I would like to test libvpx (vp9) temporal svc. For that matter I take examples/vpx_temporal_svc_encoder.c and modify it to add a decoder. If I take for example mode 3 with 3 temporal layers and 6 frame period and try to decode only layer 0 or layers 0 and 1 I get an error: "Failed to decode frame.: Corrupt frame detected. Failed to decode tile data". What am I missing ? Here is my code (had to remove unmodified part due to post limits):
/*
* Copyright (c) 2012 The WebM project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
// This is an example demonstrating how to implement a multi-layer VPx
// encoding scheme based on temporal scalability for video applications
// that benefit from a scalable bitstream.
#include <assert.h>
#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "./vpx_config.h"
#include "./y4minput.h"
#include "../vpx_ports/vpx_timer.h"
#include "vpx/vp8cx.h"
#include "vpx/vpx_encoder.h"
#include "vpx_ports/bitops.h"
#include "../tools_common.h"
#include "../video_writer.h"
#include "vpx/vpx_decoder.h"
#define ROI_MAP 0
#define zero(Dest) memset(&(Dest), 0, sizeof(Dest));
static const char *exec_name;
void usage_exit(void) { exit(EXIT_FAILURE); }
// Denoiser states for vp8, for temporal denoising.
enum denoiserStateVp8 {
kVp8DenoiserOff,
kVp8DenoiserOnYOnly,
kVp8DenoiserOnYUV,
kVp8DenoiserOnYUVAggressive,
kVp8DenoiserOnAdaptive
};
// Denoiser states for vp9, for temporal denoising.
enum denoiserStateVp9 {
kVp9DenoiserOff,
kVp9DenoiserOnYOnly,
// For SVC: denoise the top two spatial layers.
kVp9DenoiserOnYTwoSpatialLayers
};
static int mode_to_num_layers[13] = { 1, 2, 2, 3, 3, 3, 3, 5, 2, 3, 3, 3, 3 };
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;
uint32_t error_resilient = 0;
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;
#if ROI_MAP
vpx_roi_map_t roi;
#endif
vpx_svc_layer_id_t layer_id;
const VpxInterface *encoder = NULL;
struct VpxInputContext input_ctx;
struct RateControlMetrics rc;
int64_t cx_time = 0;
const int min_args_base = 13;
#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;
zero(rc.layer_target_bitrate);
memset(&layer_id, 0, sizeof(vpx_svc_layer_id_t));
memset(&input_ctx, 0, sizeof(input_ctx));
/* Setup default input stream settings */
input_ctx.framerate.numerator = 30;
input_ctx.framerate.denominator = 1;
input_ctx.only_i420 = 1;
input_ctx.bit_depth = 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> "
"<error_resilient> <threads> <mode> "
"<Rate_0> ... <Rate_nlayers-1> <bit-depth> \n",
argv[0]);
#else
//./vpx_temporal_svc_encoder ../9.yuv svc vp9 256 256 1 25 9 0 0 1 3 500000 500000 50000
die("Usage: %s <infile> <outfile> <codec_type(vp8/vp9)> <width> <height> "
"<rate_num> <rate_den> <speed> <frame_drop_threshold> "
"<error_resilient> <threads> <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 = (unsigned int)strtoul(argv[4], NULL, 0);
height = (unsigned int)strtoul(argv[5], NULL, 0);
if (width < 16 || width % 2 || height < 16 || height % 2) {
die("Invalid resolution: %d x %d", width, height);
}
layering_mode = (int)strtol(argv[12], NULL, 0);
if (layering_mode < 0 || layering_mode > 13) {
die("Invalid layering mode (0..12) %s", argv[12]);
}
if (argc != min_args + mode_to_num_layers[layering_mode]) {
die("Invalid number of arguments");
}
input_ctx.filename = argv[1];
open_input_file(&input_ctx);
#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]);
}
// Y4M reader has its own allocation.
if (input_ctx.file_type != FILE_TYPE_Y4M) {
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
// Y4M reader has its own allocation.
if (input_ctx.file_type != FILE_TYPE_Y4M) {
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 = (int)strtol(argv[6], NULL, 0);
cfg.g_timebase.den = (int)strtol(argv[7], NULL, 0);
speed = (int)strtol(argv[8], NULL, 0);
if (speed < 0) {
die("Invalid speed setting: must be positive");
}
if (strncmp(encoder->name, "vp9", 3) == 0 && speed > 9) {
warn("Mapping speed %d to speed 9.\n", speed);
}
for (i = min_args_base;
(int)i < min_args_base + mode_to_num_layers[layering_mode]; ++i) {
rc.layer_target_bitrate[i - 13] = (int)strtol(argv[i], NULL, 0);
if (strncmp(encoder->name, "vp8", 3) == 0)
cfg.ts_target_bitrate[i - 13] = rc.layer_target_bitrate[i - 13];
else if (strncmp(encoder->name, "vp9", 3) == 0)
cfg.layer_target_bitrate[i - 13] = rc.layer_target_bitrate[i - 13];
}
// Real time parameters.
cfg.rc_dropframe_thresh = (unsigned int)strtoul(argv[9], NULL, 0);
cfg.rc_end_usage = VPX_CBR;
cfg.rc_min_quantizer = 2;
cfg.rc_max_quantizer = 56;
if (strncmp(encoder->name, "vp9", 3) == 0) cfg.rc_max_quantizer = 52;
cfg.rc_undershoot_pct = 50;
cfg.rc_overshoot_pct = 50;
cfg.rc_buf_initial_sz = 600;
cfg.rc_buf_optimal_sz = 600;
cfg.rc_buf_sz = 1000;
// Disable dynamic resizing by default.
cfg.rc_resize_allowed = 0;
// Use 1 thread as default.
cfg.g_threads = (unsigned int)strtoul(argv[11], NULL, 0);
error_resilient = (uint32_t)strtoul(argv[10], NULL, 0);
if (error_resilient != 0 && error_resilient != 1) {
die("Invalid value for error resilient (0, 1): %d.", error_resilient);
}
// Enable error resilient mode.
cfg.g_error_resilient = error_resilient;
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;
cfg.temporal_layering_mode = VP9E_TEMPORAL_LAYERING_MODE_BYPASS;
set_temporal_layer_pattern(layering_mode, &cfg, layer_flags,
&flag_periodicity);
set_rate_control_metrics(&rc, &cfg);
if (input_ctx.file_type == FILE_TYPE_Y4M) {
if (input_ctx.width != cfg.g_w || input_ctx.height != cfg.g_h) {
die("Incorrect width or height: %d x %d", cfg.g_w, cfg.g_h);
}
if (input_ctx.framerate.numerator != cfg.g_timebase.den ||
input_ctx.framerate.denominator != cfg.g_timebase.num) {
die("Incorrect framerate: numerator %d denominator %d",
cfg.g_timebase.num, cfg.g_timebase.den);
}
}
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("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, kVp8DenoiserOff);
vpx_codec_control(&codec, VP8E_SET_STATIC_THRESHOLD, 1);
vpx_codec_control(&codec, VP8E_SET_GF_CBR_BOOST_PCT, 0);
#if ROI_MAP
set_roi_map(encoder->name, &cfg, &roi);
if (vpx_codec_control(&codec, VP8E_SET_ROI_MAP, &roi))
die_codec(&codec, "Failed to set ROI map");
#endif
} else if (strncmp(encoder->name, "vp9", 3) == 0) {
vpx_svc_extra_cfg_t svc_params;
memset(&svc_params, 0, sizeof(svc_params));
vpx_codec_control(&codec, VP9E_SET_POSTENCODE_DROP, 0);
vpx_codec_control(&codec, VP9E_SET_DISABLE_OVERSHOOT_MAXQ_CBR, 0);
vpx_codec_control(&codec, VP8E_SET_CPUUSED, speed);
vpx_codec_control(&codec, VP9E_SET_AQ_MODE, 3);
vpx_codec_control(&codec, VP9E_SET_GF_CBR_BOOST_PCT, 0);
vpx_codec_control(&codec, VP9E_SET_FRAME_PARALLEL_DECODING, 0);
vpx_codec_control(&codec, VP9E_SET_FRAME_PERIODIC_BOOST, 0);
vpx_codec_control(&codec, VP9E_SET_NOISE_SENSITIVITY, kVp9DenoiserOff);
vpx_codec_control(&codec, VP8E_SET_STATIC_THRESHOLD, 1);
vpx_codec_control(&codec, VP9E_SET_TUNE_CONTENT, 0);
vpx_codec_control(&codec, VP9E_SET_TILE_COLUMNS, get_msb(cfg.g_threads));
vpx_codec_control(&codec, VP9E_SET_DISABLE_LOOPFILTER, 0);
#if ROI_MAP
set_roi_map(encoder->name, &cfg, &roi);
if (vpx_codec_control(&codec, VP9E_SET_ROI_MAP, &roi))
die_codec(&codec, "Failed to set ROI map");
vpx_codec_control(&codec, VP9E_SET_AQ_MODE, 0);
#endif
if (cfg.g_threads > 1)
vpx_codec_control(&codec, VP9E_SET_ROW_MT, 1);
else
vpx_codec_control(&codec, VP9E_SET_ROW_MT, 0);
if (vpx_codec_control(&codec, VP9E_SET_SVC, layering_mode > 0 ? 1 : 0))
die_codec(&codec, "Failed to set SVC");
for (i = 0; i < cfg.ts_number_layers; ++i) {
svc_params.max_quantizers[i] = cfg.rc_max_quantizer;
svc_params.min_quantizers[i] = cfg.rc_min_quantizer;
}
svc_params.scaling_factor_num[0] = cfg.g_h;
svc_params.scaling_factor_den[0] = cfg.g_h;
vpx_codec_control(&codec, VP9E_SET_SVC_PARAMETERS, &svc_params);
}
if (strncmp(encoder->name, "vp8", 3) == 0) {
vpx_codec_control(&codec, VP8E_SET_SCREEN_CONTENT_MODE, 0);
}
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 = 1000;
vpx_codec_control(&codec, VP8E_SET_MAX_INTRA_BITRATE_PCT,
max_intra_size_pct);
}
const VpxInterface *decoder = NULL;
decoder = get_vpx_decoder_by_name( "vp9" );
if (!decoder) die("Unsupported codec.");
vpx_codec_ctx_t codec_dec;
vpx_codec_dec_cfg_t cfg_dec;
cfg_dec.w = width;
cfg_dec.h = height;
cfg_dec.threads = 1;
if (vpx_codec_dec_init( &codec_dec, decoder->codec_interface(), &cfg_dec, 0))
die_codec( &codec_dec, "Failed to initialize decoder.");
if(!codec.iface)
die_codec( &codec_dec, "vvv no iface" );
vpx_image_t *img_dec;
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];
layer_id.temporal_layer_id_per_spatial[0] = layer_id.temporal_layer_id;
if (strncmp(encoder->name, "vp9", 3) == 0) {
vpx_codec_control(&codec, VP9E_SET_SVC_LAYER_ID, &layer_id);
} else if (strncmp(encoder->name, "vp8", 3) == 0) {
vpx_codec_control(&codec, VP8E_SET_TEMPORAL_LAYER_ID,
layer_id.temporal_layer_id);
}
flags = layer_flags[frame_cnt % flag_periodicity];
if (layering_mode == 0) flags = 0;
frame_avail = read_frame(&input_ctx, &raw);
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 (rc.window_size == 0) rc.window_size = 15;
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;
}
}
//decode
if( layer_id.temporal_layer_id == 0 ){
vpx_codec_control(&codec_dec, VP9E_SET_SVC_LAYER_ID, &layer_id);
vpx_codec_err_t res = vpx_codec_decode(&codec_dec, pkt->data.frame.buf, pkt->data.frame.sz, NULL, 0);
if( res )
die_codec(&codec_dec, "Failed to decode frame.");
vpx_codec_iter_t iter = NULL;
img_dec = vpx_codec_get_frame(&codec_dec, &iter);
if( img_dec == NULL )
die_codec(&codec_dec, "No frames were decoded.");
if( vpx_codec_get_frame(&codec_dec, &iter) != NULL )
die_codec(&codec_dec, "More than on frame decoded.");
printf( "decoded frame\n" );
}
break;
default: break;
}
}
++frame_cnt;
pts += frame_duration;
}
close_input_file(&input_ctx);
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]);
if (input_ctx.file_type != FILE_TYPE_Y4M) {
vpx_img_free(&raw);
}
#if ROI_MAP
free(roi.roi_map);
#endif
return EXIT_SUCCESS;
}
Related
I have been working on a game that is a Stacker. Everything works fine, but after you have played and started over the blocks from the previous game still remain there.
Could anyone help me with this problem?
#include <SDL/SDL.h>
#include <SDL/SDL_image.h>
#include <SDL/SDL_mixer.h>
#include <stdio.h>
#include <stdlib.h>
#define BASE_TIME_INTERVAL 80
#define SPEED_INCREASE 1.03
#define GAME_ROWS 15
#define GAME_COLUMNS 7
#define MIN(a, b) ((a < b) ? a : b)
#define MAX(a, b) ((a > b) ? a : b)
int array_matrix[GAME_ROWS][GAME_COLUMNS];
//void save();
int credit = 0;
int in = 0;
int out = 0;
int inGame = 0;
static SDL_Surface *screen;
//Sprites
static SDL_Surface *square;
static SDL_Surface *background;
static SDL_Surface *grid;
static SDL_Surface *main_ui;
static SDL_Surface *award;
//Text Credits Sprites
static SDL_Surface *credits;
static SDL_Surface *number;
//Sounds
Mix_Chunk *soundPlace = NULL;
Mix_Chunk *soundGameOver = NULL;
Mix_Chunk *soundCredit = NULL;
FILE * fptr;
void print_board() {
int i, j;
SDL_Rect src, dest;
for (i = 0; i < 15; i++) {
for (j = 0; j < 7 ; j++) {
if (array_matrix[i][j] == 1) {
src.x = 0;
src.y = 0;
src.w = 65;
src.h = 65;
dest.x = j * 67 + 227;
dest.y = i * 67 + 240;
dest.w = 65;
dest.h = 65;
SDL_BlitSurface(square, &src, screen, &dest);
}
}
}
}
void update_board(int x_pos, int length, int level) {
int underflow_ammt = length - 1;
int j;
if (x_pos < underflow_ammt)
length = length - (underflow_ammt - x_pos);
x_pos = MAX(0, x_pos-underflow_ammt);
for (j = 0; j < GAME_COLUMNS; j++)
array_matrix[GAME_ROWS - level][j] = 0;
for (j = x_pos; j < x_pos + length; j++) {
array_matrix[GAME_ROWS - level][MIN(j, GAME_COLUMNS-1)] = 1;
}
}
int get_new_length(int level) {
int i;
int length = 0;
for (i = 0; i < GAME_COLUMNS; i++) {
if (array_matrix[GAME_ROWS - level][i] == 1) {
length++;
if (level != 1) {
if (array_matrix[GAME_ROWS - (level - 1)][i] != 1) {
array_matrix[GAME_ROWS - level][i] = 0;
length--;
}
}
}
if ((level == 4 && length == 3) || (level == 10 && length == 2)) {
length--;
}
}
return length;
}
void draw_background(){
SDL_BlitSurface(background, NULL, screen, NULL);
SDL_BlitSurface(main_ui, NULL, screen, NULL);
SDL_BlitSurface(award, NULL, screen, NULL);
SDL_BlitSurface(grid, NULL, screen, NULL);
}
void draw_credits(){
SDL_Rect destCredits;
SDL_Rect destNumber;
switch (credit)
{
case 0:
number = IMG_Load("assets/0.png");
break;
case 1:
number = IMG_Load("assets/1.png");
break;
case 2:
number = IMG_Load("assets/2.png");
break;
case 3:
number = IMG_Load("assets/3.png");
break;
case 4:
number = IMG_Load("assets/4.png");
break;
case 5:
number = IMG_Load("assets/5.png");
break;
case 6:
number = IMG_Load("assets/6.png");
break;
case 7:
number = IMG_Load("assets/7.png");
break;
case 8:
number = IMG_Load("assets/8.png");
break;
case 9:
number = IMG_Load("assets/9.png");
break;
}
if (number == NULL) {
printf("Unable to load number png.\n");
}
destCredits.x = 300;
destCredits.y = 1300;
destNumber.x = 550;
destNumber.y = 1305;
SDL_BlitSurface(credits, NULL, screen, &destCredits);
SDL_BlitSurface(number, NULL, screen, &destNumber);
}
void game_loop() {
int time_delay = BASE_TIME_INTERVAL;
int left_or_right = 1;
int current_level = 1;
int length = 3;
int x_pos = 0;
int quit = 0;
int underflow_ammt = length - 1;
SDL_Event event;
while (!quit) {
while (SDL_PollEvent(&event)) {
switch(event.type) {
case SDL_KEYDOWN:
if (event.key.keysym.sym == SDLK_SPACE && inGame == 1) {
length = get_new_length(current_level);
underflow_ammt = length - 1;
if (current_level >= 15 || length == 0) {
Mix_PlayChannel( -1, soundGameOver, 0 );
inGame = 0;
time_delay = BASE_TIME_INTERVAL;
left_or_right = 1;
current_level = 1;
length = 3;
x_pos = 0;
underflow_ammt = length - 1;
}
else{
Mix_PlayChannel( -1, soundPlace, 0 );
current_level++;
time_delay = time_delay/SPEED_INCREASE;
}
}
if (event.key.keysym.sym == SDLK_2 && credit < 9){
credit++;
in++;
save();
Mix_PlayChannel( -1, soundCredit, 0 );
}
if (event.key.keysym.sym == SDLK_1 && credit > 0 && inGame == 0){
credit--;
out++;
save();
Mix_PlayChannel( -1, soundCredit, 0 );
inGame = 1;
}
if (event.key.keysym.sym == SDLK_ESCAPE){
quit = 1;
exit(0);
}
break;
case SDL_QUIT:
quit = 1;
exit(0);
break;
}
}
if (!quit) {
SDL_FillRect(screen, NULL, 0x000000);
if (x_pos >= GAME_COLUMNS + (underflow_ammt - 1))
left_or_right = -1;
if (x_pos <= 0)
left_or_right = 1;
update_board(x_pos, length, current_level);
draw_background();
if (inGame == 1)
print_board();
draw_credits();
SDL_Flip(screen);
x_pos = x_pos + left_or_right;
SDL_Delay(time_delay);
}
}
}
/*
void save(){
fptr = fopen("data.xml", "w");
fprintf(fptr,"<?xml version=\"1.0\" encoding=\"utf-8\"?>\n");
fprintf(fptr,"<Data>\n");
fprintf(fptr," <credits>%d</credits>\n",credit);
fprintf(fptr," <IN>%d</IN>\n",in);
fprintf(fptr," <OUT>%d</OUT>\n",out);
fprintf(fptr,"</Data>\n");
fclose(fptr);
}
*/
void init(){
if (SDL_Init(SDL_INIT_VIDEO) != 0) {
printf("Unable to initialize SDL: %s\n", SDL_GetError());
}
screen = SDL_SetVideoMode(768, 1366, 16, SDL_DOUBLEBUF | SDL_HWSURFACE | SDL_FULLSCREEN);
if (screen == NULL) {
printf("Unable to set video mode: %s\n", SDL_GetError());
}
int imgFlags = IMG_INIT_PNG;
if( !( IMG_Init( imgFlags ) & imgFlags ) )
{
printf( "SDL_image could not initialize! SDL_image Error: %s\n", IMG_GetError() );
}
if (SDL_Init(SDL_INIT_AUDIO) < 0){
printf("Unable to set audio mode: %s\n", SDL_GetError());
}
//Initialize SDL_mixer
if( Mix_OpenAudio( 44100, MIX_DEFAULT_FORMAT, 2, 2048 ) < 0 )
printf( "SDL_mixer could not initialize! SDL_mixer Error: %s\n", Mix_GetError());
}
void loadMedia(){
background = IMG_Load("assets/background.png");
if (background == NULL)
printf("Unable to load background png.\n");
square = IMG_Load("assets/square.png");
if (square == NULL)
printf("Unable to load square png.\n");
credits = IMG_Load("assets/credits.png");
if (credits == NULL)
printf("Unable to load credits png.\n");
grid = IMG_Load("assets/grid.png");
if (grid == NULL)
printf("Unable to load grid png.\n");
main_ui = IMG_Load("assets/main_ui.png");
if (main_ui == NULL)
printf("Unable to load main_ui png.\n");
soundPlace = Mix_LoadWAV("assets/place.wav");
if(soundPlace == NULL)
printf( "Failed to load place sound effect! SDL_mixer Error: %s\n", Mix_GetError() );
soundGameOver = Mix_LoadWAV("assets/gameover.wav");
if(soundGameOver == NULL)
printf( "Failed to load gameover sound effect! SDL_mixer Error: %s\n", Mix_GetError() );
soundCredit = Mix_LoadWAV("assets/credit.wav");
if( soundCredit == NULL )
printf( "Failed to load credit sound effect! SDL_mixer Error: %s\n", Mix_GetError() );
}
void close(){
SDL_FreeSurface(square);
SDL_FreeSurface(background);
SDL_FreeSurface(credits);
SDL_FreeSurface(grid);
SDL_FreeSurface(main_ui);
SDL_FreeSurface(number);
Mix_FreeChunk(soundPlace);
Mix_FreeChunk(soundGameOver);
Mix_FreeChunk(soundCredit);
Mix_Quit();
IMG_Quit();
SDL_Quit();
}
int main(int argc, char *argv[])
{
init();
loadMedia();
game_loop();
close();
return 0;
}
I using:
SDL 1.2.15
SDL_image 1.2.12
SDL_mixer 1.2.12
I try to use ffmpeg and h264 codec to translate the video in realtime. But at the state of decoding encoded frame, I get some "bad" image.
Init encoder and decoder:
VCSession *vc_new_x264(Logger *log, ToxAV *av, uint32_t friend_number, toxav_video_receive_frame_cb *cb, void *cb_data,
VCSession *vc)
{
// ------ ffmpeg encoder ------
AVCodec *codec2 = NULL;
vc->h264_encoder_ctx = NULL;//AVCodecContext type
codec2 = NULL;
avcodec_register_all();
codec2 = avcodec_find_encoder(AV_CODEC_ID_H264);
if (codec2 == NULL)
{
LOGGER_WARNING(log, "h264: not find encoder");
}
vc->h264_encoder_ctx = avcodec_alloc_context3(codec2);
vc->h264_out_pic2 = av_packet_alloc();
vc->h264_encoder_ctx->bit_rate = 10 *1000 * 1000;
vc->h264_encoder_ctx->width = 800;
vc->h264_encoder_ctx->height = 600;
vc->h264_enc_width = vc->h264_encoder_ctx->width;
vc->h264_enc_height = vc->h264_encoder_ctx->height;
vc->h264_encoder_ctx->time_base = (AVRational) {
1, 30
};
vc->h264_encoder_ctx->gop_size = 30;
vc->h264_encoder_ctx->max_b_frames = 1;
vc->h264_encoder_ctx->pix_fmt = AV_PIX_FMT_YUV420P;
av_opt_set(vc->h264_encoder_ctx->priv_data, "preset", "veryfast", 0);
av_opt_set(vc->h264_encoder_ctx->priv_data, "annex_b", "1", 0);
av_opt_set(vc->h264_encoder_ctx->priv_data, "repeat_headers", "1", 0);
av_opt_set(vc->h264_encoder_ctx->priv_data, "tune", "zerolatency", 0);
av_opt_set_int(vc->h264_encoder_ctx->priv_data, "zerolatency", 1, 0);
vc->h264_encoder_ctx->time_base.num = 1;
vc->h264_encoder_ctx->time_base.den = 1000;
vc->h264_encoder_ctx->framerate = (AVRational) {
1000, 40
};
AVDictionary *opts = NULL;
if (avcodec_open2(vc->h264_encoder_ctx, codec2, &opts) < 0) {
LOGGER_ERROR(log, "could not open codec H264 on encoder");
}
av_dict_free(&opts);
AVCodec *codec = NULL;
vc->h264_decoder_ctx = NULL;// AVCodecContext - type
codec = NULL;
codec = avcodec_find_decoder(AV_CODEC_ID_H264);
if (!codec) {
LOGGER_WARNING(log, "codec not found H264 on decoder");
}
vc->h264_decoder_ctx = avcodec_alloc_context3(codec);
if (codec->capabilities & AV_CODEC_CAP_TRUNCATED) {
vc->h264_decoder_ctx->flags |= AV_CODEC_FLAG_TRUNCATED; /* we do not send complete frames */
}
if (codec->capabilities & AV_CODEC_FLAG_LOW_DELAY) {
vc->h264_decoder_ctx->flags |= AV_CODEC_FLAG_LOW_DELAY;
}
vc->h264_decoder_ctx->flags |= AV_CODEC_FLAG2_SHOW_ALL;
vc->h264_decoder_ctx->refcounted_frames = 0;
vc->h264_decoder_ctx->delay = 0;
vc->h264_decoder_ctx->sw_pix_fmt = AV_PIX_FMT_YUV420P;
av_opt_set_int(vc->h264_decoder_ctx->priv_data, "delay", 0, AV_OPT_SEARCH_CHILDREN);
vc->h264_decoder_ctx->time_base = (AVRational) {
40, 1000
};
vc->h264_decoder_ctx->framerate = (AVRational) {
1000, 40
};
if (avcodec_open2(vc->h264_decoder_ctx, codec, NULL) < 0) {
LOGGER_WARNING(log, "could not open codec H264 on decoder");
}
vc->h264_decoder_ctx->refcounted_frames = 0;
return vc;
}
Encoding (in this function i encode frame and for debugging decode and save him in file):
uint32_t encode_frame_h264_p(ToxAV *av, uint32_t friend_number, uint16_t width, uint16_t height,
const uint8_t *y,
const uint8_t *u, const uint8_t *v, ToxAVCall *call,
uint64_t *video_frame_record_timestamp,
int vpx_encode_flags,
x264_nal_t **nal,
int *i_frame_size)
{
AVFrame *frame;
int ret;
uint32_t result = 1;
frame = av_frame_alloc();
frame->format = call->video->h264_encoder_ctx->pix_fmt;
frame->width = width;
frame->height = height;
ret = av_frame_get_buffer(frame, 32);
if (ret < 0) {
LOGGER_ERROR(av->m->log, "av_frame_get_buffer:Could not allocate the video frame data");
}
/* make sure the frame data is writable */
ret = av_frame_make_writable(frame);
if (ret < 0) {
LOGGER_ERROR(av->m->log, "av_frame_make_writable:ERROR");
}
frame->pts = (int64_t)(*video_frame_record_timestamp);
// copy YUV frame data into buffers
memcpy(frame->data[0], y, width * height);
memcpy(frame->data[1], u, (width / 2) * (height / 2));
memcpy(frame->data[2], v, (width / 2) * (height / 2));
// encode the frame
ret = avcodec_send_frame(call->video->h264_encoder_ctx, frame);
if (ret < 0) {
LOGGER_ERROR(av->m->log, "Error sending a frame for encoding:ERROR");
}
ret = avcodec_receive_packet(call->video->h264_encoder_ctx, call->video->h264_out_pic2);
if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF) {
*i_frame_size = 0;
} else if (ret < 0) {
*i_frame_size = 0;
// fprintf(stderr, "Error during encoding\n");
} else {
// Decoded encoded frame and save him to file
saveInFile(call->video->h264_decoder_ctx, frame, call->video->h264_out_pic2, "/home/user/testSave");
// printf("Write packet %3"PRId64" (size=%5d)\n", call->video->h264_out_pic2->pts, call->video->h264_out_pic2->size);
// fwrite(call->video->h264_out_pic2->data, 1, call->video->h264_out_pic2->size, outfile);
global_encoder_delay_counter++;
if (global_encoder_delay_counter > 60) {
global_encoder_delay_counter = 0;
LOGGER_DEBUG(av->m->log, "enc:delay=%ld",
(long int)(frame->pts - (int64_t)call->video->h264_out_pic2->pts)
);
}
*i_frame_size = call->video->h264_out_pic2->size;
*video_frame_record_timestamp = (uint64_t)call->video->h264_out_pic2->pts;
result = 0;
}
av_frame_free(&frame);
return result;
}
Decode and save frame code:
void saveInFile(AVCodecContext *dec_ctx, AVFrame *frame, AVPacket *pkt, const char *filename)
{
if (!pkt)
return;
char buf[1024];
int ret;
static int curNumber = 0;
ret = avcodec_send_packet(dec_ctx, pkt);
if (ret < 0 && ret != AVERROR_EOF)
{
fprintf(stderr, "Error sending a packet for decoding'\n");
if ( ret == AVERROR(EAGAIN))
return;
if (ret == AVERROR(EINVAL))
return;
if (ret == AVERROR(ENOMEM))
return;
}
ret = avcodec_receive_frame(dec_ctx, frame);
if (ret == AVERROR(EAGAIN) )
return;
if (ret == AVERROR_EOF)
{
return;
}
else if (ret < 0)
{
fprintf(stderr, "Error during decoding\n");
}
printf("saving frame %3d\n", dec_ctx->frame_number);
sprintf(buf, "%s%d", filename, curNumber);
curNumber++;
pgm_save(frame->data[0], frame->linesize[0], frame->width, frame->height, buf);
}
void pgm_save(unsigned char* buf, int wrap, int xsize, int ysize, char *filename)
{
FILE *f;
int i;
f = fopen(filename, "w");
fprintf(f, "P5\n%d %d\n%d\n", xsize, ysize, 255);
for (i =0; i < ysize; i++)
fwrite(buf + i* wrap, 1, xsize, f);
fclose(f);
}
After this manipulation I have smth like that:
I'm trying to transcode video files into a standard format. This is the main loop I have:
for(frame_count = 0; av_read_frame(input_ctx, &in_packet) >= 0; frame_count++) {
if(in_packet.stream_index == video_stream_index) {
decodedPacket = 0;
rc = avcodec_decode_video2(video_in_codec, inputFrame, &decodedPacket, &in_packet);
if(decodedPacket) {
out_frames++;
rc = sws_scale(sws_ctx, inputFrame->data, inputFrame->linesize, 0, video_out_codec->height, outputFrame->data, outputFrame->linesize);
if(rc != video_out_codec->height) {
puts("scaling error");
}
outputFrame->pts = (1.0 / 30.0) * 90.0 * video_out_codec->frame_number;
rc = avcodec_encode_video2(video_out_codec, &out_packet, outputFrame, &encodedPacket);
if(rc != 0) {
puts("encoding error");
}
if(encodedPacket) {
if (video_out_stream->codec->coded_frame->key_frame) { // deprecated, what to use instead
out_packet.flags |= AV_PKT_FLAG_KEY;
}
out_packet.stream_index = 0;
rc = av_interleaved_write_frame(output_ctx, &out_packet);
if(rc != 0) {
puts("frame write error");
}
av_free_packet(&out_packet);
memset(&out_packet, 0, sizeof(AVPacket));
packet_count++;
}
av_free_packet(&in_packet);
}
}
...
When I run this I get strange values for tbr/tbn/tbc reported by ffprobe and the video is a mess.
I've tried adding code as per this answer e.g.
if (out_packet.pts != AV_NOPTS_VALUE) {
out_packet.pts = av_rescale_q(out_packet.pts, video_out_stream->codec->time_base, video_out_stream->time_base);
}
if (out_packet.pts != AV_NOPTS_VALUE) {
out_packet.dts = av_rescale_q(out_packet.dts, video_out_stream->codec->time_base, video_out_stream->time_base);
}
...but then I get errors such as this is the debug output:
[mp4 # 0x1038aba00] Delay between the first packet and last packet in the muxing queue is 10100000 > 10000000: forcing output
I'm setting the timebase correctly (I think)...
video_out_stream = avformat_new_stream(output_ctx, videoEncoder);
video_out_stream->id = 0;
video_out_stream->time_base.den = 30;
video_out_stream->time_base.num = 1;
video_out_codec = video_out_stream->codec;
avcodec_get_context_defaults3(video_out_codec, videoEncoder);
video_out_codec->codec_id = AV_CODEC_ID_H264;
video_out_codec->bit_rate = 2048;
video_out_codec->width = 854;
video_out_codec->height = 480;
video_out_codec->time_base.den = 30;
video_out_codec->time_base.num = 1;
video_out_codec->gop_size = 30;
video_out_codec->pix_fmt = AV_PIX_FMT_YUV420P;
Any thoughts about how I can calculate the correct output frame pts and encoded packet dts/pts? I'm clearly doing something wrong.
I am writing a wrapper code around latest ffmpeg library. I am supplying MP4 files from local system. My problem is that I am unable to get any decoded frames when I use av_decode_video2(). The return value comes out to be negative. I have used av_read_frame() which returns 0. I googled about the problem I am facing but no where could I find the correct explanation. Please give me insight here. Pasting the pseudo code here.
av_init_packet(avpkt);
picture=av_frame_alloc();
pFrameRGB=av_frame_alloc();
codec = avcodec_find_decoder(CODEC_ID_H264);
c= avcodec_alloc_context3(codec)
avcodec_open2(decoderLibraryData->c, decoderLibraryData->codec, NULL)
FormatContext = avformat_alloc_context();
char *pUrl ="./1.MP4";
iRet = avformat_open_input(atContext, pUrl, pFmt, NULL);
if(FormatContext == NULL)
{
printf("could not assign any memory !!!!!!!!! \n");
}
avformat_find_stream_info(FormatContext, NULL);
while(av_read_frame(FormatContext,avpkt) >= 0)
{
len = avcodec_decode_video2(c, picture, &got_picture,avpkt);
printf("CODEC MANAGER len %d Frame decompressed %d \n",len,got_picture);
if (len <= 0)
{
return ERROR;
}
}
}
if(lastHeight != 0 && lastWidth != 0)
{
if(lastWidth != c->width || lastHeight != c->height )
{
av_free(buffer);
buffer = NULL;
lastWidth = c->width;
lastHeight = c->height;
}
}
else
{
lastWidth = c->width;
lastHeight = c->height;
}
decodeFlag = 1;
if(!buffer)
{
int numBytes;
v_mutex_lock(globalCodecLock);
switch(inPixFormat)
{
case RGB:
// Determine required buffer size and allocate buffer
numBytes=avpicture_get_size(PIX_FMT_RGB24, c->width, c->height);
buffer=(uint8_t *)av_malloc(numBytes*sizeof(uint8_t));
avpicture_fill((AVPicture *)pFrameRGB,buffer,PIX_FMT_RGB24,c->width, c->height);
if(cntxt)
sws_freeContext(cntxt);
cntxt = sws_getContext(c->width, c->height, c->pix_fmt,
c->width, c->height, PIX_FMT_RGB24, SWS_BICUBIC, NULL, NULL, NULL);
break;
}
v_mutex_unlock(globalCodecLock);
if(cntxt == NULL)
{
printf("sws_getContext error\n");
return ERROR;
}
}
{
sws_scale(cntxt, picture->data, picture->linesize, 0, c->height, pFrameRGB->data, pFrameRGB->linesize);
if(rgbBuff)
{
if(c->width <= *width && c->height <= *height)
{
saveFrame(pFrameRGB, c->width, c->height, rgbBuff,inPixFormat);
*width = c->width;
*height = c->height;
rs = SUCCESS;
break;
}
else
{
rs = VA_LOWBUFFERSIZE;
}
}
else
{
rs = VA_LOWBUFFERSIZE;
}
}
if(width)
{
*width = c->width;
}
if(height)
{
*height = c->height;
}
if(rs == VA_LOWBUFFERSIZE)
{
break;
}
I am getting the return value of av_read_frame as 0 but av_decode_video2 returns value in negative. I am not able to get any clue here.
Make sure you have called
av_register_all();
or
avcodec_register_all();
at the beginning of your app.
Also it seems the problem is from calling avformat_find_stream_info. Test with the following code:
AVPacket avpkt;
av_init_packet(&avpkt);
AVFrame* picture = av_frame_alloc();
AVFrame* pFrameRGB = av_frame_alloc();
AVFormatContext* c2 = avformat_alloc_context();
char *pUrl = "C:/Sample Videos/20-06-34.MP4";
int video_stream_index = 0;
AVInputFormat* pFmt;
int iRet = avformat_open_input(&c2, pUrl, pFmt, NULL);
AVStream* stream = c2->streams[video_stream_index];
AVCodec* codec = avcodec_find_decoder(stream->codec->codec_id);
avcodec_open2(stream->codec, codec, NULL);
if (c2 == NULL)
{
printf("could not assign any memory !!!!!!!!! \n");
}
while (av_read_frame(c2, &avpkt) >= 0)
{
int got_picture;
int len = avcodec_decode_video2(stream->codec, picture, &got_picture, &avpkt);
printf("CODEC MANAGER len %d Frame decompressed %d \n", len, got_picture);
if (len <= 0)
{
return ERROR;
}
}
How can I cut video with FFmpeg C API? From 00:10:00 to 00:20:00 for example.
What functions I need to use?
I convert video using this function:
int convert(char *file) {
AVFrame *frame;
AVPacket inPacket, outPacket;
if(avio_open(&outFormatContext->pb, file, AVIO_FLAG_WRITE) < 0) {
fprintf(stderr, "convert(): cannot open out file\n");
return 0;
}
avformat_write_header(outFormatContext, NULL);
frame = avcodec_alloc_frame();
av_init_packet(&inPacket);
while(av_read_frame(inFormatContext, &inPacket) >= 0) {
if(inPacket.stream_index == inVideoStreamIndex) {
avcodec_decode_video2(inCodecContext, frame, &frameFinished, &inPacket);
if(frameFinished) {
av_init_packet(&outPacket);
avcodec_encode_video2(outCodecContext, &outPacket, frame, &outputed);
if(outputed) {
if (av_write_frame(outFormatContext, &outPacket) != 0) {
fprintf(stderr, "convert(): error while writing video frame\n");
return 0;
}
}
av_free_packet(&outPacket);
}
}
}
av_write_trailer(outFormatContext);
av_free_packet(&inPacket);
return 1;
}
As Wagner Patriota says, "if you just wanna cut the video, you don't need to reencode the video if you want". Here is the code based on ffmpeg remuxing.c example that you don't need to reencode the video.
#include <libavutil/timestamp.h>
#include <libavformat/avformat.h>
static void log_packet(const AVFormatContext *fmt_ctx, const AVPacket *pkt, const char *tag)
{
AVRational *time_base = &fmt_ctx->streams[pkt->stream_index]->time_base;
printf("%s: pts:%s pts_time:%s dts:%s dts_time:%s duration:%s duration_time:%s stream_index:%d\n",
tag,
av_ts2str(pkt->pts), av_ts2timestr(pkt->pts, time_base),
av_ts2str(pkt->dts), av_ts2timestr(pkt->dts, time_base),
av_ts2str(pkt->duration), av_ts2timestr(pkt->duration, time_base),
pkt->stream_index);
}
int cut_video(double from_seconds, double end_seconds, const char* in_filename, const char* out_filename) {
AVOutputFormat *ofmt = NULL;
AVFormatContext *ifmt_ctx = NULL, *ofmt_ctx = NULL;
AVPacket pkt;
int ret, i;
av_register_all();
if ((ret = avformat_open_input(&ifmt_ctx, in_filename, 0, 0)) < 0) {
fprintf(stderr, "Could not open input file '%s'", in_filename);
goto end;
}
if ((ret = avformat_find_stream_info(ifmt_ctx, 0)) < 0) {
fprintf(stderr, "Failed to retrieve input stream information");
goto end;
}
av_dump_format(ifmt_ctx, 0, in_filename, 0);
avformat_alloc_output_context2(&ofmt_ctx, NULL, NULL, out_filename);
if (!ofmt_ctx) {
fprintf(stderr, "Could not create output context\n");
ret = AVERROR_UNKNOWN;
goto end;
}
ofmt = ofmt_ctx->oformat;
for (i = 0; i < ifmt_ctx->nb_streams; i++) {
AVStream *in_stream = ifmt_ctx->streams[i];
AVStream *out_stream = avformat_new_stream(ofmt_ctx, in_stream->codec->codec);
if (!out_stream) {
fprintf(stderr, "Failed allocating output stream\n");
ret = AVERROR_UNKNOWN;
goto end;
}
ret = avcodec_copy_context(out_stream->codec, in_stream->codec);
if (ret < 0) {
fprintf(stderr, "Failed to copy context from input to output stream codec context\n");
goto end;
}
out_stream->codec->codec_tag = 0;
if (ofmt_ctx->oformat->flags & AVFMT_GLOBALHEADER)
out_stream->codec->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
}
av_dump_format(ofmt_ctx, 0, out_filename, 1);
if (!(ofmt->flags & AVFMT_NOFILE)) {
ret = avio_open(&ofmt_ctx->pb, out_filename, AVIO_FLAG_WRITE);
if (ret < 0) {
fprintf(stderr, "Could not open output file '%s'", out_filename);
goto end;
}
}
ret = avformat_write_header(ofmt_ctx, NULL);
if (ret < 0) {
fprintf(stderr, "Error occurred when opening output file\n");
goto end;
}
// int indexs[8] = {0};
// int64_t start_from = 8*AV_TIME_BASE;
ret = av_seek_frame(ifmt_ctx, -1, from_seconds*AV_TIME_BASE, AVSEEK_FLAG_ANY);
if (ret < 0) {
fprintf(stderr, "Error seek\n");
goto end;
}
int64_t *dts_start_from = malloc(sizeof(int64_t) * ifmt_ctx->nb_streams);
memset(dts_start_from, 0, sizeof(int64_t) * ifmt_ctx->nb_streams);
int64_t *pts_start_from = malloc(sizeof(int64_t) * ifmt_ctx->nb_streams);
memset(pts_start_from, 0, sizeof(int64_t) * ifmt_ctx->nb_streams);
while (1) {
AVStream *in_stream, *out_stream;
ret = av_read_frame(ifmt_ctx, &pkt);
if (ret < 0)
break;
in_stream = ifmt_ctx->streams[pkt.stream_index];
out_stream = ofmt_ctx->streams[pkt.stream_index];
log_packet(ifmt_ctx, &pkt, "in");
if (av_q2d(in_stream->time_base) * pkt.pts > end_seconds) {
av_free_packet(&pkt);
break;
}
if (dts_start_from[pkt.stream_index] == 0) {
dts_start_from[pkt.stream_index] = pkt.dts;
printf("dts_start_from: %s\n", av_ts2str(dts_start_from[pkt.stream_index]));
}
if (pts_start_from[pkt.stream_index] == 0) {
pts_start_from[pkt.stream_index] = pkt.pts;
printf("pts_start_from: %s\n", av_ts2str(pts_start_from[pkt.stream_index]));
}
/* copy packet */
pkt.pts = av_rescale_q_rnd(pkt.pts - pts_start_from[pkt.stream_index], in_stream->time_base, out_stream->time_base, AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
pkt.dts = av_rescale_q_rnd(pkt.dts - dts_start_from[pkt.stream_index], in_stream->time_base, out_stream->time_base, AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
if (pkt.pts < 0) {
pkt.pts = 0;
}
if (pkt.dts < 0) {
pkt.dts = 0;
}
pkt.duration = (int)av_rescale_q((int64_t)pkt.duration, in_stream->time_base, out_stream->time_base);
pkt.pos = -1;
log_packet(ofmt_ctx, &pkt, "out");
printf("\n");
ret = av_interleaved_write_frame(ofmt_ctx, &pkt);
if (ret < 0) {
fprintf(stderr, "Error muxing packet\n");
break;
}
av_free_packet(&pkt);
}
free(dts_start_from);
free(pts_start_from);
av_write_trailer(ofmt_ctx);
end:
avformat_close_input(&ifmt_ctx);
/* close output */
if (ofmt_ctx && !(ofmt->flags & AVFMT_NOFILE))
avio_closep(&ofmt_ctx->pb);
avformat_free_context(ofmt_ctx);
if (ret < 0 && ret != AVERROR_EOF) {
fprintf(stderr, "Error occurred: %s\n", av_err2str(ret));
return 1;
}
return 0;
}
If you just wanna cut the video, you don't need to reencode the video if you want.
So I am supposing you wanna cut and reencode for some reason. So, based on your code:
Observe you must to have access to the video AVStream* structure... I named it as inVideoStream.
int convert_and_cut(char *file, float starttime, float endtime) {
AVFrame *frame;
AVPacket inPacket, outPacket;
if(avio_open(&outFormatContext->pb, file, AVIO_FLAG_WRITE) < 0) {
fprintf(stderr, "convert(): cannot open out file\n");
return 0;
}
// seek to the start time you wish.
// BEGIN
AVRational default_timebase;
default_timebase.num = 1;
default_timebase.den = AV_TIME_BASE;
// suppose you have access to the "inVideoStream" of course
int64_t starttime_int64 = av_rescale_q((int64_t)( starttime * AV_TIME_BASE ), default_timebase, inVideoStream->time_base);
int64_t endtime_int64 = av_rescale_q((int64_t)( endtime * AV_TIME_BASE ), default_timebase, inVideoStream->time_base);
if(avformat_seek_file(inFormatContext, inVideoStreamIndex, INT64_MIN, starttime_int64, INT64_MAX, 0) < 0) {
// error... do something...
return 0; // usually 0 is used for success in C, but I am following your code.
}
avcodec_flush_buffers( inVideoStream->codec );
// END
avformat_write_header(outFormatContext, NULL);
frame = avcodec_alloc_frame();
av_init_packet(&inPacket);
// you used avformat_seek_file() to seek CLOSE to the point you want... in order to give precision to your seek,
// just go on reading the packets and checking the packets PTS (presentation timestamp)
while(av_read_frame(inFormatContext, &inPacket) >= 0) {
if(inPacket.stream_index == inVideoStreamIndex) {
avcodec_decode_video2(inCodecContext, frame, &frameFinished, &inPacket);
// this line guarantees you are getting what you really want.
if(frameFinished && frame->pkt_pts >= starttime_int64 && frame->pkt_pts <= endtime_int64) {
av_init_packet(&outPacket);
avcodec_encode_video2(outCodecContext, &outPacket, frame, &outputed);
if(outputed) {
if (av_write_frame(outFormatContext, &outPacket) != 0) {
fprintf(stderr, "convert(): error while writing video frame\n");
return 0;
}
}
av_free_packet(&outPacket);
}
// exit the loop if you got the frames you want.
if(frame->pkt_pts > endtime_int64) {
break;
}
}
}
av_write_trailer(outFormatContext);
av_free_packet(&inPacket);
return 1;
}
In addition, as ustin says, the code based on ffmpeg remuxing.c example.
add out_stream->time_base = in_stream->time_base; before out_stream->codec->codec_tag = 0;
add int64_t *start_from = NULL; after int stream_mapping_size = 0;
add start_from = av_mallocz_array(stream_mapping_size, sizeof(*start_from)); after stream_mapping = av_mallocz_array(stream_mapping_size, sizeof(*stream_mapping));
add start_from[stream_mapping[i]] = INT16_MIN; after stream_mapping[i] = stream_index++;
add
if (start_from[pkt.stream_index] == INT16_MIN) {
int64_t dts = pkt.dts;
int64_t pts = pkt.pts;
int64_t min_ts = dts > pts ? pts : dts;
start_from[pkt.stream_index] = min_ts < 0 ? 0 : min_ts;
}
After
if (pkt.stream_index >= stream_mapping_size ||
stream_mapping[pkt.stream_index] < 0) {
av_packet_unref(&pkt);
continue;
}
change justin code, about pts/dst set value to this:
pkt.pts = av_rescale_q_rnd(pkt.pts - start_from[pkt.stream_index], in_stream->time_base, out_stream->time_base, AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
pkt.dts = av_rescale_q_rnd(pkt.dts - start_from[pkt.stream_index], in_stream->time_base, out_stream->time_base, AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX);
add free(start_from); before end:
Tips: start/end timestamp like justin says