + return EXIT_SUCCESS; } diff --git a/examples/twopass_encoder.c b/examples/twopass_encoder.c index b0f0426c6..93b6150a5 100644 --- a/examples/twopass_encoder.c +++ b/examples/twopass_encoder.c @@ -50,218 +50,172 @@ #include <stdio.h> #include <stdlib.h> -#include <stdarg.h> #include <string.h> + #define VPX_CODEC_DISABLE_COMPAT 1 -#include "vpx/vpx_encoder.h" #include "vpx/vp8cx.h" -#define interface (vpx_codec_vp8_cx()) -#define fourcc 0x30385056 - -#define IVF_FILE_HDR_SZ (32) -#define IVF_FRAME_HDR_SZ (12) - -static void mem_put_le16(char *mem, unsigned int val) { - mem[0] = val; - mem[1] = val>>8; -} - -static void mem_put_le32(char *mem, unsigned int val) { - mem[0] = val; - mem[1] = val>>8; - mem[2] = val>>16; - mem[3] = val>>24; -} +#include "vpx/vpx_encoder.h" -static void die(const char *fmt, ...) { - va_list ap; +#include "./tools_common.h" +#include "./video_writer.h" - va_start(ap, fmt); - vprintf(fmt, ap); - if(fmt[strlen(fmt)-1] != '\n') - printf("\n"); - exit(EXIT_FAILURE); -} +#define interface (vpx_codec_vp8_cx()) -static void die_codec(vpx_codec_ctx_t *ctx, const char *s) { - const char *detail = vpx_codec_error_detail(ctx); +static const char *exec_name; - printf("%s: %s\n", s, vpx_codec_error(ctx)); - if(detail) - printf(" %s\n",detail); - exit(EXIT_FAILURE); +void usage_exit() { + fprintf(stderr, "Usage: %s <width> <height> <infile> <outfile>\n", exec_name); + exit(EXIT_FAILURE); } -static int read_frame(FILE *f, vpx_image_t *img) { - size_t nbytes, to_read; - int res = 1; - - to_read = img->w*img->h*3/2; - nbytes = fread(img->planes[0], 1, to_read, f); - if(nbytes != to_read) { - res = 0; - if(nbytes > 0) - printf("Warning: Read partial frame. Check your width & height!\n"); +static void get_frame_stats(vpx_codec_ctx_t *ctx, + const vpx_image_t *img, + vpx_codec_pts_t pts, + uint64_t duration, + vpx_enc_frame_flags_t flags, + uint64_t deadline, + vpx_fixed_buf_t *stats) { + vpx_codec_iter_t iter = NULL; + const vpx_codec_cx_pkt_t *pkt = NULL; + const vpx_codec_err_t res = vpx_codec_encode(ctx, img, pts, duration, flags, + deadline); + if (res != VPX_CODEC_OK) + die_codec(ctx, "Failed to get frame stats."); + + while ((pkt = vpx_codec_get_cx_data(ctx, &iter)) != NULL) { + if (pkt->kind == VPX_CODEC_STATS_PKT) { + const uint8_t *const pkt_buf = pkt->data.twopass_stats.buf; + const size_t pkt_size = pkt->data.twopass_stats.sz; + stats->buf = realloc(stats->buf, stats->sz + pkt_size); + memcpy((uint8_t *)stats->buf + stats->sz, pkt_buf, pkt_size); + stats->sz += pkt_size; } - return res; + } } -static void write_ivf_file_header(FILE *outfile, - const vpx_codec_enc_cfg_t *cfg, - int frame_cnt) { - char header[32]; - - if(cfg->g_pass != VPX_RC_ONE_PASS && cfg->g_pass != VPX_RC_LAST_PASS) - return; - header[0] = 'D'; - header[1] = 'K'; - header[2] = 'I'; - header[3] = 'F'; - mem_put_le16(header+4, 0); /* version */ - mem_put_le16(header+6, 32); /* headersize */ - mem_put_le32(header+8, fourcc); /* headersize */ - mem_put_le16(header+12, cfg->g_w); /* width */ - mem_put_le16(header+14, cfg->g_h); /* height */ - mem_put_le32(header+16, cfg->g_timebase.den); /* rate */ - mem_put_le32(header+20, cfg->g_timebase.num); /* scale */ - mem_put_le32(header+24, frame_cnt); /* length */ - mem_put_le32(header+28, 0); /* unused */ - - (void) fwrite(header, 1, 32, outfile); +static void encode_frame(vpx_codec_ctx_t *ctx, + const vpx_image_t *img, + vpx_codec_pts_t pts, + uint64_t duration, + vpx_enc_frame_flags_t flags, + uint64_t deadline, + VpxVideoWriter *writer) { + vpx_codec_iter_t iter = NULL; + const vpx_codec_cx_pkt_t *pkt = NULL; + const vpx_codec_err_t res = vpx_codec_encode(ctx, img, pts, duration, flags, + deadline); + if (res != VPX_CODEC_OK) + die_codec(ctx, "Failed to encode frame."); + + while ((pkt = vpx_codec_get_cx_data(ctx, &iter)) != NULL) { + if (pkt->kind == VPX_CODEC_CX_FRAME_PKT) { + const int keyframe = (pkt->data.frame.flags & VPX_FRAME_IS_KEY) != 0; + + if (!vpx_video_writer_write_frame(writer, pkt->data.frame.buf, + pkt->data.frame.sz, + pkt->data.frame.pts)) + die_codec(ctx, "Failed to write compressed frame."); + printf(keyframe ? "K" : "."); + fflush(stdout); + } + } } +int main(int argc, char **argv) { + FILE *infile = NULL; + VpxVideoWriter *writer = NULL; + vpx_codec_ctx_t codec; + vpx_codec_enc_cfg_t cfg; + vpx_image_t raw; + vpx_codec_err_t res; + vpx_fixed_buf_t stats = {0}; + VpxVideoInfo info = {0}; + int pass; + const int fps = 30; // TODO(dkovalev) add command line argument + const int bitrate = 200; // kbit/s TODO(dkovalev) add command line argument + + if (argc != 5) + die("Invalid number of arguments."); + + info.codec_fourcc = VP8_FOURCC; + info.time_base.numerator = 1; + info.time_base.denominator = fps; + info.frame_width = strtol(argv[1], NULL, 0); + info.frame_height = strtol(argv[2], NULL, 0); + + if (info.frame_width <= 0 || + info.frame_height <= 0 || + (info.frame_width % 2) != 0 || + (info.frame_height % 2) != 0) { + die("Invalid frame size: %dx%d", info.frame_width, info.frame_height); + } + + if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, info.frame_width, + info.frame_height, 1)) { + die("Failed to allocate image", info.frame_width, info.frame_height); + } + + writer = vpx_video_writer_open(argv[4], kContainerIVF, &info); + if (!writer) + die("Failed to open %s for writing", argv[4]); + + printf("Using %s\n", vpx_codec_iface_name(interface)); + + res = vpx_codec_enc_config_default(interface, &cfg, 0); + if (res) + die_codec(&codec, "Failed to get default codec config."); + + cfg.g_w = info.frame_width; + cfg.g_h = info.frame_height; + cfg.g_timebase.num = info.time_base.numerator; + cfg.g_timebase.den = info.time_base.denominator; + cfg.rc_target_bitrate = bitrate; + + for (pass = 0; pass < 2; ++pass) { + int frame_count = 0; + + if (pass == 0) { + cfg.g_pass = VPX_RC_FIRST_PASS; + } else { + cfg.g_pass = VPX_RC_LAST_PASS; + cfg.rc_twopass_stats_in = stats; + } -static void write_ivf_frame_header(FILE *outfile, - const vpx_codec_cx_pkt_t *pkt) -{ - char header[12]; - vpx_codec_pts_t pts; - - if(pkt->kind != VPX_CODEC_CX_FRAME_PKT) - return; + if (!(infile = fopen(argv[3], "rb"))) + die("Failed to open %s for reading", argv[3]); - pts = pkt->data.frame.pts; - mem_put_le32(header, pkt->data.frame.sz); - mem_put_le32(header+4, pts&0xFFFFFFFF); - mem_put_le32(header+8, pts >> 32); + if (vpx_codec_enc_init(&codec, interface, &cfg, 0)) + die_codec(&codec, "Failed to initialize encoder"); - (void) fwrite(header, 1, 12, outfile); -} + while (vpx_img_read(&raw, infile)) { + ++frame_count; -int main(int argc, char **argv) { - FILE *infile, *outfile; - vpx_codec_ctx_t codec; - vpx_codec_enc_cfg_t cfg; - int frame_cnt = 0; - vpx_image_t raw; - vpx_codec_err_t res; - long width; - long height; - int frame_avail; - int got_data; - int flags = 0; - int pass; - vpx_fixed_buf_t stats = {0}; - - /* Open files */ - if(argc!=5) - die("Usage: %s <width> <height> <infile> <outfile>\n", argv[0]); - width = strtol(argv[1], NULL, 0); - height = strtol(argv[2], NULL, 0); - if(width < 16 || width%2 || height <16 || height%2) - die("Invalid resolution: %ldx%ld", width, height); - if(!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, width, height, 1)) - die("Faile to allocate image", width, height); - if(!(outfile = fopen(argv[4], "wb"))) - die("Failed to open %s for writing", argv[4]); - - printf("Using %s\n",vpx_codec_iface_name(interface)); - - /* Populate encoder configuration */ - res = vpx_codec_enc_config_default(interface, &cfg, 0); - if(res) { - printf("Failed to get config: %s\n", vpx_codec_err_to_string(res)); - return EXIT_FAILURE; + if (pass == 0) { + get_frame_stats(&codec, &raw, frame_count, 1, 0, VPX_DL_BEST_QUALITY, + &stats); + } else { + encode_frame(&codec, &raw, frame_count, 1, 0, VPX_DL_BEST_QUALITY, + writer); + } } - /* Update the default configuration with our settings */ - cfg.rc_target_bitrate = width * height * cfg.rc_target_bitrate - / cfg.g_w / cfg.g_h; - cfg.g_w = width; - cfg.g_h = height; - - write_ivf_file_header(outfile, &cfg, 0); - - for(pass=0; pass<2; pass++) { - frame_cnt = 0; - - if(pass == 0) - cfg.g_pass = VPX_RC_FIRST_PASS; - else { - cfg.g_pass = VPX_RC_LAST_PASS; - cfg.rc_twopass_stats_in = stats; - } - - /* Open input file for this encoding pass */ - if(!(infile = fopen(argv[3], "rb"))) - die("Failed to open %s for reading", argv[3]); - - /* Initialize codec */ - if(vpx_codec_enc_init(&codec, interface, &cfg, 0)) - die_codec(&codec, "Failed to initialize encoder"); - - frame_avail = 1; - got_data = 0; - while(frame_avail || got_data) { - vpx_codec_iter_t iter = NULL; - const vpx_codec_cx_pkt_t *pkt; - - frame_avail = read_frame(infile, &raw); - if(vpx_codec_encode(&codec, frame_avail? &raw : NULL, frame_cnt, - 1, flags, VPX_DL_BEST_QUALITY)) - die_codec(&codec, "Failed to encode frame"); - got_data = 0; - while( (pkt = vpx_codec_get_cx_data(&codec, &iter)) ) { - got_data = 1; - switch(pkt->kind) { - case VPX_CODEC_CX_FRAME_PKT: - write_ivf_frame_header(outfile, pkt); - (void) fwrite(pkt->data.frame.buf, 1, pkt->data.frame.sz, - outfile); - break; - case VPX_CODEC_STATS_PKT: - stats.buf = realloc(stats.buf, stats.sz - + pkt->data.twopass_stats.sz); - if(!stats.buf) - die("Memory reallocation failed.\n"); - memcpy((char*)stats.buf + stats.sz, - pkt->data.twopass_stats.buf, - pkt->data.twopass_stats.sz); - stats.sz += pkt->data.twopass_stats.sz; - break; - default: - break; - } - printf(pkt->kind == VPX_CODEC_CX_FRAME_PKT - && (pkt->data.frame.flags & VPX_FRAME_IS_KEY)? "K":"."); - fflush(stdout); - } - frame_cnt++; - } - printf("\n"); - fclose(infile); - printf("Pass %d complete.\n", pass+1); - if(vpx_codec_destroy(&codec)) - die_codec(&codec, "Failed to destroy codec"); + if (pass == 0) { + get_frame_stats(&codec, NULL, frame_count, 1, 0, VPX_DL_BEST_QUALITY, + &stats); + } else { + printf("\n"); } - printf("Processed %d frames.\n",frame_cnt-1); - vpx_img_free(&raw); - free(stats.buf); + fclose(infile); + printf("Pass %d complete. Processed %d frames.\n", pass + 1, frame_count); + if (vpx_codec_destroy(&codec)) + die_codec(&codec, "Failed to destroy codec."); + } + + vpx_img_free(&raw); + free(stats.buf); + + vpx_video_writer_close(writer); - /* Try to rewrite the file header with the actual frame count */ - if(!fseek(outfile, 0, SEEK_SET)) - write_ivf_file_header(outfile, &cfg, frame_cnt-1); - fclose(outfile); - return EXIT_SUCCESS; + return EXIT_SUCCESS; } diff --git a/examples/vp8_set_maps.c b/examples/vp8_set_maps.c index 242788fd4..4c0e8a0ba 100644 --- a/examples/vp8_set_maps.c +++ b/examples/vp8_set_maps.c @@ -201,7 +201,7 @@ int main(int argc, char **argv) { if(frame_cnt + 1 == 22) { vpx_roi_map_t roi; - int i; + unsigned int i; roi.rows = cfg.g_h/16; roi.cols = cfg.g_w/16; @@ -232,7 +232,7 @@ int main(int argc, char **argv) { free(roi.roi_map); } else if(frame_cnt + 1 == 33) { vpx_active_map_t active; - int i; + unsigned int i; active.rows = cfg.g_h/16; active.cols = cfg.g_w/16; diff --git a/examples/vpx_temporal_scalable_patterns.c b/examples/vpx_temporal_scalable_patterns.c new file mode 100644 index 000000000..11d331bd8 --- /dev/null +++ b/examples/vpx_temporal_scalable_patterns.c @@ -0,0 +1,551 @@ +/* + * Copyright (c) 2012 The WebM project authors. All Rights Reserved. + * + * Use of this source code is governed by a BSD-style license + * that can be found in the LICENSE file in the root of the source + * tree. An additional intellectual property rights grant can be found + * in the file PATENTS. All contributing project authors may + * be found in the AUTHORS file in the root of the source tree. + */ + +// This is an example demonstrating how to implement a multi-layer VP9 +// encoding scheme based on temporal scalability for video applications +// that benefit from a scalable bitstream. + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#define VPX_CODEC_DISABLE_COMPAT 1 +#include "vpx/vp8cx.h" +#include "vpx/vpx_encoder.h" + +#include "./tools_common.h" +#include "./video_writer.h" + +static const char *exec_name; + +void usage_exit() { + exit(EXIT_FAILURE); +} + +static int mode_to_num_layers[12] = {1, 2, 2, 3, 3, 3, 3, 5, 2, 3, 3, 3}; + +// Temporal scaling parameters: +// NOTE: The 3 prediction frames cannot be used interchangeably due to +// differences in the way they are handled throughout the code. The +// frames should be allocated to layers in the order LAST, GF, ARF. +// Other combinations work, but may produce slightly inferior results. +static void set_temporal_layer_pattern(int layering_mode, + vpx_codec_enc_cfg_t *cfg, + int *layer_flags, + int *flag_periodicity) { + switch (layering_mode) { + case 0: { + // 1-layer. + int ids[1] = {0}; + cfg->ts_periodicity = 1; + *flag_periodicity = 1; + cfg->ts_number_layers = 1; + cfg->ts_rate_decimator[0] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // Update L only. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_UPD_GF | + VP8_EFLAG_NO_UPD_ARF; + break; + } + case 1: { + // 2-layers, 2-frame period. + int ids[2] = {0, 1}; + cfg->ts_periodicity = 2; + *flag_periodicity = 2; + cfg->ts_number_layers = 2; + cfg->ts_rate_decimator[0] = 2; + cfg->ts_rate_decimator[1] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); +#if 1 + // 0=L, 1=GF, Intra-layer prediction enabled. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_UPD_GF | + VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF; + layer_flags[1] = VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_REF_ARF; +#else + // 0=L, 1=GF, Intra-layer prediction disabled. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_UPD_GF | + VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF; + layer_flags[1] = VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_REF_LAST; +#endif + break; + } + case 2: { + // 2-layers, 3-frame period. + int ids[3] = {0, 1, 1}; + cfg->ts_periodicity = 3; + *flag_periodicity = 3; + cfg->ts_number_layers = 2; + cfg->ts_rate_decimator[0] = 3; + cfg->ts_rate_decimator[1] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, Intra-layer prediction enabled. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_REF_GF | + VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + layer_flags[1] = + layer_flags[2] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF | + VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST; + break; + } + case 3: { + // 3-layers, 6-frame period. + int ids[6] = {0, 2, 2, 1, 2, 2}; + cfg->ts_periodicity = 6; + *flag_periodicity = 6; + cfg->ts_number_layers = 3; + cfg->ts_rate_decimator[0] = 6; + cfg->ts_rate_decimator[1] = 3; + cfg->ts_rate_decimator[2] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, 2=ARF, Intra-layer prediction enabled. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_REF_GF | + VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + layer_flags[3] = VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_ARF | + VP8_EFLAG_NO_UPD_LAST; + layer_flags[1] = + layer_flags[2] = + layer_flags[4] = + layer_flags[5] = VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_LAST; + break; + } + case 4: { + // 3-layers, 4-frame period. + int ids[4] = {0, 2, 1, 2}; + cfg->ts_periodicity = 4; + *flag_periodicity = 4; + cfg->ts_number_layers = 3; + cfg->ts_rate_decimator[0] = 4; + cfg->ts_rate_decimator[1] = 2; + cfg->ts_rate_decimator[2] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, 2=ARF, Intra-layer prediction disabled. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_REF_GF | + VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + layer_flags[2] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF | + VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST; + layer_flags[1] = + layer_flags[3] = VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + break; + } + case 5: { + // 3-layers, 4-frame period. + int ids[4] = {0, 2, 1, 2}; + cfg->ts_periodicity = 4; + *flag_periodicity = 4; + cfg->ts_number_layers = 3; + cfg->ts_rate_decimator[0] = 4; + cfg->ts_rate_decimator[1] = 2; + cfg->ts_rate_decimator[2] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, 2=ARF, Intra-layer prediction enabled in layer 1, disabled + // in layer 2. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_REF_GF | + VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + layer_flags[2] = VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_ARF; + layer_flags[1] = + layer_flags[3] = VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + break; + } + case 6: { + // 3-layers, 4-frame period. + int ids[4] = {0, 2, 1, 2}; + cfg->ts_periodicity = 4; + *flag_periodicity = 4; + cfg->ts_number_layers = 3; + cfg->ts_rate_decimator[0] = 4; + cfg->ts_rate_decimator[1] = 2; + cfg->ts_rate_decimator[2] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, 2=ARF, Intra-layer prediction enabled. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_REF_GF | + VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + layer_flags[2] = VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_ARF; + layer_flags[1] = + layer_flags[3] = VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF; + break; + } + case 7: { + // NOTE: Probably of academic interest only. + // 5-layers, 16-frame period. + int ids[16] = {0, 4, 3, 4, 2, 4, 3, 4, 1, 4, 3, 4, 2, 4, 3, 4}; + cfg->ts_periodicity = 16; + *flag_periodicity = 16; + cfg->ts_number_layers = 5; + cfg->ts_rate_decimator[0] = 16; + cfg->ts_rate_decimator[1] = 8; + cfg->ts_rate_decimator[2] = 4; + cfg->ts_rate_decimator[3] = 2; + cfg->ts_rate_decimator[4] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + layer_flags[0] = VPX_EFLAG_FORCE_KF; + layer_flags[1] = + layer_flags[3] = + layer_flags[5] = + layer_flags[7] = + layer_flags[9] = + layer_flags[11] = + layer_flags[13] = + layer_flags[15] = VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF | + VP8_EFLAG_NO_UPD_ARF; + layer_flags[2] = + layer_flags[6] = + layer_flags[10] = + layer_flags[14] = VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_GF; + layer_flags[4] = + layer_flags[12] = VP8_EFLAG_NO_REF_LAST | VP8_EFLAG_NO_UPD_ARF; + layer_flags[8] = VP8_EFLAG_NO_REF_LAST | VP8_EFLAG_NO_REF_GF; + break; + } + case 8: { + // 2-layers, with sync point at first frame of layer 1. + int ids[2] = {0, 1}; + cfg->ts_periodicity = 2; + *flag_periodicity = 8; + cfg->ts_number_layers = 2; + cfg->ts_rate_decimator[0] = 2; + cfg->ts_rate_decimator[1] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF. + // ARF is used as predictor for all frames, and is only updated on + // key frame. Sync point every 8 frames. + + // Layer 0: predict from L and ARF, update L and G. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_REF_GF | + VP8_EFLAG_NO_UPD_ARF; + // Layer 1: sync point: predict from L and ARF, and update G. + layer_flags[1] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_ARF; + // Layer 0, predict from L and ARF, update L. + layer_flags[2] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_UPD_GF | + VP8_EFLAG_NO_UPD_ARF; + // Layer 1: predict from L, G and ARF, and update G. + layer_flags[3] = VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_ENTROPY; + // Layer 0. + layer_flags[4] = layer_flags[2]; + // Layer 1. + layer_flags[5] = layer_flags[3]; + // Layer 0. + layer_flags[6] = layer_flags[4]; + // Layer 1. + layer_flags[7] = layer_flags[5]; + break; + } + case 9: { + // 3-layers: Sync points for layer 1 and 2 every 8 frames. + int ids[4] = {0, 2, 1, 2}; + cfg->ts_periodicity = 4; + *flag_periodicity = 8; + cfg->ts_number_layers = 3; + cfg->ts_rate_decimator[0] = 4; + cfg->ts_rate_decimator[1] = 2; + cfg->ts_rate_decimator[2] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, 2=ARF. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_REF_GF | + VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + layer_flags[1] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF | + VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF; + layer_flags[2] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF | + VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_ARF; + layer_flags[3] = + layer_flags[5] = VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF; + layer_flags[4] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_REF_ARF | + VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF; + layer_flags[6] = VP8_EFLAG_NO_REF_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_ARF; + layer_flags[7] = VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_GF | + VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_ENTROPY; + break; + } + case 10: { + // 3-layers structure where ARF is used as predictor for all frames, + // and is only updated on key frame. + // Sync points for layer 1 and 2 every 8 frames. + + int ids[4] = {0, 2, 1, 2}; + cfg->ts_periodicity = 4; + *flag_periodicity = 8; + cfg->ts_number_layers = 3; + cfg->ts_rate_decimator[0] = 4; + cfg->ts_rate_decimator[1] = 2; + cfg->ts_rate_decimator[2] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, 2=ARF. + // Layer 0: predict from L and ARF; update L and G. + layer_flags[0] = VPX_EFLAG_FORCE_KF | VP8_EFLAG_NO_UPD_ARF | + VP8_EFLAG_NO_REF_GF; + // Layer 2: sync point: predict from L and ARF; update none. + layer_flags[1] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_UPD_GF | + VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST | + VP8_EFLAG_NO_UPD_ENTROPY; + // Layer 1: sync point: predict from L and ARF; update G. + layer_flags[2] = VP8_EFLAG_NO_REF_GF | VP8_EFLAG_NO_UPD_ARF | + VP8_EFLAG_NO_UPD_LAST; + // Layer 2: predict from L, G, ARF; update none. + layer_flags[3] = VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF | + VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_ENTROPY; + // Layer 0: predict from L and ARF; update L. + layer_flags[4] = VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF | + VP8_EFLAG_NO_REF_GF; + // Layer 2: predict from L, G, ARF; update none. + layer_flags[5] = layer_flags[3]; + // Layer 1: predict from L, G, ARF; update G. + layer_flags[6] = VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST; + // Layer 2: predict from L, G, ARF; update none. + layer_flags[7] = layer_flags[3]; + break; + } + case 11: + default: { + // 3-layers structure as in case 10, but no sync/refresh points for + // layer 1 and 2. + int ids[4] = {0, 2, 1, 2}; + cfg->ts_periodicity = 4; + *flag_periodicity = 8; + cfg->ts_number_layers = 3; + cfg->ts_rate_decimator[0] = 4; + cfg->ts_rate_decimator[1] = 2; + cfg->ts_rate_decimator[2] = 1; + memcpy(cfg->ts_layer_id, ids, sizeof(ids)); + // 0=L, 1=GF, 2=ARF. + // Layer 0: predict from L and ARF; update L. + layer_flags[0] = VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF | + VP8_EFLAG_NO_REF_GF; + layer_flags[4] = layer_flags[0]; + // Layer 1: predict from L, G, ARF; update G. + layer_flags[2] = VP8_EFLAG_NO_UPD_ARF | VP8_EFLAG_NO_UPD_LAST; + layer_flags[6] = layer_flags[2]; + // Layer 2: predict from L, G, ARF; update none. + layer_flags[1] = VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF | + VP8_EFLAG_NO_UPD_LAST | VP8_EFLAG_NO_UPD_ENTROPY; + layer_flags[3] = layer_flags[1]; + layer_flags[5] = layer_flags[1]; + layer_flags[7] = layer_flags[1]; + break; + } + } +} + +int main(int argc, char **argv) { + VpxVideoWriter *outfile[VPX_TS_MAX_LAYERS]; + vpx_codec_ctx_t codec; + vpx_codec_enc_cfg_t cfg; + int frame_cnt = 0; + vpx_image_t raw; + vpx_codec_err_t res; + unsigned int width; + unsigned int height; + int frame_avail; + int got_data; + int flags = 0; + int i; + int pts = 0; // PTS starts at 0. + int frame_duration = 1; // 1 timebase tick per frame. + int layering_mode = 0; + int frames_in_layer[VPX_TS_MAX_LAYERS] = {0}; + int layer_flags[VPX_TS_MAX_PERIODICITY] = {0}; + int flag_periodicity = 1; + int max_intra_size_pct; + vpx_svc_layer_id_t layer_id = {0, 0}; + char *codec_type; + vpx_codec_iface_t *(*interface)(void); + unsigned int fourcc; + struct VpxInputContext input_ctx = {0}; + + exec_name = argv[0]; + // Check usage and arguments. + if (argc < 10) { + die("Usage: %s <infile> <outfile> <codec_type(vp8/vp9)> <width> <height> " + "<rate_num> <rate_den> <mode> <Rate_0> ... <Rate_nlayers-1> \n", + argv[0]); + } + + codec_type = argv[3]; + if (strncmp(codec_type, "vp9", 3) == 0) { +#if CONFIG_VP9_ENCODER + interface = vpx_codec_vp9_cx; + fourcc = VP9_FOURCC; +#else + die("Encoder vp9 selected but not configured"); +#endif + } else { +#if CONFIG_VP8_ENCODER + interface = vpx_codec_vp8_cx; + fourcc = VP8_FOURCC; +#else + die("Encoder vp8 selected but not configured"); +#endif + } + printf("Using %s\n", vpx_codec_iface_name(interface())); + + width = strtol(argv[4], NULL, 0); + height = strtol(argv[5], NULL, 0); + if (width < 16 || width % 2 || height < 16 || height % 2) { + die("Invalid resolution: %d x %d", width, height); + } + + layering_mode = strtol(argv[8], NULL, 0); + if (layering_mode < 0 || layering_mode > 11) { + die("Invalid mode (0..11) %s", argv[8]); + } + + if (argc != 9 + mode_to_num_layers[layering_mode]) { + die("Invalid number of arguments"); + } + + if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, width, height, 32)) { + die("Failed to allocate image", width, height); + } + + // Populate encoder configuration. + res = vpx_codec_enc_config_default(interface(), &cfg, 0); + if (res) { + printf("Failed to get config: %s\n", vpx_codec_err_to_string(res)); + return EXIT_FAILURE; + } + + // Update the default configuration with our settings. + cfg.g_w = width; + cfg.g_h = height; + + // Timebase format e.g. 30fps: numerator=1, demoninator = 30. + cfg.g_timebase.num = strtol(argv[6], NULL, 0); + cfg.g_timebase.den = strtol(argv[7], NULL, 0); + + for (i = 9; i < 9 + mode_to_num_layers[layering_mode]; ++i) { + cfg.ts_target_bitrate[i - 9] = strtol(argv[i], NULL, 0); + } + + // Real time parameters. + cfg.rc_dropframe_thresh = 0; + cfg.rc_end_usage = VPX_CBR; + cfg.rc_resize_allowed = 0; + cfg.rc_min_quantizer = 2; + cfg.rc_max_quantizer = 56; + cfg.rc_undershoot_pct = 100; + cfg.rc_overshoot_pct = 15; + cfg.rc_buf_initial_sz = 500; + cfg.rc_buf_optimal_sz = 600; + cfg.rc_buf_sz = 1000; + + // Enable error resilient mode. + cfg.g_error_resilient = 1; + cfg.g_lag_in_frames = 0; + cfg.kf_mode = VPX_KF_DISABLED; + + // Disable automatic keyframe placement. + cfg.kf_min_dist = cfg.kf_max_dist = 3000; + + // Default setting for bitrate: used in special case of 1 layer (case 0). + cfg.rc_target_bitrate = cfg.ts_target_bitrate[0]; + + set_temporal_layer_pattern(layering_mode, + &cfg, + layer_flags, + &flag_periodicity); + + // Open input file. + input_ctx.filename = argv[1]; + if (!(input_ctx.file = fopen(input_ctx.filename, "rb"))) { + die("Failed to open %s for reading", argv[1]); + } + + // Open an output file for each stream. + for (i = 0; i < cfg.ts_number_layers; ++i) { + char file_name[PATH_MAX]; + VpxVideoInfo info; + info.codec_fourcc = fourcc; + info.frame_width = cfg.g_w; + info.frame_height = cfg.g_h; + info.time_base.numerator = cfg.g_timebase.num; + info.time_base.denominator = cfg.g_timebase.den; + + snprintf(file_name, sizeof(file_name), "%s_%d.ivf", argv[2], i); + outfile[i] = vpx_video_writer_open(file_name, kContainerIVF, &info); + if (!outfile[i]) + die("Failed to open %s for writing", file_name); + } + // No spatial layers in this encoder. + cfg.ss_number_layers = 1; + + // Initialize codec. + if (vpx_codec_enc_init(&codec, interface(), &cfg, 0)) + die_codec(&codec, "Failed to initialize encoder"); + + vpx_codec_control(&codec, VP8E_SET_CPUUSED, -6); + vpx_codec_control(&codec, VP8E_SET_NOISE_SENSITIVITY, 1); + if (strncmp(codec_type, "vp9", 3) == 0) { + vpx_codec_control(&codec, VP8E_SET_CPUUSED, 3); + vpx_codec_control(&codec, VP8E_SET_NOISE_SENSITIVITY, 0); + if (vpx_codec_control(&codec, VP9E_SET_SVC, 1)) { + die_codec(&codec, "Failed to set SVC"); + } + } + vpx_codec_control(&codec, VP8E_SET_STATIC_THRESHOLD, 1); + vpx_codec_control(&codec, VP8E_SET_TOKEN_PARTITIONS, 1); + max_intra_size_pct = (int) (((double)cfg.rc_buf_optimal_sz * 0.5) + * ((double) cfg.g_timebase.den / cfg.g_timebase.num) / 10.0); + vpx_codec_control(&codec, VP8E_SET_MAX_INTRA_BITRATE_PCT, max_intra_size_pct); + + frame_avail = 1; + while (frame_avail || got_data) { + vpx_codec_iter_t iter = NULL; + const vpx_codec_cx_pkt_t *pkt; + // Update the temporal layer_id. No spatial layers in this test. + layer_id.spatial_layer_id = 0; + layer_id.temporal_layer_id = + cfg.ts_layer_id[frame_cnt % cfg.ts_periodicity]; + vpx_codec_control(&codec, VP9E_SET_SVC_LAYER_ID, &layer_id); + flags = layer_flags[frame_cnt % flag_periodicity]; + frame_avail = !read_yuv_frame(&input_ctx, &raw); + if (vpx_codec_encode(&codec, frame_avail? &raw : NULL, pts, 1, flags, + VPX_DL_REALTIME)) { + die_codec(&codec, "Failed to encode frame"); + } + // Reset KF flag. + if (layering_mode != 7) { + layer_flags[0] &= ~VPX_EFLAG_FORCE_KF; + } + got_data = 0; + while ( (pkt = vpx_codec_get_cx_data(&codec, &iter)) ) { + got_data = 1; + switch (pkt->kind) { + case VPX_CODEC_CX_FRAME_PKT: + for (i = cfg.ts_layer_id[frame_cnt % cfg.ts_periodicity]; + i < cfg.ts_number_layers; ++i) { + vpx_video_writer_write_frame(outfile[i], pkt->data.frame.buf, + pkt->data.frame.sz, pts); + ++frames_in_layer[i]; + } + break; + default: + break; + } + } + ++frame_cnt; + pts += frame_duration; + } + fclose(input_ctx.file); + printf("Processed %d frames: \n", frame_cnt - 1); + if (vpx_codec_destroy(&codec)) + die_codec(&codec, "Failed to destroy codec"); + + // Try to rewrite the output file headers with the actual frame count. + for (i = 0; i < cfg.ts_number_layers; ++i) + vpx_video_writer_close(outfile[i]); + + return EXIT_SUCCESS; +} |