int main(int argc, char **argv) {
FILE *infile = NULL;
vpx_codec_ctx_t codec = {0};
vpx_codec_enc_cfg_t cfg = {0};
int frame_count = 0;
vpx_image_t raw = {0};
vpx_codec_err_t res;
VpxVideoInfo info = {0};
VpxVideoWriter *writer = NULL;
const VpxInterface *encoder = NULL;
const int fps = 2; // 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");
encoder = get_vpx_encoder_by_name("vp8"); // only vp8 for now
if (!encoder)
die("Unsupported codec.");
info.codec_fourcc = encoder->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(encoder->interface()));
res = vpx_codec_enc_config_default(encoder->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, encoder->interface(), &cfg, 0))
die_codec(&codec, "Failed to initialize encoder");
while (vpx_img_read(&raw, infile)) {
++frame_count;
if (frame_count == 22) {
set_roi_map(&cfg, &codec);
} else if (frame_count == 33) {
set_active_map(&cfg, &codec);
} else if (frame_count == 44) {
unset_active_map(&cfg, &codec);
}
encode_frame(&codec, &raw, frame_count, writer);
}
encode_frame(&codec, NULL, -1, writer);
printf("\n");
fclose(infile);
printf("Processed %d frames.\n", frame_count);
vpx_img_free(&raw);
if (vpx_codec_destroy(&codec))
die_codec(&codec, "Failed to destroy codec.");
vpx_video_writer_close(writer);
return EXIT_SUCCESS;
}
int main(int argc, char **argv) {
FILE *infile, *outfile;
char infile_path[] = "vsamples/HIMOTION4_3_30_VGA.yuv";
char outfile_path[] = "vsamples/ivf_test1.ivf";
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=640;
long height=480;
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(outfile_path, "wb")))
die("Failed to open %s for writing", outfile_path);
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; //
} //
/* Update the default configuration with our settings */ // //
cfg.g_w = width; //
cfg.g_h = height;
cfg.rc_target_bitrate = 800;
cfg.rc_buf_sz = 300;
cfg.rc_overshoot_pct = 1000;
cfg.g_pass = VPX_RC_ONE_PASS;
write_ivf_file_header(outfile, &cfg, 0);
/* Open input file for this encoding pass */
if(!(infile = fopen(infile_path, "rb")))
die("Failed to open %s for reading", infile_path);
/* 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);
printf("Processed %d frames.\n",frame_cnt-1);
vpx_img_free(&raw); //
if(vpx_codec_destroy(&codec)) //
die_codec(&codec, "Failed to destroy codec"); //
/* 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;
}
int main_loop(int argc, const char **argv_) {
vpx_codec_ctx_t decoder;
char *fn = NULL;
uint8_t *buf = NULL;
size_t bytes_in_buffer = 0, buffer_size = 0;
FILE *infile;
int frame_in = 0, frame_out = 0, flipuv = 0, noblit = 0;
int do_md5 = 0, progress = 0;
int stop_after = 0, postproc = 0, summary = 0, quiet = 1;
int arg_skip = 0;
int ec_enabled = 0;
const VpxInterface *interface = NULL;
const VpxInterface *fourcc_interface = NULL;
unsigned long dx_time = 0;
struct arg arg;
char **argv, **argi, **argj;
int single_file;
int use_y4m = 1;
vpx_codec_dec_cfg_t cfg = {0};
#if CONFIG_VP8_DECODER
vp8_postproc_cfg_t vp8_pp_cfg = {0};
int vp8_dbg_color_ref_frame = 0;
int vp8_dbg_color_mb_modes = 0;
int vp8_dbg_color_b_modes = 0;
int vp8_dbg_display_mv = 0;
#endif
int frames_corrupted = 0;
int dec_flags = 0;
int do_scale = 0;
vpx_image_t *scaled_img = NULL;
int frame_avail, got_data;
int num_external_frame_buffers = 0;
struct ExternalFrameBufferList ext_fb_list = {0};
const char *outfile_pattern = NULL;
char outfile_name[PATH_MAX] = {0};
FILE *outfile = NULL;
MD5Context md5_ctx;
unsigned char md5_digest[16];
struct VpxDecInputContext input = {0};
struct VpxInputContext vpx_input_ctx = {0};
struct WebmInputContext webm_ctx = {0};
input.vpx_input_ctx = &vpx_input_ctx;
input.webm_ctx = &webm_ctx;
/* Parse command line */
exec_name = argv_[0];
argv = argv_dup(argc - 1, argv_ + 1);
for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step) {
memset(&arg, 0, sizeof(arg));
arg.argv_step = 1;
if (arg_match(&arg, &codecarg, argi)) {
interface = get_vpx_decoder_by_name(arg.val);
if (!interface)
die("Error: Unrecognized argument (%s) to --codec\n", arg.val);
} else if (arg_match(&arg, &looparg, argi)) {
// no-op
} else if (arg_match(&arg, &outputfile, argi))
outfile_pattern = arg.val;
else if (arg_match(&arg, &use_yv12, argi)) {
use_y4m = 0;
flipuv = 1;
} else if (arg_match(&arg, &use_i420, argi)) {
use_y4m = 0;
flipuv = 0;
} else if (arg_match(&arg, &flipuvarg, argi))
flipuv = 1;
else if (arg_match(&arg, &noblitarg, argi))
noblit = 1;
else if (arg_match(&arg, &progressarg, argi))
progress = 1;
else if (arg_match(&arg, &limitarg, argi))
stop_after = arg_parse_uint(&arg);
else if (arg_match(&arg, &skiparg, argi))
arg_skip = arg_parse_uint(&arg);
else if (arg_match(&arg, &postprocarg, argi))
postproc = 1;
else if (arg_match(&arg, &md5arg, argi))
do_md5 = 1;
else if (arg_match(&arg, &summaryarg, argi))
summary = 1;
else if (arg_match(&arg, &threadsarg, argi))
cfg.threads = arg_parse_uint(&arg);
else if (arg_match(&arg, &verbosearg, argi))
quiet = 0;
else if (arg_match(&arg, &scalearg, argi))
do_scale = 1;
else if (arg_match(&arg, &fb_arg, argi))
num_external_frame_buffers = arg_parse_uint(&arg);
#if CONFIG_VP8_DECODER
else if (arg_match(&arg, &addnoise_level, argi)) {
postproc = 1;
vp8_pp_cfg.post_proc_flag |= VP8_ADDNOISE;
vp8_pp_cfg.noise_level = arg_parse_uint(&arg);
} else if (arg_match(&arg, &demacroblock_level, argi)) {
postproc = 1;
vp8_pp_cfg.post_proc_flag |= VP8_DEMACROBLOCK;
vp8_pp_cfg.deblocking_level = arg_parse_uint(&arg);
} else if (arg_match(&arg, &deblock, argi)) {
postproc = 1;
vp8_pp_cfg.post_proc_flag |= VP8_DEBLOCK;
} else if (arg_match(&arg, &mfqe, argi)) {
postproc = 1;
vp8_pp_cfg.post_proc_flag |= VP8_MFQE;
} else if (arg_match(&arg, &pp_debug_info, argi)) {
unsigned int level = arg_parse_uint(&arg);
postproc = 1;
vp8_pp_cfg.post_proc_flag &= ~0x7;
if (level)
vp8_pp_cfg.post_proc_flag |= level;
} else if (arg_match(&arg, &pp_disp_ref_frame, argi)) {
unsigned int flags = arg_parse_int(&arg);
if (flags) {
postproc = 1;
vp8_dbg_color_ref_frame = flags;
}
} else if (arg_match(&arg, &pp_disp_mb_modes, argi)) {
unsigned int flags = arg_parse_int(&arg);
if (flags) {
postproc = 1;
vp8_dbg_color_mb_modes = flags;
}
} else if (arg_match(&arg, &pp_disp_b_modes, argi)) {
unsigned int flags = arg_parse_int(&arg);
if (flags) {
postproc = 1;
vp8_dbg_color_b_modes = flags;
}
} else if (arg_match(&arg, &pp_disp_mvs, argi)) {
unsigned int flags = arg_parse_int(&arg);
if (flags) {
postproc = 1;
vp8_dbg_display_mv = flags;
}
} else if (arg_match(&arg, &error_concealment, argi)) {
ec_enabled = 1;
}
#endif
else
argj++;
}
/* Check for unrecognized options */
for (argi = argv; *argi; argi++)
if (argi[0][0] == '-' && strlen(argi[0]) > 1)
die("Error: Unrecognized option %s\n", *argi);
/* Handle non-option arguments */
fn = argv[0];
if (!fn)
usage_exit();
/* Open file */
infile = strcmp(fn, "-") ? fopen(fn, "rb") : set_binary_mode(stdin);
if (!infile) {
fprintf(stderr, "Failed to open file '%s'", strcmp(fn, "-") ? fn : "stdin");
return EXIT_FAILURE;
}
#if CONFIG_OS_SUPPORT
/* Make sure we don't dump to the terminal, unless forced to with -o - */
if (!outfile_pattern && isatty(fileno(stdout)) && !do_md5 && !noblit) {
fprintf(stderr,
"Not dumping raw video to your terminal. Use '-o -' to "
"override.\n");
return EXIT_FAILURE;
}
#endif
input.vpx_input_ctx->file = infile;
if (file_is_ivf(input.vpx_input_ctx))
input.vpx_input_ctx->file_type = FILE_TYPE_IVF;
else if (file_is_webm(input.webm_ctx, input.vpx_input_ctx))
input.vpx_input_ctx->file_type = FILE_TYPE_WEBM;
else if (file_is_raw(input.vpx_input_ctx))
input.vpx_input_ctx->file_type = FILE_TYPE_RAW;
else {
fprintf(stderr, "Unrecognized input file type.\n");
return EXIT_FAILURE;
}
outfile_pattern = outfile_pattern ? outfile_pattern : "-";
single_file = is_single_file(outfile_pattern);
if (!noblit && single_file) {
generate_filename(outfile_pattern, outfile_name, PATH_MAX,
vpx_input_ctx.width, vpx_input_ctx.height, 0);
if (do_md5)
MD5Init(&md5_ctx);
else
outfile = open_outfile(outfile_name);
}
if (use_y4m && !noblit) {
if (!single_file) {
fprintf(stderr, "YUV4MPEG2 not supported with output patterns,"
" try --i420 or --yv12.\n");
return EXIT_FAILURE;
}
if (vpx_input_ctx.file_type == FILE_TYPE_WEBM) {
if (webm_guess_framerate(input.webm_ctx, input.vpx_input_ctx)) {
fprintf(stderr, "Failed to guess framerate -- error parsing "
"webm file?\n");
return EXIT_FAILURE;
}
}
}
fourcc_interface = get_vpx_decoder_by_fourcc(vpx_input_ctx.fourcc);
if (interface && fourcc_interface && interface != fourcc_interface)
warn("Header indicates codec: %s\n", fourcc_interface->name);
else
interface = fourcc_interface;
if (!interface)
interface = get_vpx_decoder_by_index(0);
dec_flags = (postproc ? VPX_CODEC_USE_POSTPROC : 0) |
(ec_enabled ? VPX_CODEC_USE_ERROR_CONCEALMENT : 0);
if (vpx_codec_dec_init(&decoder, interface->interface(), &cfg, dec_flags)) {
fprintf(stderr, "Failed to initialize decoder: %s\n",
vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
if (!quiet)
fprintf(stderr, "%s\n", decoder.name);
#if CONFIG_VP8_DECODER
if (vp8_pp_cfg.post_proc_flag
&& vpx_codec_control(&decoder, VP8_SET_POSTPROC, &vp8_pp_cfg)) {
fprintf(stderr, "Failed to configure postproc: %s\n", vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
if (vp8_dbg_color_ref_frame
&& vpx_codec_control(&decoder, VP8_SET_DBG_COLOR_REF_FRAME, vp8_dbg_color_ref_frame)) {
fprintf(stderr, "Failed to configure reference block visualizer: %s\n", vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
if (vp8_dbg_color_mb_modes
&& vpx_codec_control(&decoder, VP8_SET_DBG_COLOR_MB_MODES, vp8_dbg_color_mb_modes)) {
fprintf(stderr, "Failed to configure macro block visualizer: %s\n", vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
if (vp8_dbg_color_b_modes
&& vpx_codec_control(&decoder, VP8_SET_DBG_COLOR_B_MODES, vp8_dbg_color_b_modes)) {
fprintf(stderr, "Failed to configure block visualizer: %s\n", vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
if (vp8_dbg_display_mv
&& vpx_codec_control(&decoder, VP8_SET_DBG_DISPLAY_MV, vp8_dbg_display_mv)) {
fprintf(stderr, "Failed to configure motion vector visualizer: %s\n", vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
#endif
if (arg_skip)
fprintf(stderr, "Skipping first %d frames.\n", arg_skip);
while (arg_skip) {
if (read_frame(&input, &buf, &bytes_in_buffer, &buffer_size))
break;
arg_skip--;
}
if (num_external_frame_buffers > 0) {
ext_fb_list.num_external_frame_buffers = num_external_frame_buffers;
ext_fb_list.ext_fb = (struct ExternalFrameBuffer *)calloc(
num_external_frame_buffers, sizeof(*ext_fb_list.ext_fb));
if (vpx_codec_set_frame_buffer_functions(
&decoder, get_vp9_frame_buffer, release_vp9_frame_buffer,
&ext_fb_list)) {
fprintf(stderr, "Failed to configure external frame buffers: %s\n",
vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
}
frame_avail = 1;
got_data = 0;
/* Decode file */
while (frame_avail || got_data) {
vpx_codec_iter_t iter = NULL;
vpx_image_t *img;
struct vpx_usec_timer timer;
int corrupted;
frame_avail = 0;
if (!stop_after || frame_in < stop_after) {
if (!read_frame(&input, &buf, &bytes_in_buffer, &buffer_size)) {
frame_avail = 1;
frame_in++;
vpx_usec_timer_start(&timer);
if (vpx_codec_decode(&decoder, buf, bytes_in_buffer, NULL, 0)) {
const char *detail = vpx_codec_error_detail(&decoder);
warn("Failed to decode frame %d: %s",
frame_in, vpx_codec_error(&decoder));
if (detail)
warn("Additional information: %s", detail);
goto fail;
}
vpx_usec_timer_mark(&timer);
dx_time += (unsigned int)vpx_usec_timer_elapsed(&timer);
}
}
vpx_usec_timer_start(&timer);
got_data = 0;
if ((img = vpx_codec_get_frame(&decoder, &iter))) {
++frame_out;
got_data = 1;
}
vpx_usec_timer_mark(&timer);
dx_time += (unsigned int)vpx_usec_timer_elapsed(&timer);
if (vpx_codec_control(&decoder, VP8D_GET_FRAME_CORRUPTED, &corrupted)) {
warn("Failed VP8_GET_FRAME_CORRUPTED: %s", vpx_codec_error(&decoder));
goto fail;
}
frames_corrupted += corrupted;
if (progress)
show_progress(frame_in, frame_out, dx_time);
if (!noblit && img) {
const int PLANES_YUV[] = {VPX_PLANE_Y, VPX_PLANE_U, VPX_PLANE_V};
const int PLANES_YVU[] = {VPX_PLANE_Y, VPX_PLANE_V, VPX_PLANE_U};
const int *planes = flipuv ? PLANES_YVU : PLANES_YUV;
if (do_scale) {
if (frame_out == 1) {
// If the output frames are to be scaled to a fixed display size then
// use the width and height specified in the container. If either of
// these is set to 0, use the display size set in the first frame
// header. If that is unavailable, use the raw decoded size of the
// first decoded frame.
int display_width = vpx_input_ctx.width;
int display_height = vpx_input_ctx.height;
if (!display_width || !display_height) {
int display_size[2];
if (vpx_codec_control(&decoder, VP9D_GET_DISPLAY_SIZE,
display_size)) {
// As last resort use size of first frame as display size.
display_width = img->d_w;
display_height = img->d_h;
} else {
display_width = display_size[0];
display_height = display_size[1];
}
}
scaled_img = vpx_img_alloc(NULL, VPX_IMG_FMT_I420, display_width,
display_height, 16);
}
if (img->d_w != scaled_img->d_w || img->d_h != scaled_img->d_h) {
vpx_image_scale(img, scaled_img, kFilterBox);
img = scaled_img;
}
}
if (single_file) {
if (use_y4m) {
char buf[Y4M_BUFFER_SIZE] = {0};
size_t len = 0;
if (frame_out == 1) {
// Y4M file header
len = y4m_write_file_header(buf, sizeof(buf),
vpx_input_ctx.width,
vpx_input_ctx.height,
&vpx_input_ctx.framerate, img->fmt);
if (do_md5) {
MD5Update(&md5_ctx, (md5byte *)buf, len);
} else {
fputs(buf, outfile);
}
}
// Y4M frame header
len = y4m_write_frame_header(buf, sizeof(buf));
if (do_md5) {
MD5Update(&md5_ctx, (md5byte *)buf, len);
} else {
fputs(buf, outfile);
}
}
if (do_md5) {
update_image_md5(img, planes, &md5_ctx);
} else {
write_image_file(img, planes, outfile);
}
} else {
generate_filename(outfile_pattern, outfile_name, PATH_MAX,
img->d_w, img->d_h, frame_in);
if (do_md5) {
MD5Init(&md5_ctx);
update_image_md5(img, planes, &md5_ctx);
MD5Final(md5_digest, &md5_ctx);
print_md5(md5_digest, outfile_name);
} else {
outfile = open_outfile(outfile_name);
write_image_file(img, planes, outfile);
fclose(outfile);
}
}
}
if (stop_after && frame_in >= stop_after)
break;
}
if (summary || progress) {
show_progress(frame_in, frame_out, dx_time);
fprintf(stderr, "\n");
}
if (frames_corrupted)
fprintf(stderr, "WARNING: %d frames corrupted.\n", frames_corrupted);
fail:
if (vpx_codec_destroy(&decoder)) {
fprintf(stderr, "Failed to destroy decoder: %s\n",
vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
if (!noblit && single_file) {
if (do_md5) {
MD5Final(md5_digest, &md5_ctx);
print_md5(md5_digest, outfile_name);
} else {
fclose(outfile);
}
}
if (input.vpx_input_ctx->file_type == FILE_TYPE_WEBM)
webm_free(input.webm_ctx);
else
free(buf);
if (scaled_img) vpx_img_free(scaled_img);
for (i = 0; i < ext_fb_list.num_external_frame_buffers; ++i) {
free(ext_fb_list.ext_fb[i].data);
}
free(ext_fb_list.ext_fb);
fclose(infile);
free(argv);
return frames_corrupted ? EXIT_FAILURE : EXIT_SUCCESS;
}
// TODO(tomfinegan): Improve command line parsing and add args for bitrate/fps.
int main(int argc, char **argv) {
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, 0, 0, { 0, 0 } };
VpxVideoWriter *writer = NULL;
const VpxInterface *encoder = NULL;
const int fps = 30;
const int bitrate = 200;
int keyframe_interval = 0;
int max_frames = 0;
int frames_encoded = 0;
const char *codec_arg = NULL;
const char *width_arg = NULL;
const char *height_arg = NULL;
const char *infile_arg = NULL;
const char *outfile_arg = NULL;
const char *keyframe_interval_arg = NULL;
exec_name = argv[0];
if (argc != 9) die("Invalid number of arguments");
codec_arg = argv[1];
width_arg = argv[2];
height_arg = argv[3];
infile_arg = argv[4];
outfile_arg = argv[5];
keyframe_interval_arg = argv[6];
max_frames = (int)strtol(argv[8], NULL, 0);
encoder = get_vpx_encoder_by_name(codec_arg);
if (!encoder) die("Unsupported codec.");
info.codec_fourcc = encoder->fourcc;
info.frame_width = (int)strtol(width_arg, NULL, 0);
info.frame_height = (int)strtol(height_arg, 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.");
}
keyframe_interval = (int)strtol(keyframe_interval_arg, NULL, 0);
if (keyframe_interval < 0) die("Invalid keyframe interval value.");
printf("Using %s\n", vpx_codec_iface_name(encoder->codec_interface()));
res = vpx_codec_enc_config_default(encoder->codec_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;
cfg.g_error_resilient = (vpx_codec_er_flags_t)strtoul(argv[7], NULL, 0);
writer = vpx_video_writer_open(outfile_arg, kContainerIVF, &info);
if (!writer) die("Failed to open %s for writing.", outfile_arg);
if (!(infile = fopen(infile_arg, "rb")))
die("Failed to open %s for reading.", infile_arg);
if (vpx_codec_enc_init(&codec, encoder->codec_interface(), &cfg, 0))
die_codec(&codec, "Failed to initialize encoder");
// Encode frames.
while (vpx_img_read(&raw, infile)) {
int flags = 0;
if (keyframe_interval > 0 && frame_count % keyframe_interval == 0)
flags |= VPX_EFLAG_FORCE_KF;
encode_frame(&codec, &raw, frame_count++, flags, writer);
frames_encoded++;
if (max_frames > 0 && frames_encoded >= max_frames) break;
}
// Flush encoder.
while (encode_frame(&codec, NULL, -1, 0, writer)) {
}
printf("\n");
fclose(infile);
printf("Processed %d frames.\n", frame_count);
vpx_img_free(&raw);
if (vpx_codec_destroy(&codec)) die_codec(&codec, "Failed to destroy codec.");
vpx_video_writer_close(writer);
return EXIT_SUCCESS;
}
int main(int argc, char **argv) {
FILE *infile = NULL;
// Encoder
vpx_codec_ctx_t ecodec = {0};
vpx_codec_enc_cfg_t cfg = {0};
unsigned int frame_in = 0;
vpx_image_t raw;
vpx_codec_err_t res;
VpxVideoInfo info = {0};
VpxVideoWriter *writer = NULL;
const VpxInterface *encoder = NULL;
// Test encoder/decoder mismatch.
int test_decode = 1;
// Decoder
vpx_codec_ctx_t dcodec;
unsigned int frame_out = 0;
// The frame number to set reference frame on
unsigned int update_frame_num = 0;
int mismatch_seen = 0;
const int fps = 30;
const int bitrate = 500;
const char *width_arg = NULL;
const char *height_arg = NULL;
const char *infile_arg = NULL;
const char *outfile_arg = NULL;
unsigned int limit = 0;
exec_name = argv[0];
if (argc < 6)
die("Invalid number of arguments");
width_arg = argv[1];
height_arg = argv[2];
infile_arg = argv[3];
outfile_arg = argv[4];
encoder = get_vpx_encoder_by_name("vp9");
if (!encoder)
die("Unsupported codec.");
update_frame_num = atoi(argv[5]);
// In VP9, the reference buffers (cm->buffer_pool->frame_bufs[i].buf) are
// allocated while calling vpx_codec_encode(), thus, setting reference for
// 1st frame isn't supported.
if (update_frame_num <= 1)
die("Couldn't parse frame number '%s'\n", argv[5]);
if (argc > 6) {
limit = atoi(argv[6]);
if (update_frame_num > limit)
die("Update frame number couldn't larger than limit\n");
}
info.codec_fourcc = encoder->fourcc;
info.frame_width = strtol(width_arg, NULL, 0);
info.frame_height = strtol(height_arg, 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(encoder->codec_interface()));
res = vpx_codec_enc_config_default(encoder->codec_interface(), &cfg, 0);
if (res)
die_codec(&ecodec, "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;
cfg.g_lag_in_frames = 3;
writer = vpx_video_writer_open(outfile_arg, kContainerIVF, &info);
if (!writer)
die("Failed to open %s for writing.", outfile_arg);
if (!(infile = fopen(infile_arg, "rb")))
die("Failed to open %s for reading.", infile_arg);
if (vpx_codec_enc_init(&ecodec, encoder->codec_interface(), &cfg, 0))
die_codec(&ecodec, "Failed to initialize encoder");
// Disable alt_ref.
if (vpx_codec_control(&ecodec, VP8E_SET_ENABLEAUTOALTREF, 0))
die_codec(&ecodec, "Failed to set enable auto alt ref");
if (test_decode) {
const VpxInterface *decoder = get_vpx_decoder_by_name("vp9");
if (vpx_codec_dec_init(&dcodec, decoder->codec_interface(), NULL, 0))
die_codec(&dcodec, "Failed to initialize decoder.");
}
// Encode frames.
while (vpx_img_read(&raw, infile)) {
if (limit && frame_in >= limit)
break;
if (update_frame_num > 1 && frame_out + 1 == update_frame_num) {
vpx_ref_frame_t ref;
ref.frame_type = VP8_LAST_FRAME;
ref.img = raw;
// Set reference frame in encoder.
if (vpx_codec_control(&ecodec, VP8_SET_REFERENCE, &ref))
die_codec(&ecodec, "Failed to set reference frame");
printf(" <SET_REF>");
// If set_reference in decoder is commented out, the enc/dec mismatch
// would be seen.
if (test_decode) {
if (vpx_codec_control(&dcodec, VP8_SET_REFERENCE, &ref))
die_codec(&dcodec, "Failed to set reference frame");
}
}
encode_frame(&ecodec, &cfg, &raw, frame_in, writer, test_decode,
&dcodec, &frame_out, &mismatch_seen);
frame_in++;
if (mismatch_seen)
break;
}
// Flush encoder.
if (!mismatch_seen)
while (encode_frame(&ecodec, &cfg, NULL, frame_in, writer, test_decode,
&dcodec, &frame_out, &mismatch_seen)) {}
printf("\n");
fclose(infile);
printf("Processed %d frames.\n", frame_out);
if (test_decode) {
if (!mismatch_seen)
printf("Encoder/decoder results are matching.\n");
else
printf("Encoder/decoder results are NOT matching.\n");
}
if (test_decode)
if (vpx_codec_destroy(&dcodec))
die_codec(&dcodec, "Failed to destroy decoder");
vpx_img_free(&raw);
if (vpx_codec_destroy(&ecodec))
die_codec(&ecodec, "Failed to destroy encoder.");
vpx_video_writer_close(writer);
return EXIT_SUCCESS;
}
int main(int argc, char **argv)
{
FILE *infile, *outfile[NUM_ENCODERS];
vpx_codec_ctx_t codec[NUM_ENCODERS];
vpx_codec_enc_cfg_t cfg[NUM_ENCODERS];
vpx_codec_pts_t frame_cnt = 0;
vpx_image_t raw[NUM_ENCODERS];
vpx_codec_err_t res[NUM_ENCODERS];
int i;
long width;
long height;
int frame_avail;
int got_data;
int flags = 0;
/*Currently, only realtime mode is supported in multi-resolution encoding.*/
int arg_deadline = VPX_DL_REALTIME;
/* Set show_psnr to 1/0 to show/not show PSNR. Choose show_psnr=0 if you
don't need to know PSNR, which will skip PSNR calculation and save
encoding time. */
int show_psnr = 0;
uint64_t psnr_sse_total[NUM_ENCODERS] = {0};
uint64_t psnr_samples_total[NUM_ENCODERS] = {0};
double psnr_totals[NUM_ENCODERS][4] = {{0,0}};
int psnr_count[NUM_ENCODERS] = {0};
/* Set the required target bitrates for each resolution level. */
unsigned int target_bitrate[NUM_ENCODERS]={1400, 500, 100};
/* Enter the frame rate of the input video */
int framerate = 30;
/* Set down-sampling factor for each resolution level.
dsf[0] controls down sampling from level 0 to level 1;
dsf[1] controls down sampling from level 1 to level 2;
dsf[2] is not used. */
vpx_rational_t dsf[NUM_ENCODERS] = {{2, 1}, {2, 1}, {1, 1}};
if(argc!= (5+NUM_ENCODERS))
die("Usage: %s <width> <height> <infile> <outfile(s)> <output psnr?>\n",
argv[0]);
printf("Using %s\n",vpx_codec_iface_name(interface));
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);
/* Open input video file for encoding */
if(!(infile = fopen(argv[3], "rb")))
die("Failed to open %s for reading", argv[3]);
/* Open output file for each encoder to output bitstreams */
for (i=0; i< NUM_ENCODERS; i++)
{
if(!(outfile[i] = fopen(argv[i+4], "wb")))
die("Failed to open %s for writing", argv[i+4]);
}
show_psnr = strtol(argv[NUM_ENCODERS + 4], NULL, 0);
/* Populate default encoder configuration */
for (i=0; i< NUM_ENCODERS; i++)
{
res[i] = vpx_codec_enc_config_default(interface, &cfg[i], 0);
if(res[i]) {
printf("Failed to get config: %s\n", vpx_codec_err_to_string(res[i]));
return EXIT_FAILURE;
}
}
/*
* Update the default configuration according to needs of the application.
*/
/* Highest-resolution encoder settings */
cfg[0].g_w = width;
cfg[0].g_h = height;
cfg[0].g_threads = 1; /* number of threads used */
cfg[0].rc_dropframe_thresh = 0;
cfg[0].rc_end_usage = VPX_CBR;
cfg[0].rc_resize_allowed = 0;
cfg[0].rc_min_quantizer = 4;
cfg[0].rc_max_quantizer = 56;
cfg[0].rc_undershoot_pct = 98;
cfg[0].rc_overshoot_pct = 100;
cfg[0].rc_buf_initial_sz = 500;
cfg[0].rc_buf_optimal_sz = 600;
cfg[0].rc_buf_sz = 1000;
//cfg[0].rc_dropframe_thresh = 10;
cfg[0].g_error_resilient = 1; /* Enable error resilient mode */
cfg[0].g_lag_in_frames = 0;
/* Disable automatic keyframe placement */
//cfg[0].kf_mode = VPX_KF_DISABLED;
cfg[0].kf_min_dist = cfg[0].kf_max_dist = 1000;
cfg[0].rc_target_bitrate = target_bitrate[0]; /* Set target bitrate */
cfg[0].g_timebase.num = 1; /* Set fps */
cfg[0].g_timebase.den = framerate;
/* Other-resolution encoder settings */
for (i=1; i< NUM_ENCODERS; i++)
{
memcpy(&cfg[i], &cfg[0], sizeof(vpx_codec_enc_cfg_t));
cfg[i].g_threads = 1; /* number of threads used */
cfg[i].rc_target_bitrate = target_bitrate[i];
/* Note: Width & height of other-resolution encoders are calculated
* from the highest-resolution encoder's size and the corresponding
* down_sampling_factor.
*/
{
unsigned int iw = cfg[i-1].g_w*dsf[i-1].den + dsf[i-1].num - 1;
unsigned int ih = cfg[i-1].g_h*dsf[i-1].den + dsf[i-1].num - 1;
cfg[i].g_w = iw/dsf[i-1].num;
cfg[i].g_h = ih/dsf[i-1].num;
}
/* Make width & height to be multiplier of 2. */
// Should support odd size ???
if((cfg[i].g_w)%2)cfg[i].g_w++;
if((cfg[i].g_h)%2)cfg[i].g_h++;
}
/* Allocate image for each encoder */
for (i=0; i< NUM_ENCODERS; i++)
if(!vpx_img_alloc(&raw[i], VPX_IMG_FMT_I420, cfg[i].g_w, cfg[i].g_h, 32))
die("Failed to allocate image", cfg[i].g_w, cfg[i].g_h);
if (raw[0].stride[VPX_PLANE_Y] == raw[0].d_w)
read_frame_p = read_frame;
else
read_frame_p = read_frame_by_row;
for (i=0; i< NUM_ENCODERS; i++)
write_ivf_file_header(outfile[i], &cfg[i], 0);
/* Initialize multi-encoder */
if(vpx_codec_enc_init_multi(&codec[0], interface, &cfg[0], NUM_ENCODERS,
(show_psnr ? VPX_CODEC_USE_PSNR : 0), &dsf[0]))
die_codec(&codec[0], "Failed to initialize encoder");
/* The extra encoding configuration parameters can be set as follows. */
/* Set encoding speed */
for ( i=0; i<NUM_ENCODERS; i++)
{
int speed = -6;
if(vpx_codec_control(&codec[i], VP8E_SET_CPUUSED, speed))
die_codec(&codec[i], "Failed to set cpu_used");
}
/* Set static thresh for highest-resolution encoder. Set it to 1000 for
* better performance. */
{
unsigned int static_thresh = 1000;
if(vpx_codec_control(&codec[0], VP8E_SET_STATIC_THRESHOLD, static_thresh))
die_codec(&codec[0], "Failed to set static threshold");
}
/* Set static thresh = 0 for other encoders for better quality */
for ( i=1; i<NUM_ENCODERS; i++)
{
unsigned int static_thresh = 0;
if(vpx_codec_control(&codec[i], VP8E_SET_STATIC_THRESHOLD, static_thresh))
die_codec(&codec[i], "Failed to set static threshold");
}
frame_avail = 1;
got_data = 0;
while(frame_avail || got_data)
{
vpx_codec_iter_t iter[NUM_ENCODERS]={NULL};
const vpx_codec_cx_pkt_t *pkt[NUM_ENCODERS];
flags = 0;
frame_avail = read_frame_p(infile, &raw[0]);
if(frame_avail)
{
for ( i=1; i<NUM_ENCODERS; i++)
{
/*Scale the image down a number of times by downsampling factor*/
/* FilterMode 1 or 2 give better psnr than FilterMode 0. */
I420Scale(raw[i-1].planes[VPX_PLANE_Y], raw[i-1].stride[VPX_PLANE_Y],
raw[i-1].planes[VPX_PLANE_U], raw[i-1].stride[VPX_PLANE_U],
raw[i-1].planes[VPX_PLANE_V], raw[i-1].stride[VPX_PLANE_V],
raw[i-1].d_w, raw[i-1].d_h,
raw[i].planes[VPX_PLANE_Y], raw[i].stride[VPX_PLANE_Y],
raw[i].planes[VPX_PLANE_U], raw[i].stride[VPX_PLANE_U],
raw[i].planes[VPX_PLANE_V], raw[i].stride[VPX_PLANE_V],
raw[i].d_w, raw[i].d_h, 1);
}
}
/* Encode each frame at multi-levels */
if(vpx_codec_encode(&codec[0], frame_avail? &raw[0] : NULL,
frame_cnt, 1, flags, arg_deadline))
die_codec(&codec[0], "Failed to encode frame");
for (i=NUM_ENCODERS-1; i>=0 ; i--)
{
got_data = 0;
while( (pkt[i] = vpx_codec_get_cx_data(&codec[i], &iter[i])) )
{
got_data = 1;
switch(pkt[i]->kind) {
case VPX_CODEC_CX_FRAME_PKT:
write_ivf_frame_header(outfile[i], pkt[i]);
if(fwrite(pkt[i]->data.frame.buf, 1, pkt[i]->data.frame.sz,
outfile[i]));
break;
case VPX_CODEC_PSNR_PKT:
if (show_psnr)
{
int j;
psnr_sse_total[i] += pkt[i]->data.psnr.sse[0];
psnr_samples_total[i] += pkt[i]->data.psnr.samples[0];
for (j = 0; j < 4; j++)
{
//fprintf(stderr, "%.3lf ", pkt[i]->data.psnr.psnr[j]);
psnr_totals[i][j] += pkt[i]->data.psnr.psnr[j];
}
psnr_count[i]++;
}
break;
default:
break;
}
printf(pkt[i]->kind == VPX_CODEC_CX_FRAME_PKT
&& (pkt[i]->data.frame.flags & VPX_FRAME_IS_KEY)? "K":".");
fflush(stdout);
}
}
frame_cnt++;
}
printf("\n");
fclose(infile);
for (i=0; i< NUM_ENCODERS; i++)
{
printf("Processed %ld frames.\n",(long int)frame_cnt-1);
/* Calculate PSNR and print it out */
if ( (show_psnr) && (psnr_count[i]>0) )
{
int j;
double ovpsnr = vp8_mse2psnr(psnr_samples_total[i], 255.0,
psnr_sse_total[i]);
fprintf(stderr, "\n ENC%d PSNR (Overall/Avg/Y/U/V)", i);
fprintf(stderr, " %.3lf", ovpsnr);
for (j = 0; j < 4; j++)
{
fprintf(stderr, " %.3lf", psnr_totals[i][j]/psnr_count[i]);
}
}
if(vpx_codec_destroy(&codec[i]))
die_codec(&codec[i], "Failed to destroy codec");
/* Try to rewrite the file header with the actual frame count */
if(!fseek(outfile[i], 0, SEEK_SET))
write_ivf_file_header(outfile[i], &cfg[i], frame_cnt-1);
fclose(outfile[i]);
vpx_img_free(&raw[i]);
}
return EXIT_SUCCESS;
}
int main(int argc, char **argv) {
int frame_cnt = 0;
FILE *outfile = NULL;
vpx_codec_ctx_t codec;
vpx_codec_err_t res;
VpxVideoReader *reader = NULL;
const VpxInterface *decoder = NULL;
const VpxVideoInfo *info = NULL;
exec_name = argv[0];
if (argc != 3) die("Invalid number of arguments.");
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]);
info = vpx_video_reader_get_info(reader);
decoder = get_vpx_decoder_by_fourcc(info->codec_fourcc);
if (!decoder) die("Unknown input codec.");
printf("Using %s\n", vpx_codec_iface_name(decoder->codec_interface()));
res = vpx_codec_dec_init(&codec, decoder->codec_interface(), NULL,
VPX_CODEC_USE_POSTPROC);
if (res == VPX_CODEC_INCAPABLE)
die_codec(&codec, "Postproc not supported by this decoder.");
if (res) die_codec(&codec, "Failed to initialize decoder.");
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_reader_get_frame(reader, &frame_size);
++frame_cnt;
if (frame_cnt % 30 == 1) {
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.");
} 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.");
};
// Decode the frame with 15ms deadline
if (vpx_codec_decode(&codec, frame, (unsigned int)frame_size, NULL, 15000))
die_codec(&codec, "Failed to decode frame");
while ((img = vpx_codec_get_frame(&codec, &iter)) != NULL) {
vpx_img_write(img, outfile);
}
}
printf("Processed %d frames.\n", frame_cnt);
if (vpx_codec_destroy(&codec)) die_codec(&codec, "Failed to destroy codec");
printf("Play: ffplay -f rawvideo -pix_fmt yuv420p -s %dx%d %s\n",
info->frame_width, info->frame_height, argv[2]);
vpx_video_reader_close(reader);
fclose(outfile);
return EXIT_SUCCESS;
}
static void enc_postprocess(MSFilter *f) {
EncState *s=(EncState*)f->data;
if (s->ready) vpx_codec_destroy(&s->codec);
s->ready=FALSE;
}
int main(int argc, const char **argv_)
{
vpx_codec_ctx_t encoder;
const char *in_fn = NULL, *out_fn = NULL, *stats_fn = NULL;
int i;
FILE *infile, *outfile;
vpx_codec_enc_cfg_t cfg;
vpx_codec_err_t res;
int pass, one_pass_only = 0;
stats_io_t stats;
vpx_image_t raw;
const struct codec_item *codec = codecs;
int frame_avail, got_data;
struct arg arg;
char **argv, **argi, **argj;
int arg_usage = 0, arg_passes = 1, arg_deadline = 0;
int arg_ctrls[ARG_CTRL_CNT_MAX][2], arg_ctrl_cnt = 0;
int arg_limit = 0;
static const arg_def_t **ctrl_args = no_args;
static const int *ctrl_args_map = NULL;
int verbose = 0, show_psnr = 0;
int arg_use_i420 = 1;
int arg_have_timebase = 0;
unsigned long cx_time = 0;
unsigned int file_type, fourcc;
y4m_input y4m;
exec_name = argv_[0];
if (argc < 3)
usage_exit();
/* First parse the codec and usage values, because we want to apply other
* parameters on top of the default configuration provided by the codec.
*/
argv = argv_dup(argc - 1, argv_ + 1);
for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step)
{
arg.argv_step = 1;
if (arg_match(&arg, &codecarg, argi))
{
int j, k = -1;
for (j = 0; j < sizeof(codecs) / sizeof(codecs[0]); j++)
if (!strcmp(codecs[j].name, arg.val))
k = j;
if (k >= 0)
codec = codecs + k;
else
die("Error: Unrecognized argument (%s) to --codec\n",
arg.val);
}
else if (arg_match(&arg, &passes, argi))
{
arg_passes = arg_parse_uint(&arg);
if (arg_passes < 1 || arg_passes > 2)
die("Error: Invalid number of passes (%d)\n", arg_passes);
}
else if (arg_match(&arg, &pass_arg, argi))
{
one_pass_only = arg_parse_uint(&arg);
if (one_pass_only < 1 || one_pass_only > 2)
die("Error: Invalid pass selected (%d)\n", one_pass_only);
}
else if (arg_match(&arg, &fpf_name, argi))
stats_fn = arg.val;
else if (arg_match(&arg, &usage, argi))
arg_usage = arg_parse_uint(&arg);
else if (arg_match(&arg, &deadline, argi))
arg_deadline = arg_parse_uint(&arg);
else if (arg_match(&arg, &best_dl, argi))
arg_deadline = VPX_DL_BEST_QUALITY;
else if (arg_match(&arg, &good_dl, argi))
arg_deadline = VPX_DL_GOOD_QUALITY;
else if (arg_match(&arg, &rt_dl, argi))
arg_deadline = VPX_DL_REALTIME;
else if (arg_match(&arg, &use_yv12, argi))
{
arg_use_i420 = 0;
}
else if (arg_match(&arg, &use_i420, argi))
{
arg_use_i420 = 1;
}
else if (arg_match(&arg, &verbosearg, argi))
verbose = 1;
else if (arg_match(&arg, &limit, argi))
arg_limit = arg_parse_uint(&arg);
else if (arg_match(&arg, &psnrarg, argi))
show_psnr = 1;
else
argj++;
}
/* Ensure that --passes and --pass are consistent. If --pass is set and --passes=2,
* ensure --fpf was set.
*/
if (one_pass_only)
{
/* DWIM: Assume the user meant passes=2 if pass=2 is specified */
if (one_pass_only > arg_passes)
{
fprintf(stderr, "Warning: Assuming --pass=%d implies --passes=%d\n",
one_pass_only, one_pass_only);
arg_passes = one_pass_only;
}
if (arg_passes == 2 && !stats_fn)
die("Must specify --fpf when --pass=%d and --passes=2\n", one_pass_only);
}
/* Populate encoder configuration */
res = vpx_codec_enc_config_default(codec->iface, &cfg, arg_usage);
if (res)
{
fprintf(stderr, "Failed to get config: %s\n",
vpx_codec_err_to_string(res));
return EXIT_FAILURE;
}
/* Now parse the remainder of the parameters. */
for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step)
{
arg.argv_step = 1;
if (0);
else if (arg_match(&arg, &threads, argi))
cfg.g_threads = arg_parse_uint(&arg);
else if (arg_match(&arg, &profile, argi))
cfg.g_profile = arg_parse_uint(&arg);
else if (arg_match(&arg, &width, argi))
cfg.g_w = arg_parse_uint(&arg);
else if (arg_match(&arg, &height, argi))
cfg.g_h = arg_parse_uint(&arg);
else if (arg_match(&arg, &timebase, argi))
{
cfg.g_timebase = arg_parse_rational(&arg);
arg_have_timebase = 1;
}
else if (arg_match(&arg, &error_resilient, argi))
cfg.g_error_resilient = arg_parse_uint(&arg);
else if (arg_match(&arg, &lag_in_frames, argi))
cfg.g_lag_in_frames = arg_parse_uint(&arg);
else if (arg_match(&arg, &dropframe_thresh, argi))
cfg.rc_dropframe_thresh = arg_parse_uint(&arg);
else if (arg_match(&arg, &resize_allowed, argi))
cfg.rc_resize_allowed = arg_parse_uint(&arg);
else if (arg_match(&arg, &resize_up_thresh, argi))
cfg.rc_resize_up_thresh = arg_parse_uint(&arg);
else if (arg_match(&arg, &resize_down_thresh, argi))
cfg.rc_resize_down_thresh = arg_parse_uint(&arg);
else if (arg_match(&arg, &resize_down_thresh, argi))
cfg.rc_resize_down_thresh = arg_parse_uint(&arg);
else if (arg_match(&arg, &end_usage, argi))
cfg.rc_end_usage = arg_parse_uint(&arg);
else if (arg_match(&arg, &target_bitrate, argi))
cfg.rc_target_bitrate = arg_parse_uint(&arg);
else if (arg_match(&arg, &min_quantizer, argi))
cfg.rc_min_quantizer = arg_parse_uint(&arg);
else if (arg_match(&arg, &max_quantizer, argi))
cfg.rc_max_quantizer = arg_parse_uint(&arg);
else if (arg_match(&arg, &undershoot_pct, argi))
cfg.rc_undershoot_pct = arg_parse_uint(&arg);
else if (arg_match(&arg, &overshoot_pct, argi))
cfg.rc_overshoot_pct = arg_parse_uint(&arg);
else if (arg_match(&arg, &buf_sz, argi))
cfg.rc_buf_sz = arg_parse_uint(&arg);
else if (arg_match(&arg, &buf_initial_sz, argi))
cfg.rc_buf_initial_sz = arg_parse_uint(&arg);
else if (arg_match(&arg, &buf_optimal_sz, argi))
cfg.rc_buf_optimal_sz = arg_parse_uint(&arg);
else if (arg_match(&arg, &bias_pct, argi))
{
cfg.rc_2pass_vbr_bias_pct = arg_parse_uint(&arg);
if (arg_passes < 2)
fprintf(stderr,
"Warning: option %s ignored in one-pass mode.\n",
arg.name);
}
else if (arg_match(&arg, &minsection_pct, argi))
{
cfg.rc_2pass_vbr_minsection_pct = arg_parse_uint(&arg);
if (arg_passes < 2)
fprintf(stderr,
"Warning: option %s ignored in one-pass mode.\n",
arg.name);
}
else if (arg_match(&arg, &maxsection_pct, argi))
{
cfg.rc_2pass_vbr_maxsection_pct = arg_parse_uint(&arg);
if (arg_passes < 2)
fprintf(stderr,
"Warning: option %s ignored in one-pass mode.\n",
arg.name);
}
else if (arg_match(&arg, &kf_min_dist, argi))
cfg.kf_min_dist = arg_parse_uint(&arg);
else if (arg_match(&arg, &kf_max_dist, argi))
cfg.kf_max_dist = arg_parse_uint(&arg);
else if (arg_match(&arg, &kf_disabled, argi))
cfg.kf_mode = VPX_KF_DISABLED;
else
argj++;
}
/* Handle codec specific options */
#if CONFIG_VP8_ENCODER
if (codec->iface == &vpx_codec_vp8_cx_algo)
{
ctrl_args = vp8_args;
ctrl_args_map = vp8_arg_ctrl_map;
}
#endif
for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step)
{
int match = 0;
arg.argv_step = 1;
for (i = 0; ctrl_args[i]; i++)
{
if (arg_match(&arg, ctrl_args[i], argi))
{
match = 1;
if (arg_ctrl_cnt < ARG_CTRL_CNT_MAX)
{
arg_ctrls[arg_ctrl_cnt][0] = ctrl_args_map[i];
arg_ctrls[arg_ctrl_cnt][1] = arg_parse_int(&arg);
arg_ctrl_cnt++;
}
}
}
if (!match)
argj++;
}
/* Check for unrecognized options */
for (argi = argv; *argi; argi++)
if (argi[0][0] == '-' && argi[0][1])
die("Error: Unrecognized option %s\n", *argi);
/* Handle non-option arguments */
in_fn = argv[0];
out_fn = argv[1];
if (!in_fn || !out_fn)
usage_exit();
memset(&stats, 0, sizeof(stats));
for (pass = one_pass_only ? one_pass_only - 1 : 0; pass < arg_passes; pass++)
{
int frames_in = 0, frames_out = 0;
unsigned long nbytes = 0;
struct detect_buffer detect;
/* Parse certain options from the input file, if possible */
infile = strcmp(in_fn, "-") ? fopen(in_fn, "rb") : stdin;
if (!infile)
{
fprintf(stderr, "Failed to open input file\n");
return EXIT_FAILURE;
}
fread(detect.buf, 1, 4, infile);
detect.valid = 0;
if (file_is_y4m(infile, &y4m, detect.buf))
{
if (y4m_input_open(&y4m, infile, detect.buf, 4) >= 0)
{
file_type = FILE_TYPE_Y4M;
cfg.g_w = y4m.pic_w;
cfg.g_h = y4m.pic_h;
/* Use the frame rate from the file only if none was specified
* on the command-line.
*/
if (!arg_have_timebase)
{
cfg.g_timebase.num = y4m.fps_d;
cfg.g_timebase.den = y4m.fps_n;
}
arg_use_i420 = 0;
}
else
{
fprintf(stderr, "Unsupported Y4M stream.\n");
return EXIT_FAILURE;
}
}
else if (file_is_ivf(infile, &fourcc, &cfg.g_w, &cfg.g_h, detect.buf))
{
file_type = FILE_TYPE_IVF;
switch (fourcc)
{
case 0x32315659:
arg_use_i420 = 0;
break;
case 0x30323449:
arg_use_i420 = 1;
break;
default:
fprintf(stderr, "Unsupported fourcc (%08x) in IVF\n", fourcc);
return EXIT_FAILURE;
}
}
else
{
file_type = FILE_TYPE_RAW;
detect.valid = 1;
}
#define SHOW(field) fprintf(stderr, " %-28s = %d\n", #field, cfg.field)
if (verbose && pass == 0)
{
fprintf(stderr, "Codec: %s\n", vpx_codec_iface_name(codec->iface));
fprintf(stderr, "Source file: %s Format: %s\n", in_fn,
arg_use_i420 ? "I420" : "YV12");
fprintf(stderr, "Destination file: %s\n", out_fn);
fprintf(stderr, "Encoder parameters:\n");
SHOW(g_usage);
SHOW(g_threads);
SHOW(g_profile);
SHOW(g_w);
SHOW(g_h);
SHOW(g_timebase.num);
SHOW(g_timebase.den);
SHOW(g_error_resilient);
SHOW(g_pass);
SHOW(g_lag_in_frames);
SHOW(rc_dropframe_thresh);
SHOW(rc_resize_allowed);
SHOW(rc_resize_up_thresh);
SHOW(rc_resize_down_thresh);
SHOW(rc_end_usage);
SHOW(rc_target_bitrate);
SHOW(rc_min_quantizer);
SHOW(rc_max_quantizer);
SHOW(rc_undershoot_pct);
SHOW(rc_overshoot_pct);
SHOW(rc_buf_sz);
SHOW(rc_buf_initial_sz);
SHOW(rc_buf_optimal_sz);
SHOW(rc_2pass_vbr_bias_pct);
SHOW(rc_2pass_vbr_minsection_pct);
SHOW(rc_2pass_vbr_maxsection_pct);
SHOW(kf_mode);
SHOW(kf_min_dist);
SHOW(kf_max_dist);
}
if(pass == (one_pass_only ? one_pass_only - 1 : 0)) {
if (file_type == FILE_TYPE_Y4M)
/*The Y4M reader does its own allocation.
Just initialize this here to avoid problems if we never read any
frames.*/
memset(&raw, 0, sizeof(raw));
else
vpx_img_alloc(&raw, arg_use_i420 ? VPX_IMG_FMT_I420 : VPX_IMG_FMT_YV12,
cfg.g_w, cfg.g_h, 1);
// This was added so that ivfenc will create monotically increasing
// timestamps. Since we create new timestamps for alt-reference frames
// we need to make room in the series of timestamps. Since there can
// only be 1 alt-ref frame ( current bitstream) multiplying by 2
// gives us enough room.
cfg.g_timebase.den *= 2;
}
outfile = strcmp(out_fn, "-") ? fopen(out_fn, "wb") : stdout;
if (!outfile)
{
fprintf(stderr, "Failed to open output file\n");
return EXIT_FAILURE;
}
if (stats_fn)
{
if (!stats_open_file(&stats, stats_fn, pass))
{
fprintf(stderr, "Failed to open statistics store\n");
return EXIT_FAILURE;
}
}
else
{
if (!stats_open_mem(&stats, pass))
{
fprintf(stderr, "Failed to open statistics store\n");
return EXIT_FAILURE;
}
}
cfg.g_pass = arg_passes == 2
? pass ? VPX_RC_LAST_PASS : VPX_RC_FIRST_PASS
: VPX_RC_ONE_PASS;
#if VPX_ENCODER_ABI_VERSION > (1 + VPX_CODEC_ABI_VERSION)
if (pass)
{
cfg.rc_twopass_stats_in = stats_get(&stats);
}
#endif
write_ivf_file_header(outfile, &cfg, codec->fourcc, 0);
/* Construct Encoder Context */
vpx_codec_enc_init(&encoder, codec->iface, &cfg,
show_psnr ? VPX_CODEC_USE_PSNR : 0);
ctx_exit_on_error(&encoder, "Failed to initialize encoder");
/* Note that we bypass the vpx_codec_control wrapper macro because
* we're being clever to store the control IDs in an array. Real
* applications will want to make use of the enumerations directly
*/
for (i = 0; i < arg_ctrl_cnt; i++)
{
if (vpx_codec_control_(&encoder, arg_ctrls[i][0], arg_ctrls[i][1]))
fprintf(stderr, "Error: Tried to set control %d = %d\n",
arg_ctrls[i][0], arg_ctrls[i][1]);
ctx_exit_on_error(&encoder, "Failed to control codec");
}
frame_avail = 1;
got_data = 0;
while (frame_avail || got_data)
{
vpx_codec_iter_t iter = NULL;
const vpx_codec_cx_pkt_t *pkt;
struct vpx_usec_timer timer;
if (!arg_limit || frames_in < arg_limit)
{
frame_avail = read_frame(infile, &raw, file_type, &y4m,
&detect);
if (frame_avail)
frames_in++;
fprintf(stderr,
"\rPass %d/%d frame %4d/%-4d %7ldB \033[K", pass + 1,
arg_passes, frames_in, frames_out, nbytes);
}
else
frame_avail = 0;
vpx_usec_timer_start(&timer);
// since we halved our timebase we need to double the timestamps
// and duration we pass in.
vpx_codec_encode(&encoder, frame_avail ? &raw : NULL, (frames_in - 1) * 2,
2, 0, arg_deadline);
vpx_usec_timer_mark(&timer);
cx_time += vpx_usec_timer_elapsed(&timer);
ctx_exit_on_error(&encoder, "Failed to encode frame");
got_data = 0;
while ((pkt = vpx_codec_get_cx_data(&encoder, &iter)))
{
got_data = 1;
switch (pkt->kind)
{
case VPX_CODEC_CX_FRAME_PKT:
frames_out++;
fprintf(stderr, " %6luF",
(unsigned long)pkt->data.frame.sz);
write_ivf_frame_header(outfile, pkt);
fwrite(pkt->data.frame.buf, 1, pkt->data.frame.sz, outfile);
nbytes += pkt->data.raw.sz;
break;
case VPX_CODEC_STATS_PKT:
frames_out++;
fprintf(stderr, " %6luS",
(unsigned long)pkt->data.twopass_stats.sz);
stats_write(&stats,
pkt->data.twopass_stats.buf,
pkt->data.twopass_stats.sz);
nbytes += pkt->data.raw.sz;
break;
case VPX_CODEC_PSNR_PKT:
if (show_psnr)
{
int i;
for (i = 0; i < 4; i++)
fprintf(stderr, "%.3lf ", pkt->data.psnr.psnr[i]);
}
break;
default:
break;
}
}
fflush(stdout);
}
/* this bitrate calc is simplified and relies on the fact that this
* application uses 1/timebase for framerate.
*/
fprintf(stderr,
"\rPass %d/%d frame %4d/%-4d %7ldB %7ldb/f %7"PRId64"b/s"
" %7lu %s (%.2f fps)\033[K", pass + 1,
arg_passes, frames_in, frames_out, nbytes, nbytes * 8 / frames_in,
nbytes * 8 *(int64_t)cfg.g_timebase.den/2/ cfg.g_timebase.num / frames_in,
cx_time > 9999999 ? cx_time / 1000 : cx_time,
cx_time > 9999999 ? "ms" : "us",
(float)frames_in * 1000000.0 / (float)cx_time);
vpx_codec_destroy(&encoder);
fclose(infile);
if (!fseek(outfile, 0, SEEK_SET))
write_ivf_file_header(outfile, &cfg, codec->fourcc, frames_out);
fclose(outfile);
stats_close(&stats);
fprintf(stderr, "\n");
if (one_pass_only)
break;
}
vpx_img_free(&raw);
free(argv);
return EXIT_SUCCESS;
}
vpx_codec_err_t vpx_codec_enc_init_multi_ver(vpx_codec_ctx_t *ctx,
vpx_codec_iface_t *iface,
vpx_codec_enc_cfg_t *cfg,
int num_enc,
vpx_codec_flags_t flags,
vpx_rational_t *dsf,
int ver) {
vpx_codec_err_t res = VPX_CODEC_OK;
if (ver != VPX_ENCODER_ABI_VERSION)
res = VPX_CODEC_ABI_MISMATCH;
else if (!ctx || !iface || !cfg || (num_enc > 16 || num_enc < 1))
res = VPX_CODEC_INVALID_PARAM;
else if (iface->abi_version != VPX_CODEC_INTERNAL_ABI_VERSION)
res = VPX_CODEC_ABI_MISMATCH;
else if (!(iface->caps & VPX_CODEC_CAP_ENCODER))
res = VPX_CODEC_INCAPABLE;
else if ((flags & VPX_CODEC_USE_XMA) && !(iface->caps & VPX_CODEC_CAP_XMA))
res = VPX_CODEC_INCAPABLE;
else if ((flags & VPX_CODEC_USE_PSNR)
&& !(iface->caps & VPX_CODEC_CAP_PSNR))
res = VPX_CODEC_INCAPABLE;
else if ((flags & VPX_CODEC_USE_OUTPUT_PARTITION)
&& !(iface->caps & VPX_CODEC_CAP_OUTPUT_PARTITION))
res = VPX_CODEC_INCAPABLE;
else {
int i;
void *mem_loc = NULL;
if (!(res = iface->enc.mr_get_mem_loc(cfg, &mem_loc))) {
for (i = 0; i < num_enc; i++) {
vpx_codec_priv_enc_mr_cfg_t mr_cfg;
/* Validate down-sampling factor. */
if (dsf->num < 1 || dsf->num > 4096 || dsf->den < 1 ||
dsf->den > dsf->num) {
res = VPX_CODEC_INVALID_PARAM;
break;
}
mr_cfg.mr_low_res_mode_info = mem_loc;
mr_cfg.mr_total_resolutions = num_enc;
mr_cfg.mr_encoder_id = num_enc - 1 - i;
mr_cfg.mr_down_sampling_factor.num = dsf->num;
mr_cfg.mr_down_sampling_factor.den = dsf->den;
/* Force Key-frame synchronization. Namely, encoder at higher
* resolution always use the same frame_type chosen by the
* lowest-resolution encoder.
*/
if (mr_cfg.mr_encoder_id)
cfg->kf_mode = VPX_KF_DISABLED;
ctx->iface = iface;
ctx->name = iface->name;
ctx->priv = NULL;
ctx->init_flags = flags;
ctx->config.enc = cfg;
res = ctx->iface->init(ctx, &mr_cfg);
if (res) {
const char *error_detail =
ctx->priv ? ctx->priv->err_detail : NULL;
/* Destroy current ctx */
ctx->err_detail = error_detail;
vpx_codec_destroy(ctx);
/* Destroy already allocated high-level ctx */
while (i) {
ctx--;
ctx->err_detail = error_detail;
vpx_codec_destroy(ctx);
i--;
}
}
if (ctx->priv)
ctx->priv->iface = ctx->iface;
if (res)
break;
ctx++;
cfg++;
dsf++;
}
ctx--;
}
}
return SAVE_STATUS(ctx, res);
}
int main(int argc, char **argv) {
FILE *infile, *outfile[MAX_LAYERS];
vpx_codec_ctx_t codec;
vpx_codec_enc_cfg_t cfg;
int frame_cnt = 0;
vpx_image_t raw;
int frame_avail = 1;
int got_data = 0;
int i;
int frames_in_layer[MAX_LAYERS] = {0};
clock_t before;
clock_t after;
int pts = 0; /* PTS starts at 0 */
int frame_duration = 1; /* 1 timebase tick per frame */
parse_command_line(argc, argv, &cfg);
// Allocate image buffer
if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, cfg.g_w, cfg.g_h, 32))
die("Failed to allocate image", cfg.g_w, cfg.g_h);
set_default_configuration(&cfg);
/* Open input file */
if (!(infile = fopen(input_filename, "rb")))
die("Failed to open %s for reading", argv[1]);
/* Open output file */
for (i = 0; i < number_spatial_layers; i++) {
char file_name[512];
snprintf(file_name, sizeof(file_name), "%s_%d.ivf", output_filename, i);
if (!(outfile[i] = fopen(file_name, "wb")))
die("Failed to open %s for writing", file_name);
write_ivf_file_header(outfile[i], &cfg, 0);
}
initialize_codec(&codec, &cfg);
// skip initial frames
for (i = 0; i < number_frames_to_skip; i++) {
read_frame(infile, &raw);
}
before = clock();
// Encoding frames
while ((frame_avail || got_data) &&
frame_cnt <= number_frames_to_code * number_spatial_layers) {
int flags = 0;
vpx_codec_iter_t iter = NULL;
const vpx_codec_cx_pkt_t *pkt;
int layer = calculate_layer(frame_cnt, number_spatial_layers);
int is_I_frame_in_layer =
(((frame_cnt - 1) / number_spatial_layers % key_period) == 0);
int is_dummy = (frame_cnt == 0);
if (is_dummy) { // Dummy frame
flags = VPX_EFLAG_FORCE_KF;
frame_avail = read_dummy_frame(&raw);
} else { // Regular frame
// Read a new frame only at the base layer
if (layer == 0) frame_avail = read_frame(infile, &raw);
switch_to_layer(layer, cfg.g_w, cfg.g_h, &codec);
flags = get_flag(is_I_frame_in_layer, layer, INTER_LAYER_PREDICTION_I);
}
// Actual Encoding
if (vpx_codec_encode(&codec, frame_avail ? &raw : NULL, pts, 1, flags,
VPX_DL_REALTIME))
die_codec(&codec, "Failed to encode frame");
got_data = 0;
// Process data / Get PSNR statistics
while ((pkt = vpx_codec_get_cx_data(&codec, &iter))) {
got_data = 1;
switch (pkt->kind) {
case VPX_CODEC_CX_FRAME_PKT:
for (i = layer; i < number_spatial_layers; i++) {
write_ivf_frame_header(outfile[i], pkt);
(void)fwrite(pkt->data.frame.buf, 1, pkt->data.frame.sz,
outfile[i]);
frames_in_layer[i]++;
}
break;
case VPX_CODEC_PSNR_PKT:
if (frame_cnt != 0)
printf(
"Processed Frame %d, layer %d, PSNR(Total/Y/U/V): "
"%2.3f %2.3f %2.3f %2.3f \n",
(frame_cnt - 1) / number_spatial_layers + 1, layer,
pkt->data.psnr.psnr[0], pkt->data.psnr.psnr[1],
pkt->data.psnr.psnr[2], pkt->data.psnr.psnr[3]);
break;
default:
break;
}
}
frame_cnt++;
// TODO(ivan): Modify ts later if(!layer)
pts += frame_duration;
}
// end while
after = clock();
printf("Processed %d frames in different resolutions in %ld ms.\n",
frame_cnt - 1, (int)(after - before) / (CLOCKS_PER_SEC / 1000));
fclose(infile);
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 < number_spatial_layers; i++) {
if (!fseek(outfile[i], 0, SEEK_SET)) {
write_ivf_file_header(outfile[i], &cfg, frames_in_layer[i]);
}
fclose(outfile[i]);
}
return EXIT_SUCCESS;
}
static void enc_postprocess(MSFilter *f) {
EncState *s = (EncState *)f->data;
if (s->ready) vpx_codec_destroy(&s->codec);
vp8rtpfmt_packer_uninit(&s->packer);
s->ready = FALSE;
}
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 update_frame_num = 0; //
/* Open files */
if(argc!=6) //
die("Usage: %s <width> <height> <infile> <outfile> <frame>\n", //
argv[0]); //
//
update_frame_num = atoi(argv[5]); //
if(!update_frame_num) //
die("Couldn't parse frame number '%s'\n", argv[5]); //
//
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;
}
/* 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(frame_cnt + 1 == update_frame_num) { //
vpx_ref_frame_t ref; //
//
ref.frame_type = VP8_LAST_FRAME; //
ref.img = raw; //
//
if(vpx_codec_control(&codec, VP8_SET_REFERENCE, &ref)) //
die_codec(&codec, "Failed to set reference frame"); //
} //
//
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);
if(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);
printf("Processed %d frames.\n",frame_cnt-1);
if(vpx_codec_destroy(&codec))
die_codec(&codec, "Failed to destroy codec");
/* 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;
}
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};
const VpxInterface *encoder = NULL;
int pass;
const int fps = 30; // TODO(dkovalev) add command line argument
const int bitrate = 200; // kbit/s TODO(dkovalev) add command line argument
const char *const codec_arg = argv[1];
const char *const width_arg = argv[2];
const char *const height_arg = argv[3];
const char *const infile_arg = argv[4];
const char *const outfile_arg = argv[5];
exec_name = argv[0];
if (argc != 6)
die("Invalid number of arguments.");
encoder = get_vpx_encoder_by_name(codec_arg);
if (!encoder)
die("Unsupported codec.");
info.codec_fourcc = encoder->fourcc;
info.time_base.numerator = 1;
info.time_base.denominator = fps;
info.frame_width = strtol(width_arg, NULL, 0);
info.frame_height = strtol(height_arg, 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(outfile_arg, kContainerIVF, &info);
if (!writer)
die("Failed to open %s for writing", outfile_arg);
printf("Using %s\n", vpx_codec_iface_name(encoder->interface()));
res = vpx_codec_enc_config_default(encoder->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;
}
if (!(infile = fopen(infile_arg, "rb")))
die("Failed to open %s for reading", infile_arg);
if (vpx_codec_enc_init(&codec, encoder->interface(), &cfg, 0))
die_codec(&codec, "Failed to initialize encoder");
while (vpx_img_read(&raw, infile)) {
++frame_count;
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);
}
}
if (pass == 0) {
get_frame_stats(&codec, NULL, frame_count, 1, 0, VPX_DL_BEST_QUALITY,
&stats);
} else {
printf("\n");
}
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);
return EXIT_SUCCESS;
}
void matrixCalc(void *inputs, void *outputs, long in_bitRate)
{
//Validate the input
TML::Matrix i(inputs,0);
if (i.dims() != 2)
throw "Input should be a 2D matrix";
if (!i.isChar())
throw "Input should have character data";
if (i.planes() != 4)
throw "Input needs 4 planes";
if (i.dim(0) % 64 != 0)
throw "Width needs to be a multiple of 64";
if (i.dim(1) % 2 != 0)
throw "Height needs to be a multiple of 2";
if (i.dim(0) != m_cfg.g_w || i.dim(1) != m_cfg.g_h || m_cfg.rc_target_bitrate != in_bitRate)
{
vpx_img_free(&m_raw);
vpx_img_alloc(&m_raw, VPX_IMG_FMT_I420, i.dim(0), i.dim(1), 1);
vpx_codec_destroy(&m_codec);
vpx_codec_enc_config_default(vpx_codec_vp8_cx(), &m_cfg, 0);
m_cfg.rc_target_bitrate = in_bitRate;
m_cfg.g_w = i.dim(0);
m_cfg.g_h = i.dim(1);
vpx_codec_enc_init(&m_codec, vpx_codec_vp8_cx(), &m_cfg, 0);
}
//ARGB -> YYYY U V
int x;
const int N = 32;
const int Uoffset = i.dim(0)*i.dim(1);
const int Voffset = Uoffset + Uoffset/4;
const int w = i.dim(0);
const int h = i.dim(1);
const int sy = i.stride(1);
unsigned char *data = (unsigned char*)i.data();
int y;
unsigned char *buffer = m_raw.planes[0];
//RRRR
__v16qi rShuffle = {1, -1,-1,-1,
5, -1,-1,-1,
9, -1,-1,-1,
13, -1,-1,-1
};
__v16qi gShuffle = {2, -1,-1,-1,
6, -1,-1,-1,
10,-1,-1,-1,
14,-1,-1,-1
};
__v16qi bShuffle = {3,-1,-1,-1,
7,-1,-1,-1,
11,-1,-1,-1,
15,-1,-1,-1
};
//Shuffle so elements are moved to front/back
__v16qi _aShuffle = {
0, 4, 8, 12,
-1, -1, -1, -1,
-1, -1, -1, -1,
-1, -1, -1, -1
};
__v16qi _bShuffle = {
-1, -1, -1, -1,
0, 4, 8, 12,
-1, -1, -1, -1,
-1, -1, -1, -1
};
__v16qi _cShuffle = {
-1, -1, -1, -1,
-1, -1, -1, -1,
0, 4, 8, 12,
-1, -1, -1, -1
};
__v16qi _dShuffle = {
-1, -1, -1, -1,
-1, -1, -1, -1,
-1, -1, -1, -1,
0, 4, 8, 12
};
__v8hi R2Y = { 27, 27, 27, 27, 27, 27, 27, 27};
__v8hi G2Y = { 91, 91, 91, 91, 91, 91, 91, 91};
__v8hi B2Y = { 9, 9, 9, 9, 9, 9, 9, 9};
__v8hi R2U = { -12,-12,-12,-12,-12,-12,-12,-12};
__v8hi G2U = { -43,-43,-43,-43,-43,-43,-43,-43};
__v8hi B2U = { 55, 55, 55, 55, 55, 55, 55, 55};
__v8hi R2V = { 78, 78, 78, 78, 78, 78, 78, 78};
__v8hi G2V = { -71,-71,-71,-71,-71,-71,-71,-71};
__v8hi B2V = { -7, -7, -7, -7, -7, -7, -7, -7};
__v8hi m127 = {127,127,127,127,127,127,127,127};
__v8hi zero = { 0, 0, 0, 0, 0, 0, 0, 0};
__v8hi two55 = {255,255,255,255,255,255,255,255};
for (y=0; y<h; y+=2)
{
for (x=0; x<w; x+=N)
{
__v8hi tY[N/8];
__v8hi bY[N/8];
__v8hi tU[N/8];
__v8hi bU[N/8];
__v8hi tV[N/8];
__v8hi bV[N/8];
//Step 1: Convert to YUV
int n;
for (n=0; n<N; n+=8) //Read 8x per lane
{
__v16qi tARGBx4_l = _mm_load_si128((__m128i*)(data + y*w*4 + x*4 + n*4));
__v16qi tARGBx4_r = _mm_load_si128((__m128i*)(data + y*w*4 + x*4 + n*4 + 16));
__v16qi bARGBx4_l = _mm_load_si128((__m128i*)(data + y*w*4 + x*4 + n*4 + sy));
__v16qi bARGBx4_r = _mm_load_si128((__m128i*)(data + y*w*4 + x*4 + n*4 + sy + 16));
// ARGB(1) ARGB(2) ARGB(3) ARGB(4) | ARGB(5) ARGB(6) ARGB(7) ARGB(8)
// => AARRGGBB(1,5) AARRGGBB(2,6) | AARRGGBB(3,7) AARRGGBB(4,8)
__v16qi tARGBx2_15 = _mm_unpacklo_epi8(tARGBx4_l, tARGBx4_r);
__v16qi tARGBx2_26 = _mm_unpackhi_epi8(tARGBx4_l, tARGBx4_r);
__v16qi bARGBx2_15 = _mm_unpacklo_epi8(bARGBx4_l, bARGBx4_r);
__v16qi bARGBx2_26 = _mm_unpackhi_epi8(bARGBx4_l, bARGBx4_r);
// AARRGGBB(1,5) AARRGGBB(2,6) | AARRGGBB(3,7) AARRGGBB(4,8)
// => AAAARRRRGGGGBBBB(1,3,5,7) | AAAARRRRGGGGBBBB(2,4,6,8)
__v16qi tARGB_1357 = _mm_unpacklo_epi8(tARGBx2_15, tARGBx2_26);
__v16qi tARGB_2468 = _mm_unpackhi_epi8(tARGBx2_15, tARGBx2_26);
__v16qi bARGB_1357 = _mm_unpacklo_epi8(bARGBx2_15, bARGBx2_26);
__v16qi bARGB_2468 = _mm_unpackhi_epi8(bARGBx2_15, bARGBx2_26);
//AAAARRRRGGGGBBBB(1,3,5,7) | AAAARRRRGGGGBBBB(2,4,6,8)
// => AAAAAAAARRRRRRRR | GGGGGGGGBBBBBBBB
__v16qi tAARR = _mm_unpacklo_epi8(tARGB_1357, tARGB_2468);
__v16qi tGGBB = _mm_unpackhi_epi8(tARGB_1357, tARGB_2468);
__v16qi bAARR = _mm_unpacklo_epi8(bARGB_1357, bARGB_2468);
__v16qi bGGBB = _mm_unpackhi_epi8(bARGB_1357, bARGB_2468);
//Unpack to 8 R's, 8 G's, and 8 B's.
__v8hi tRRRR = _mm_unpackhi_epi8(tAARR, zero);
__v8hi tGGGG = _mm_unpacklo_epi8(tGGBB, zero);
__v8hi tBBBB = _mm_unpackhi_epi8(tGGBB, zero);
__v8hi bRRRR = _mm_unpackhi_epi8(bAARR, zero);
__v8hi bGGGG = _mm_unpacklo_epi8(bGGBB, zero);
__v8hi bBBBB = _mm_unpackhi_epi8(bGGBB, zero);
//Convert to YUV (8x parallel)
__v8hi tYYYY = _mm_add_epi16(_mm_mullo_epi16(tRRRR, R2Y), _mm_add_epi16(_mm_mullo_epi16(tGGGG, G2Y), _mm_mullo_epi16(tBBBB, B2Y)));
__v8hi tUUUU = _mm_add_epi16(_mm_mullo_epi16(tRRRR, R2U), _mm_add_epi16(_mm_mullo_epi16(tGGGG, G2U), _mm_mullo_epi16(tBBBB, B2U)));
__v8hi tVVVV = _mm_add_epi16(_mm_mullo_epi16(tRRRR, R2V), _mm_add_epi16(_mm_mullo_epi16(tGGGG, G2V), _mm_mullo_epi16(tBBBB, B2V)));
__v8hi bYYYY = _mm_add_epi16(_mm_mullo_epi16(bRRRR, R2Y), _mm_add_epi16(_mm_mullo_epi16(bGGGG, G2Y), _mm_mullo_epi16(bBBBB, B2Y)));
__v8hi bUUUU = _mm_add_epi16(_mm_mullo_epi16(bRRRR, R2U), _mm_add_epi16(_mm_mullo_epi16(bGGGG, G2U), _mm_mullo_epi16(bBBBB, B2U)));
__v8hi bVVVV = _mm_add_epi16(_mm_mullo_epi16(bRRRR, R2V), _mm_add_epi16(_mm_mullo_epi16(bGGGG, G2V), _mm_mullo_epi16(bBBBB, B2V)));
tUUUU = _mm_add_epi16(_mm_srai_epi16(tUUUU, 7), m127);
tVVVV = _mm_add_epi16(_mm_srai_epi16(tVVVV, 7), m127);
bUUUU = _mm_add_epi16(_mm_srai_epi16(bUUUU, 7), m127);
bVVVV = _mm_add_epi16(_mm_srai_epi16(bVVVV, 7), m127);
//Remove the fractional portion and clamp in 0...255
tY[n/8] = _mm_min_epi16(_mm_srai_epi16(_mm_max_epi16(tYYYY,zero), 7), two55);
tU[n/8] = _mm_min_epi16(_mm_max_epi16(tUUUU,zero), two55);
tV[n/8] = _mm_min_epi16(_mm_max_epi16(tVVVV,zero), two55);
bY[n/8] = _mm_min_epi16(_mm_srai_epi16(_mm_max_epi16(bYYYY,zero), 7), two55);
bU[n/8] = _mm_min_epi16(_mm_max_epi16(bUUUU,zero), two55);
bV[n/8] = _mm_min_epi16(_mm_max_epi16(bVVVV,zero), two55);
}
// Step 2 - Write out Luma (part 1)
for (n=0; n<N; n+=16)
{
__v8hi A = tY[n/8];
__v8hi B = tY[n/8+1];
__m128i Y = _mm_packus_epi16(A,B);
_mm_storeu_si128((__m128i*)(buffer+y*w+x+n), Y);
}
for (n=0; n<N; n+=16)
{
__v8hi A = bY[n/8];
__v8hi B = bY[n/8+1];
__m128i Y = _mm_packus_epi16(A,B);
_mm_storeu_si128((__m128i*)(buffer+y*w+x+n+w), Y);
}
//Step 3 -- U and V data...
for (n=0; n<N; n+=32)
{
__m128i U16a = _mm_add_epi16(tU[n/8], bU[n/8]);
__m128i U16b = _mm_add_epi16(tU[n/8+1], bU[n/8+1]);
__m128i U16c = _mm_add_epi16(tU[n/8+2], bU[n/8+2]);
__m128i U16d = _mm_add_epi16(tU[n/8+3], bU[n/8+3]);
U16a = _mm_srli_epi16(_mm_hadd_epi16(U16a, U16b),2);
U16c = _mm_srli_epi16(_mm_hadd_epi16(U16c, U16d),2);
__m128i U = _mm_packus_epi16(U16a, U16c);
_mm_storeu_si128((__m128i*)(buffer+Uoffset+y/2*w/2 + x/2+n/2), U);
}
for (n=0; n<N; n+=32)
{
__m128i U16a = _mm_add_epi16(tV[n/8], bV[n/8]);
__m128i U16b = _mm_add_epi16(tV[n/8+1], bV[n/8+1]);
__m128i U16c = _mm_add_epi16(tV[n/8+2], bV[n/8+2]);
__m128i U16d = _mm_add_epi16(tV[n/8+3], bV[n/8+3]);
U16a = _mm_srli_epi16(_mm_hadd_epi16(U16a, U16b),2);
U16c = _mm_srli_epi16(_mm_hadd_epi16(U16c, U16d),2);
__m128i U = _mm_packus_epi16(U16a, U16c);
_mm_storeu_si128((__m128i*)(buffer+Voffset+y/2*w/2 + x/2+n/2), U);
}
}
}
m_frameCnt++;
vpx_codec_encode(&m_codec, &m_raw, m_frameCnt, 1, 0, VPX_DL_REALTIME);
vpx_codec_iter_t iter = NULL;
const vpx_codec_cx_pkt_t *pkt;
while ((pkt = vpx_codec_get_cx_data(&m_codec, &iter)))
{
if (pkt->kind == VPX_CODEC_CX_FRAME_PKT)
{
//Generate output
TML::Matrix o(outputs,0);
_jit_matrix_info m;
memset(&m, 0, sizeof(m));
m.dimcount = 2;
m.dim[0] = pkt->data.frame.sz;
m.dim[1] = 1;
m.dimstride[0] = pkt->data.frame.sz;
m.dimstride[1] = 1;
m.planecount = 1;
m.type = _jit_sym_char;
o.resizeTo(&m);
memcpy(o.data(), pkt->data.frame.buf, pkt->data.frame.sz);
break;
}
}
}
int main(int argc, char **argv) {
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;
VpxVideoWriter *writer = NULL;
const VpxInterface *encoder = NULL;
const int fps = 2; // TODO(dkovalev) add command line argument
const double bits_per_pixel_per_frame = 0.067;
exec_name = argv[0];
if (argc != 6)
die("Invalid number of arguments");
memset(&info, 0, sizeof(info));
encoder = get_vpx_encoder_by_name(argv[1]);
if (encoder == NULL) {
die("Unsupported codec.");
}
assert(encoder != NULL);
info.codec_fourcc = encoder->fourcc;
info.frame_width = strtol(argv[2], NULL, 0);
info.frame_height = strtol(argv[3], 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(encoder->codec_interface()));
res = vpx_codec_enc_config_default(encoder->codec_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 = (unsigned int)(bits_per_pixel_per_frame * cfg.g_w *
cfg.g_h * fps / 1000);
cfg.g_lag_in_frames = 0;
writer = vpx_video_writer_open(argv[5], kContainerIVF, &info);
if (!writer)
die("Failed to open %s for writing.", argv[5]);
if (!(infile = fopen(argv[4], "rb")))
die("Failed to open %s for reading.", argv[4]);
if (vpx_codec_enc_init(&codec, encoder->codec_interface(), &cfg, 0))
die_codec(&codec, "Failed to initialize encoder");
// Encode frames.
while (vpx_img_read(&raw, infile)) {
++frame_count;
if (frame_count == 22 && encoder->fourcc == VP8_FOURCC) {
set_roi_map(&cfg, &codec);
} else if (frame_count == 33) {
set_active_map(&cfg, &codec);
} else if (frame_count == 44) {
unset_active_map(&cfg, &codec);
}
encode_frame(&codec, &raw, frame_count, writer);
}
// Flush encoder.
while (encode_frame(&codec, NULL, -1, writer)) {}
printf("\n");
fclose(infile);
printf("Processed %d frames.\n", frame_count);
vpx_img_free(&raw);
if (vpx_codec_destroy(&codec))
die_codec(&codec, "Failed to destroy codec.");
vpx_video_writer_close(writer);
return EXIT_SUCCESS;
}
int main(int argc, const char **argv) {
AppInput app_input = {0};
VpxVideoWriter *writer = NULL;
VpxVideoInfo info = {0};
vpx_codec_ctx_t codec;
vpx_codec_enc_cfg_t enc_cfg;
SvcContext svc_ctx;
uint32_t i;
uint32_t frame_cnt = 0;
vpx_image_t raw;
vpx_codec_err_t res;
int pts = 0; /* PTS starts at 0 */
int frame_duration = 1; /* 1 timebase tick per frame */
FILE *infile = NULL;
memset(&svc_ctx, 0, sizeof(svc_ctx));
svc_ctx.log_print = 1;
exec_name = argv[0];
parse_command_line(argc, argv, &app_input, &svc_ctx, &enc_cfg);
// Allocate image buffer
if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, enc_cfg.g_w, enc_cfg.g_h, 32))
die("Failed to allocate image %dx%d\n", enc_cfg.g_w, enc_cfg.g_h);
if (!(infile = fopen(app_input.input_filename, "rb")))
die("Failed to open %s for reading\n", app_input.input_filename);
// Initialize codec
if (vpx_svc_init(&svc_ctx, &codec, vpx_codec_vp9_cx(), &enc_cfg) !=
VPX_CODEC_OK)
die("Failed to initialize encoder\n");
info.codec_fourcc = VP9_FOURCC;
info.time_base.numerator = enc_cfg.g_timebase.num;
info.time_base.denominator = enc_cfg.g_timebase.den;
if (vpx_svc_get_layer_resolution(&svc_ctx, svc_ctx.spatial_layers - 1,
(unsigned int *)&info.frame_width,
(unsigned int *)&info.frame_height) !=
VPX_CODEC_OK) {
die("Failed to get output resolution");
}
writer = vpx_video_writer_open(app_input.output_filename, kContainerIVF,
&info);
if (!writer)
die("Failed to open %s for writing\n", app_input.output_filename);
// skip initial frames
for (i = 0; i < app_input.frames_to_skip; ++i)
vpx_img_read(&raw, infile);
// Encode frames
while (frame_cnt < app_input.frames_to_code) {
if (!vpx_img_read(&raw, infile))
break;
res = vpx_svc_encode(&svc_ctx, &codec, &raw, pts, frame_duration,
VPX_DL_REALTIME);
printf("%s", vpx_svc_get_message(&svc_ctx));
if (res != VPX_CODEC_OK) {
die_codec(&codec, "Failed to encode frame");
}
if (vpx_svc_get_frame_size(&svc_ctx) > 0) {
vpx_video_writer_write_frame(writer,
vpx_svc_get_buffer(&svc_ctx),
vpx_svc_get_frame_size(&svc_ctx),
pts);
}
++frame_cnt;
pts += frame_duration;
}
printf("Processed %d frames\n", frame_cnt);
fclose(infile);
if (vpx_codec_destroy(&codec)) die_codec(&codec, "Failed to destroy codec");
vpx_video_writer_close(writer);
vpx_img_free(&raw);
// display average size, psnr
printf("%s", vpx_svc_dump_statistics(&svc_ctx));
vpx_svc_release(&svc_ctx);
return EXIT_SUCCESS;
}
int main(int argc, char **argv) {
int frame_cnt = 0;
FILE *outfile = NULL;
vpx_codec_ctx_t codec;
const VpxInterface *decoder = 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.");
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]);
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]);
info = vpx_video_reader_get_info(reader);
decoder = get_vpx_decoder_by_fourcc(info->codec_fourcc);
if (!decoder)
die("Unknown input codec.");
printf("Using %s\n", vpx_codec_iface_name(decoder->codec_interface()));
if (vpx_codec_dec_init(&codec, decoder->codec_interface(), NULL, 0))
die_codec(&codec, "Failed to initialize decoder.");
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_reader_get_frame(reader,
&frame_size);
if (vpx_codec_decode(&codec, frame, (unsigned int)frame_size, NULL, 0))
die_codec(&codec, "Failed to decode frame.");
++frame_cnt;
skip = (is_range && frame_cnt >= n && frame_cnt <= m) ||
(!is_range && m - (frame_cnt - 1) % m <= n);
if (!skip) {
putc('.', stdout);
while ((img = vpx_codec_get_frame(&codec, &iter)) != NULL)
vpx_img_write(img, outfile);
} else {
putc('X', stdout);
}
fflush(stdout);
}
printf("Processed %d frames.\n", frame_cnt);
if (vpx_codec_destroy(&codec))
die_codec(&codec, "Failed to destroy codec.");
printf("Play: ffplay -f rawvideo -pix_fmt yuv420p -s %dx%d %s\n",
info->frame_width, info->frame_height, argv[2]);
vpx_video_reader_close(reader);
fclose(outfile);
return EXIT_SUCCESS;
}
int main(int argc, char **argv) {
VpxVideoWriter *outfile[VPX_TS_MAX_LAYERS] = {NULL};
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 speed;
int frame_avail;
int got_data;
int flags = 0;
unsigned int i;
int pts = 0; // PTS starts at 0.
int frame_duration = 1; // 1 timebase tick per frame.
int layering_mode = 0;
int layer_flags[VPX_TS_MAX_PERIODICITY] = {0};
int flag_periodicity = 1;
vpx_svc_layer_id_t layer_id = {0, 0};
const VpxInterface *encoder = NULL;
FILE *infile = NULL;
struct RateControlMetrics rc;
int64_t cx_time = 0;
const int min_args_base = 11;
#if CONFIG_VP9_HIGHBITDEPTH
vpx_bit_depth_t bit_depth = VPX_BITS_8;
int input_bit_depth = 8;
const int min_args = min_args_base + 1;
#else
const int min_args = min_args_base;
#endif // CONFIG_VP9_HIGHBITDEPTH
double sum_bitrate = 0.0;
double sum_bitrate2 = 0.0;
double framerate = 30.0;
exec_name = argv[0];
// Check usage and arguments.
if (argc < min_args) {
#if CONFIG_VP9_HIGHBITDEPTH
die("Usage: %s <infile> <outfile> <codec_type(vp8/vp9)> <width> <height> "
"<rate_num> <rate_den> <speed> <frame_drop_threshold> <mode> "
"<Rate_0> ... <Rate_nlayers-1> <bit-depth> \n", argv[0]);
#else
die("Usage: %s <infile> <outfile> <codec_type(vp8/vp9)> <width> <height> "
"<rate_num> <rate_den> <speed> <frame_drop_threshold> <mode> "
"<Rate_0> ... <Rate_nlayers-1> \n", argv[0]);
#endif // CONFIG_VP9_HIGHBITDEPTH
}
encoder = get_vpx_encoder_by_name(argv[3]);
if (!encoder)
die("Unsupported codec.");
printf("Using %s\n", vpx_codec_iface_name(encoder->codec_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[10], NULL, 0);
if (layering_mode < 0 || layering_mode > 12) {
die("Invalid layering mode (0..12) %s", argv[10]);
}
if (argc != min_args + mode_to_num_layers[layering_mode]) {
die("Invalid number of arguments");
}
#if CONFIG_VP9_HIGHBITDEPTH
switch (strtol(argv[argc-1], NULL, 0)) {
case 8:
bit_depth = VPX_BITS_8;
input_bit_depth = 8;
break;
case 10:
bit_depth = VPX_BITS_10;
input_bit_depth = 10;
break;
case 12:
bit_depth = VPX_BITS_12;
input_bit_depth = 12;
break;
default:
die("Invalid bit depth (8, 10, 12) %s", argv[argc-1]);
}
if (!vpx_img_alloc(&raw,
bit_depth == VPX_BITS_8 ? VPX_IMG_FMT_I420 :
VPX_IMG_FMT_I42016,
width, height, 32)) {
die("Failed to allocate image", width, height);
}
#else
if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, width, height, 32)) {
die("Failed to allocate image", width, height);
}
#endif // CONFIG_VP9_HIGHBITDEPTH
// Populate encoder configuration.
res = vpx_codec_enc_config_default(encoder->codec_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;
#if CONFIG_VP9_HIGHBITDEPTH
if (bit_depth != VPX_BITS_8) {
cfg.g_bit_depth = bit_depth;
cfg.g_input_bit_depth = input_bit_depth;
cfg.g_profile = 2;
}
#endif // CONFIG_VP9_HIGHBITDEPTH
// 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);
speed = strtol(argv[8], NULL, 0);
if (speed < 0) {
die("Invalid speed setting: must be positive");
}
for (i = min_args_base;
(int)i < min_args_base + mode_to_num_layers[layering_mode];
++i) {
cfg.ts_target_bitrate[i - 11] = strtol(argv[i], NULL, 0);
}
// Real time parameters.
cfg.rc_dropframe_thresh = strtol(argv[9], NULL, 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 = 50;
cfg.rc_overshoot_pct = 50;
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_AUTO;
// Disable automatic keyframe placement.
cfg.kf_min_dist = cfg.kf_max_dist = 3000;
set_temporal_layer_pattern(layering_mode,
&cfg,
layer_flags,
&flag_periodicity);
set_rate_control_metrics(&rc, &cfg);
// Target bandwidth for the whole stream.
// Set to ts_target_bitrate for highest layer (total bitrate).
cfg.rc_target_bitrate = cfg.ts_target_bitrate[cfg.ts_number_layers - 1];
// Open input file.
if (!(infile = fopen(argv[1], "rb"))) {
die("Failed to open %s for reading", argv[1]);
}
framerate = cfg.g_timebase.den / cfg.g_timebase.num;
// 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 = encoder->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);
assert(outfile[i] != NULL);
}
// No spatial layers in this encoder.
cfg.ss_number_layers = 1;
// Initialize codec.
#if CONFIG_VP9_HIGHBITDEPTH
if (vpx_codec_enc_init(
&codec, encoder->codec_interface(), &cfg,
bit_depth == VPX_BITS_8 ? 0 : VPX_CODEC_USE_HIGHBITDEPTH))
#else
if (vpx_codec_enc_init(&codec, encoder->codec_interface(), &cfg, 0))
#endif // CONFIG_VP9_HIGHBITDEPTH
die_codec(&codec, "Failed to initialize encoder");
if (strncmp(encoder->name, "vp8", 3) == 0) {
vpx_codec_control(&codec, VP8E_SET_CPUUSED, -speed);
vpx_codec_control(&codec, VP8E_SET_NOISE_SENSITIVITY, kDenoiserOnYOnly);
} else if (strncmp(encoder->name, "vp9", 3) == 0) {
vpx_codec_control(&codec, VP8E_SET_CPUUSED, speed);
vpx_codec_control(&codec, VP9E_SET_AQ_MODE, 3);
vpx_codec_control(&codec, VP9E_SET_FRAME_PERIODIC_BOOST, 0);
vpx_codec_control(&codec, VP9E_SET_NOISE_SENSITIVITY, 0);
if (vpx_codec_control(&codec, VP9E_SET_SVC, layering_mode > 0 ? 1: 0)) {
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);
// This controls the maximum target size of the key frame.
// For generating smaller key frames, use a smaller max_intra_size_pct
// value, like 100 or 200.
{
const int max_intra_size_pct = 900;
vpx_codec_control(&codec, VP8E_SET_MAX_INTRA_BITRATE_PCT,
max_intra_size_pct);
}
frame_avail = 1;
while (frame_avail || got_data) {
struct vpx_usec_timer timer;
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];
if (strncmp(encoder->name, "vp9", 3) == 0) {
vpx_codec_control(&codec, VP9E_SET_SVC_LAYER_ID, &layer_id);
}
flags = layer_flags[frame_cnt % flag_periodicity];
frame_avail = vpx_img_read(&raw, infile);
if (frame_avail)
++rc.layer_input_frames[layer_id.temporal_layer_id];
vpx_usec_timer_start(&timer);
if (vpx_codec_encode(&codec, frame_avail? &raw : NULL, pts, 1, flags,
VPX_DL_REALTIME)) {
die_codec(&codec, "Failed to encode frame");
}
vpx_usec_timer_mark(&timer);
cx_time += vpx_usec_timer_elapsed(&timer);
// 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);
++rc.layer_tot_enc_frames[i];
rc.layer_encoding_bitrate[i] += 8.0 * pkt->data.frame.sz;
// Keep count of rate control stats per layer (for non-key frames).
if (i == cfg.ts_layer_id[frame_cnt % cfg.ts_periodicity] &&
!(pkt->data.frame.flags & VPX_FRAME_IS_KEY)) {
rc.layer_avg_frame_size[i] += 8.0 * pkt->data.frame.sz;
rc.layer_avg_rate_mismatch[i] +=
fabs(8.0 * pkt->data.frame.sz - rc.layer_pfb[i]) /
rc.layer_pfb[i];
++rc.layer_enc_frames[i];
}
}
// Update for short-time encoding bitrate states, for moving window
// of size rc->window, shifted by rc->window / 2.
// Ignore first window segment, due to key frame.
if (frame_cnt > rc.window_size) {
sum_bitrate += 0.001 * 8.0 * pkt->data.frame.sz * framerate;
if (frame_cnt % rc.window_size == 0) {
rc.window_count += 1;
rc.avg_st_encoding_bitrate += sum_bitrate / rc.window_size;
rc.variance_st_encoding_bitrate +=
(sum_bitrate / rc.window_size) *
(sum_bitrate / rc.window_size);
sum_bitrate = 0.0;
}
}
// Second shifted window.
if (frame_cnt > rc.window_size + rc.window_size / 2) {
sum_bitrate2 += 0.001 * 8.0 * pkt->data.frame.sz * framerate;
if (frame_cnt > 2 * rc.window_size &&
frame_cnt % rc.window_size == 0) {
rc.window_count += 1;
rc.avg_st_encoding_bitrate += sum_bitrate2 / rc.window_size;
rc.variance_st_encoding_bitrate +=
(sum_bitrate2 / rc.window_size) *
(sum_bitrate2 / rc.window_size);
sum_bitrate2 = 0.0;
}
}
break;
default:
break;
}
}
++frame_cnt;
pts += frame_duration;
}
fclose(infile);
printout_rate_control_summary(&rc, &cfg, frame_cnt);
printf("\n");
printf("Frame cnt and encoding time/FPS stats for encoding: %d %f %f \n",
frame_cnt,
1000 * (float)cx_time / (double)(frame_cnt * 1000000),
1000000 * (double)frame_cnt / (double)cx_time);
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]);
vpx_img_free(&raw);
return EXIT_SUCCESS;
}
/*****************************************************************************
* Open: probe the decoder
*****************************************************************************/
static int Open(vlc_object_t *p_this)
{
decoder_t *dec = (decoder_t *)p_this;
const struct vpx_codec_iface *iface;
int vp_version;
switch (dec->fmt_in.i_codec)
{
#ifdef ENABLE_VP8_DECODER
case VLC_CODEC_VP8:
iface = &vpx_codec_vp8_dx_algo;
vp_version = 8;
break;
#endif
#ifdef ENABLE_VP9_DECODER
case VLC_CODEC_VP9:
iface = &vpx_codec_vp9_dx_algo;
vp_version = 9;
break;
#endif
default:
return VLC_EGENERIC;
}
decoder_sys_t *sys = malloc(sizeof(*sys));
if (!sys)
return VLC_ENOMEM;
dec->p_sys = sys;
struct vpx_codec_dec_cfg deccfg = {
.threads = __MIN(vlc_GetCPUCount(), 16)
};
msg_Dbg(p_this, "VP%d: using libvpx version %s (build options %s)",
vp_version, vpx_codec_version_str(), vpx_codec_build_config());
if (vpx_codec_dec_init(&sys->ctx, iface, &deccfg, 0) != VPX_CODEC_OK) {
const char *error = vpx_codec_error(&sys->ctx);
msg_Err(p_this, "Failed to initialize decoder: %s\n", error);
free(sys);
return VLC_EGENERIC;;
}
dec->pf_decode_video = Decode;
dec->fmt_out.i_cat = VIDEO_ES;
dec->fmt_out.video.i_width = dec->fmt_in.video.i_width;
dec->fmt_out.video.i_height = dec->fmt_in.video.i_height;
dec->fmt_out.i_codec = VLC_CODEC_I420;
return VLC_SUCCESS;
}
/*****************************************************************************
* Close: decoder destruction
*****************************************************************************/
static void Close(vlc_object_t *p_this)
{
decoder_t *dec = (decoder_t *)p_this;
decoder_sys_t *sys = dec->p_sys;
/* Free our PTS */
const void *iter = NULL;
for (;;) {
struct vpx_image *img = vpx_codec_get_frame(&sys->ctx, &iter);
if (!img)
break;
free(img->user_priv);
}
vpx_codec_destroy(&sys->ctx);
free(sys);
}
void clean_decoder(struct vpx_context *ctx)
{
vpx_codec_destroy(&ctx->codec);
free(ctx);
}
VPXDecoder::~VPXDecoder() {
if (_initialized)
vpx_codec_destroy(&_context);
}
int main(int argc, char **argv) {
FILE *infile, *outfile;
vpx_codec_ctx_t codec;
vpx_codec_enc_cfg_t cfg;
int frame_cnt = 0;
unsigned char file_hdr[IVF_FILE_HDR_SZ];
unsigned char frame_hdr[IVF_FRAME_HDR_SZ];
vpx_image_t raw;
vpx_codec_err_t ret;
int width,height;
int y_size;
int frame_avail;
int got_data;
int flags = 0;
width = 640;
height = 360;
infile = fopen("../cuc_ieschool_640x360_yuv420p.yuv", "rb");
outfile = fopen("cuc_ieschool.ivf", "wb");
if(infile==NULL||outfile==NULL){
printf("Error open files.\n");
return -1;
}
if(!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, width, height, 1)){
printf("Fail to allocate image\n");
return -1;
}
printf("Using %s\n",vpx_codec_iface_name(interface));
//Populate encoder configuration
ret = vpx_codec_enc_config_default(interface, &cfg, 0);
if(ret) {
printf("Failed to get config: %s\n", vpx_codec_err_to_string(ret));
return -1;
}
cfg.rc_target_bitrate =800;
cfg.g_w = width;
cfg.g_h = height;
write_ivf_file_header(outfile, &cfg, 0);
//Initialize codec
if(vpx_codec_enc_init(&codec, interface, &cfg, 0)){
printf("Failed to initialize encoder\n");
return -1;
}
frame_avail = 1;
got_data = 0;
y_size=cfg.g_w*cfg.g_h;
while(frame_avail || got_data) {
vpx_codec_iter_t iter = NULL;
const vpx_codec_cx_pkt_t *pkt;
if(fread(raw.planes[0], 1, y_size*3/2, infile)!=y_size*3/2){
frame_avail=0;
}
if(frame_avail){
//Encode
ret=vpx_codec_encode(&codec,&raw,frame_cnt,1,flags,VPX_DL_REALTIME);
}else{
//Flush Encoder
ret=vpx_codec_encode(&codec,NULL,frame_cnt,1,flags,VPX_DL_REALTIME);
}
if(ret){
printf("Failed to encode frame\n");
return -1;
}
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);
fwrite(pkt->data.frame.buf, 1, pkt->data.frame.sz,outfile);
break;
default:
break;
}
}
printf("Succeed encode frame: %5d\n",frame_cnt);
frame_cnt++;
}
fclose(infile);
vpx_codec_destroy(&codec);
//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 0;
}
SoftVPX::~SoftVPX() {
vpx_codec_destroy((vpx_codec_ctx_t *)mCtx);
delete (vpx_codec_ctx_t *)mCtx;
mCtx = NULL;
}
int main(int argc, char **argv) {
vpx_codec_ctx_t decoder;
vpx_codec_iface_t *iface = ifaces[0].iface;
vpx_codec_iter_t iter;
vpx_codec_dec_cfg_t cfg;
vpx_codec_err_t res = VPX_CODEC_OK;
unsigned int alloc_sz = 0;
unsigned int w = 352;
unsigned int h = 288;
int i;
exec_name = argv[0];
for (i = 1; i < argc; i++) {
if (!strcmp(argv[i], "--codec")) {
if (i + 1 < argc) {
int j, k = -1;
i++;
for (j = 0; j < sizeof(ifaces) / sizeof(ifaces[0]); j++)
if (!strcmp(ifaces[j].name, argv[i]))
k = j;
if (k >= 0)
iface = ifaces[k].iface;
else
usage_error("Error: Unrecognized argument (%s) to --codec\n",
argv[i]);
} else
usage_error("Error: Option --codec requires argument.\n");
} else if (!strcmp(argv[i], "-v"))
verbose = 1;
else if (!strcmp(argv[i], "-h"))
if (i + 1 < argc) {
h = atoi(argv[++i]);
} else
usage_error("Error: Option -h requires argument.\n");
else if (!strcmp(argv[i], "-w"))
if (i + 1 < argc) {
w = atoi(argv[++i]);
} else
usage_error("Error: Option -w requires argument.\n");
else if (!strcmp(argv[i], "--help"))
usage_exit();
else
usage_error("Error: Unrecognized option %s\n\n", argv[i]);
}
if (argc == 1)
printf("Using built-in defaults. For options, rerun with --help\n\n");
/* XMA mode is not supported on all decoders! */
if (!(vpx_codec_get_caps(iface) & VPX_CODEC_CAP_XMA)) {
printf("%s does not support XMA mode!\n", vpx_codec_iface_name(iface));
return EXIT_FAILURE;
}
/* The codec knows how much memory to allocate based on the size of the
* encoded frames. This data can be parsed from the bitstream with
* vpx_codec_peek_stream_info() if a bitstream is available. Otherwise,
* a fixed size can be used that will be the upper limit on the frame
* size the decoder can decode.
*/
cfg.w = w;
cfg.h = h;
/* Initialize the decoder in XMA mode. */
if (vpx_codec_dec_init(&decoder, iface, &cfg, VPX_CODEC_USE_XMA)) {
printf("Failed to initialize decoder in XMA mode: %s\n",
vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
/* Iterate through the list of memory maps, allocating them with the
* requested alignment.
*/
iter = NULL;
do {
vpx_codec_mmap_t mmap;
unsigned int align;
res = vpx_codec_get_mem_map(&decoder, &mmap, &iter);
align = mmap.align ? mmap.align - 1 : 0;
if (!res) {
if (verbose)
printf("Allocating segment %u, size %lu, align %u %s\n",
mmap.id, mmap.sz, mmap.align,
mmap.flags & VPX_CODEC_MEM_ZERO ? "(ZEROED)" : "");
if (mmap.flags & VPX_CODEC_MEM_ZERO)
mmap.priv = calloc(1, mmap.sz + align);
else
mmap.priv = malloc(mmap.sz + align);
mmap.base = (void *)((((uintptr_t)mmap.priv) + align) &
~(uintptr_t)align);
mmap.dtor = my_mem_dtor;
alloc_sz += mmap.sz + align;
if (vpx_codec_set_mem_map(&decoder, &mmap, 1)) {
printf("Failed to set mmap: %s\n", vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
} else if (res != VPX_CODEC_LIST_END) {
printf("Failed to get mmap: %s\n", vpx_codec_error(&decoder));
return EXIT_FAILURE;
}
} while (res != VPX_CODEC_LIST_END);
printf("%s\n %d bytes external memory required for %dx%d.\n",
decoder.name, alloc_sz, cfg.w, cfg.h);
vpx_codec_destroy(&decoder);
return EXIT_SUCCESS;
}
int main(int argc, const char **argv) {
AppInput app_input = {0};
VpxVideoWriter *writer = NULL;
VpxVideoInfo info = {0};
vpx_codec_ctx_t codec;
vpx_codec_enc_cfg_t enc_cfg;
SvcContext svc_ctx;
uint32_t i;
uint32_t frame_cnt = 0;
vpx_image_t raw;
vpx_codec_err_t res;
int pts = 0; /* PTS starts at 0 */
int frame_duration = 1; /* 1 timebase tick per frame */
FILE *infile = NULL;
int end_of_stream = 0;
int frames_received = 0;
memset(&svc_ctx, 0, sizeof(svc_ctx));
svc_ctx.log_print = 1;
exec_name = argv[0];
parse_command_line(argc, argv, &app_input, &svc_ctx, &enc_cfg);
// Allocate image buffer
if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, enc_cfg.g_w, enc_cfg.g_h, 32))
die("Failed to allocate image %dx%d\n", enc_cfg.g_w, enc_cfg.g_h);
if (!(infile = fopen(app_input.input_filename, "rb")))
die("Failed to open %s for reading\n", app_input.input_filename);
// Initialize codec
if (vpx_svc_init(&svc_ctx, &codec, vpx_codec_vp9_cx(), &enc_cfg) !=
VPX_CODEC_OK)
die("Failed to initialize encoder\n");
info.codec_fourcc = VP9_FOURCC;
info.time_base.numerator = enc_cfg.g_timebase.num;
info.time_base.denominator = enc_cfg.g_timebase.den;
if (vpx_svc_get_layer_resolution(&svc_ctx, svc_ctx.spatial_layers - 1,
(unsigned int *)&info.frame_width,
(unsigned int *)&info.frame_height) !=
VPX_CODEC_OK) {
die("Failed to get output resolution");
}
if (!(app_input.passes == 2 && app_input.pass == 1)) {
// We don't save the bitstream for the 1st pass on two pass rate control
writer = vpx_video_writer_open(app_input.output_filename, kContainerIVF,
&info);
if (!writer)
die("Failed to open %s for writing\n", app_input.output_filename);
}
// skip initial frames
for (i = 0; i < app_input.frames_to_skip; ++i)
vpx_img_read(&raw, infile);
// Encode frames
while (!end_of_stream) {
vpx_codec_iter_t iter = NULL;
const vpx_codec_cx_pkt_t *cx_pkt;
if (frame_cnt >= app_input.frames_to_code || !vpx_img_read(&raw, infile)) {
// We need one extra vpx_svc_encode call at end of stream to flush
// encoder and get remaining data
end_of_stream = 1;
}
res = vpx_svc_encode(&svc_ctx, &codec, (end_of_stream ? NULL : &raw),
pts, frame_duration, VPX_DL_GOOD_QUALITY);
printf("%s", vpx_svc_get_message(&svc_ctx));
if (res != VPX_CODEC_OK) {
die_codec(&codec, "Failed to encode frame");
}
while ((cx_pkt = vpx_codec_get_cx_data(&codec, &iter)) != NULL) {
switch (cx_pkt->kind) {
case VPX_CODEC_CX_FRAME_PKT: {
if (cx_pkt->data.frame.sz > 0)
vpx_video_writer_write_frame(writer,
cx_pkt->data.frame.buf,
cx_pkt->data.frame.sz,
cx_pkt->data.frame.pts);
printf("SVC frame: %d, kf: %d, size: %d, pts: %d\n", frames_received,
!!(cx_pkt->data.frame.flags & VPX_FRAME_IS_KEY),
(int)cx_pkt->data.frame.sz, (int)cx_pkt->data.frame.pts);
++frames_received;
break;
}
case VPX_CODEC_STATS_PKT: {
stats_write(&app_input.rc_stats,
cx_pkt->data.twopass_stats.buf,
cx_pkt->data.twopass_stats.sz);
break;
}
default: {
break;
}
}
}
if (!end_of_stream) {
++frame_cnt;
pts += frame_duration;
}
}
printf("Processed %d frames\n", frame_cnt);
fclose(infile);
if (vpx_codec_destroy(&codec)) die_codec(&codec, "Failed to destroy codec");
if (app_input.passes == 2)
stats_close(&app_input.rc_stats, 1);
if (writer) {
vpx_video_writer_close(writer);
}
vpx_img_free(&raw);
// display average size, psnr
printf("%s", vpx_svc_dump_statistics(&svc_ctx));
vpx_svc_release(&svc_ctx);
return EXIT_SUCCESS;
}
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;
}
/* 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;
if(frame_cnt + 1 == 22) { //
vpx_roi_map_t roi; //
int i; //
//
roi.rows = cfg.g_h/16; //
roi.cols = cfg.g_w/16; //
//
roi.delta_q[0] = 0; //
roi.delta_q[1] = -2; //
roi.delta_q[2] = -4; //
roi.delta_q[3] = -6; //
//
roi.delta_lf[0] = 0; //
roi.delta_lf[1] = 1; //
roi.delta_lf[2] = 2; //
roi.delta_lf[3] = 3; //
//
roi.static_threshold[0] = 1500; //
roi.static_threshold[1] = 1000; //
roi.static_threshold[2] = 500; //
roi.static_threshold[3] = 0; //
//
/* generate an ROI map for example */ //
roi.roi_map = malloc(roi.rows * roi.cols); //
for(i=0;i<roi.rows*roi.cols;i++) //
roi.roi_map[i] = i & 3; //
//
if(vpx_codec_control(&codec, VP8E_SET_ROI_MAP, &roi)) //
die_codec(&codec, "Failed to set ROI map"); //
//
free(roi.roi_map); //
} else if(frame_cnt + 1 == 33) { //
vpx_active_map_t active; //
int i; //
//
active.rows = cfg.g_h/16; //
active.cols = cfg.g_w/16; //
//
/* generate active map for example */ //
active.active_map = malloc(active.rows * active.cols); //
for(i=0;i<active.rows*active.cols;i++) //
active.active_map[i] = i & 1; //
//
if(vpx_codec_control(&codec, VP8E_SET_ACTIVEMAP, &active)) //
die_codec(&codec, "Failed to set active map"); //
//
free(active.active_map); //
} else if(frame_cnt + 1 == 44) { //
vpx_active_map_t active; //
//
active.rows = cfg.g_h/16; //
active.cols = cfg.g_w/16; //
//
/* pass in null map to disable active_map*/ //
active.active_map = NULL; //
//
if(vpx_codec_control(&codec, VP8E_SET_ACTIVEMAP, &active)) //
die_codec(&codec, "Failed to set active map"); //
} //
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);
if(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);
printf("Processed %d frames.\n",frame_cnt-1);
if(vpx_codec_destroy(&codec))
die_codec(&codec, "Failed to destroy codec");
/* 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;
}
~information()
{
vpx_img_free(&m_raw);
vpx_codec_destroy(&m_codec);
}
VCSession *vc_new(Mono_Time *mono_time, const Logger *log, ToxAV *av, uint32_t friend_number,
toxav_video_receive_frame_cb *cb, void *cb_data)
{
VCSession *vc = (VCSession *)calloc(sizeof(VCSession), 1);
vpx_codec_err_t rc;
if (!vc) {
LOGGER_WARNING(log, "Allocation failed! Application might misbehave!");
return nullptr;
}
if (create_recursive_mutex(vc->queue_mutex) != 0) {
LOGGER_WARNING(log, "Failed to create recursive mutex!");
free(vc);
return nullptr;
}
int cpu_used_value = VP8E_SET_CPUUSED_VALUE;
vc->vbuf_raw = rb_new(VIDEO_DECODE_BUFFER_SIZE);
if (!vc->vbuf_raw) {
goto BASE_CLEANUP;
}
/*
* VPX_CODEC_USE_FRAME_THREADING
* Enable frame-based multi-threading
*
* VPX_CODEC_USE_ERROR_CONCEALMENT
* Conceal errors in decoded frames
*/
vpx_codec_dec_cfg_t dec_cfg;
dec_cfg.threads = VPX_MAX_DECODER_THREADS; // Maximum number of threads to use
dec_cfg.w = VIDEO_CODEC_DECODER_MAX_WIDTH;
dec_cfg.h = VIDEO_CODEC_DECODER_MAX_HEIGHT;
LOGGER_DEBUG(log, "Using VP8 codec for decoder (0)");
rc = vpx_codec_dec_init(vc->decoder, video_codec_decoder_interface(), &dec_cfg,
VPX_CODEC_USE_FRAME_THREADING | VPX_CODEC_USE_POSTPROC);
if (rc == VPX_CODEC_INCAPABLE) {
LOGGER_WARNING(log, "Postproc not supported by this decoder (0)");
rc = vpx_codec_dec_init(vc->decoder, video_codec_decoder_interface(), &dec_cfg, VPX_CODEC_USE_FRAME_THREADING);
}
if (rc != VPX_CODEC_OK) {
LOGGER_ERROR(log, "Init video_decoder failed: %s", vpx_codec_err_to_string(rc));
goto BASE_CLEANUP;
}
if (VIDEO_VP8_DECODER_POST_PROCESSING_ENABLED == 1) {
vp8_postproc_cfg_t pp = {VP8_DEBLOCK, 1, 0};
vpx_codec_err_t cc_res = vpx_codec_control(vc->decoder, VP8_SET_POSTPROC, &pp);
if (cc_res != VPX_CODEC_OK) {
LOGGER_WARNING(log, "Failed to turn on postproc");
} else {
LOGGER_DEBUG(log, "turn on postproc: OK");
}
} else {
vp8_postproc_cfg_t pp = {0, 0, 0};
vpx_codec_err_t cc_res = vpx_codec_control(vc->decoder, VP8_SET_POSTPROC, &pp);
if (cc_res != VPX_CODEC_OK) {
LOGGER_WARNING(log, "Failed to turn OFF postproc");
} else {
LOGGER_DEBUG(log, "Disable postproc: OK");
}
}
/* Set encoder to some initial values
*/
vpx_codec_enc_cfg_t cfg;
vc_init_encoder_cfg(log, &cfg, 1);
LOGGER_DEBUG(log, "Using VP8 codec for encoder (0.1)");
rc = vpx_codec_enc_init(vc->encoder, video_codec_encoder_interface(), &cfg, VPX_CODEC_USE_FRAME_THREADING);
if (rc != VPX_CODEC_OK) {
LOGGER_ERROR(log, "Failed to initialize encoder: %s", vpx_codec_err_to_string(rc));
goto BASE_CLEANUP_1;
}
rc = vpx_codec_control(vc->encoder, VP8E_SET_CPUUSED, cpu_used_value);
if (rc != VPX_CODEC_OK) {
LOGGER_ERROR(log, "Failed to set encoder control setting: %s", vpx_codec_err_to_string(rc));
vpx_codec_destroy(vc->encoder);
goto BASE_CLEANUP_1;
}
/*
VPX_CTRL_USE_TYPE(VP8E_SET_NOISE_SENSITIVITY, unsigned int)
control function to set noise sensitivity
0: off, 1: OnYOnly, 2: OnYUV, 3: OnYUVAggressive, 4: Adaptive
*/
/*
rc = vpx_codec_control(vc->encoder, VP8E_SET_NOISE_SENSITIVITY, 2);
if (rc != VPX_CODEC_OK) {
LOGGER_ERROR(log, "Failed to set encoder control setting: %s", vpx_codec_err_to_string(rc));
vpx_codec_destroy(vc->encoder);
goto BASE_CLEANUP_1;
}
*/
vc->linfts = current_time_monotonic(mono_time);
vc->lcfd = 60;
vc->vcb = cb;
vc->vcb_user_data = cb_data;
vc->friend_number = friend_number;
vc->av = av;
vc->log = log;
return vc;
BASE_CLEANUP_1:
vpx_codec_destroy(vc->decoder);
BASE_CLEANUP:
pthread_mutex_destroy(vc->queue_mutex);
rb_kill(vc->vbuf_raw);
free(vc);
return nullptr;
}
static switch_status_t switch_vpx_decode(switch_codec_t *codec, switch_frame_t *frame)
{
vpx_context_t *context = (vpx_context_t *)codec->private_info;
switch_size_t len;
vpx_codec_ctx_t *decoder = NULL;
switch_status_t status = SWITCH_STATUS_SUCCESS;
int is_start = 0, is_keyframe = 0, get_refresh = 0;
if (context->is_vp9) {
is_start = is_keyframe = IS_VP9_KEY_FRAME(*(unsigned char *)frame->data);
} else { // vp8
is_start = (*(unsigned char *)frame->data & 0x10);
is_keyframe = IS_VP8_KEY_FRAME((uint8_t *)frame->data);
}
// if (is_keyframe) switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "got key %d\n", is_keyframe);
if (context->need_decoder_reset != 0) {
vpx_codec_destroy(&context->decoder);
context->decoder_init = 0;
init_decoder(codec);
context->need_decoder_reset = 0;
}
if (!context->decoder_init) {
init_decoder(codec);
}
if (!context->decoder_init) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_WARNING, "VPX decoder is not initialized!\n");
return SWITCH_STATUS_FALSE;
}
decoder = &context->decoder;
// switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_INFO, "len: %d ts: %u mark:%d\n", frame->datalen, frame->timestamp, frame->m);
// context->last_received_timestamp = frame->timestamp;
context->last_received_complete_picture = frame->m ? SWITCH_TRUE : SWITCH_FALSE;
if (is_start) {
context->got_start_frame = 1;
}
if (is_keyframe) {
if (context->got_key_frame <= 0) {
context->got_key_frame = 1;
if (!is_keyframe) {
get_refresh = 1;
}
} else {
context->got_key_frame++;
}
} else if (context->got_key_frame <= 0) {
if ((--context->got_key_frame % 200) == 0) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG1, "Waiting for key frame %d\n", context->got_key_frame);
}
if (!context->got_start_frame) {
switch_goto_status(SWITCH_STATUS_MORE_DATA, end);
}
}
status = context->is_vp9 ? buffer_vp9_packets(context, frame) : buffer_vp8_packets(context, frame);
if (context->dec_iter && (frame->img = (switch_image_t *) vpx_codec_get_frame(decoder, &context->dec_iter))) {
switch_goto_status(SWITCH_STATUS_SUCCESS, end);
}
//printf("READ buf:%ld got_key:%d st:%d m:%d\n", switch_buffer_inuse(context->vpx_packet_buffer), context->got_key_frame, status, frame->m);
len = switch_buffer_inuse(context->vpx_packet_buffer);
//if (frame->m && (status != SWITCH_STATUS_SUCCESS || !len)) {
//switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_ERROR, "WTF????? %d %ld\n", status, len);
//}
if (status == SWITCH_STATUS_SUCCESS && frame->m && len) {
uint8_t *data;
int corrupted = 0;
int err;
switch_buffer_peek_zerocopy(context->vpx_packet_buffer, (void *)&data);
context->dec_iter = NULL;
err = vpx_codec_decode(decoder, data, (unsigned int)len, NULL, 0);
if (err != VPX_CODEC_OK) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_DEBUG1, "Error decoding %" SWITCH_SIZE_T_FMT " bytes, [%d:%s:%s]\n",
len, err, vpx_codec_error(decoder), vpx_codec_error_detail(decoder));
switch_goto_status(SWITCH_STATUS_RESTART, end);
}
if (vpx_codec_control(decoder, VP8D_GET_FRAME_CORRUPTED, &corrupted) != VPX_CODEC_OK) {
switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_WARNING, "VPX control error!\n");
switch_goto_status(SWITCH_STATUS_RESTART, end);
}
if (corrupted) {
frame->img = NULL;
} else {
frame->img = (switch_image_t *) vpx_codec_get_frame(decoder, &context->dec_iter);
}
switch_buffer_zero(context->vpx_packet_buffer);
if (!frame->img) {
//context->need_decoder_reset = 1;
context->got_key_frame = 0;
context->got_start_frame = 0;
status = SWITCH_STATUS_RESTART;
}
}
end:
if (status == SWITCH_STATUS_RESTART) {
switch_buffer_zero(context->vpx_packet_buffer);
//context->need_decoder_reset = 1;
context->got_key_frame = 0;
context->got_start_frame = 0;
//switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_WARNING, "RESET VPX\n");
}
if (!frame->img || status == SWITCH_STATUS_RESTART) {
status = SWITCH_STATUS_MORE_DATA;
}
if (context->got_key_frame <= 0 || get_refresh) {
switch_set_flag(frame, SFF_WAIT_KEY_FRAME);
}
return status;
}
