Related
One of the things I'm trying to achieve is parallel encoding via FFmpeg's c API. This looks to work out of the box quite nicely; however, I've changed the goal posts slightly:
In an existing application, I already have a thread pool at hand. Instead of using another thread pool via FFmpeg, I would like reuse the existing thread pool in my application. Having studied the latest FFmpeg trunk docs, it very much looks possible.
Using some FFmpeg sample code, I've created a sample application to demonstrate what I'm trying to achieve (see below). The sample app generates a video-only mpeg2 ts using the mp2v codec.
The problem I'm experiencing is that the custom 'thread_execute' or 'thread_execute2' are never invoked. This is despite the fact that the codec appears to indicate that threading is supported. Please be aware that I have not yet plumbed in the thread pool just yet. My first goal is for it to call the custom function pointer.
I've tried to get assistance on the FFmpeg mailing lists but to no avail.
#include <iostream>
#include <thread>
#include <stdlib.h>
#include <string.h>
#include <math.h>
#include <cstring>
#include <future>
extern "C"
{
#include <libavutil/avassert.h>
#include <libavutil/channel_layout.h>
#include <libavutil/opt.h>
#include <libavutil/timestamp.h>
#include <libavformat/avformat.h>
//#include <libswscale/swscale.h>
#include <libswresample/swresample.h>
}
#define STREAM_DURATION 1000.0
#define STREAM_FRAME_RATE 25 /* 25 images/s */
#define STREAM_PIX_FMT AV_PIX_FMT_YUV420P /* default pix_fmt */
#define SCALE_FLAGS SWS_BICUBIC
// a wrapper around a single output AVStream
typedef struct OutputStream {
AVStream *st;
AVCodecContext *enc;
/* pts of the next frame that will be generated */
int64_t next_pts;
int samples_count;
AVFrame *frame;
AVFrame *tmp_frame;
float t, tincr, tincr2;
struct SwsContext *sws_ctx;
struct SwrContext *swr_ctx;
} OutputStream;
/////////////////////////////////////////////////////////////////////////////
// The ffmpeg variation raises compiler warnings.
char *cb_av_ts2str(char *buf, int64_t ts)
{
std::memset(buf,0,AV_TS_MAX_STRING_SIZE);
return av_ts_make_string(buf,ts);
}
/////////////////////////////////////////////////////////////////////////////
// The ffmpeg variation raises compiler warnings.
char *cb_av_ts2timestr(char *buf, int64_t ts, AVRational *tb)
{
std::memset(buf,0,sizeof(AV_TS_MAX_STRING_SIZE));
return av_ts_make_time_string(buf,ts,tb);
}
/////////////////////////////////////////////////////////////////////////////
// The ffmpeg variation raises compiler warnings.
char *cb_av_err2str(char *errbuf, size_t errbuf_size, int errnum)
{
std::memset(errbuf,0,errbuf_size);
return av_make_error_string(errbuf,errbuf_size,errnum);
}
int thread_execute(AVCodecContext* s, int (*func)(AVCodecContext *c2, void *arg2), void* arg, int* ret, int count, int size)
{
// Do it all serially for now
std::cout << "thread_execute" << std::endl;
for (int k = 0; k < count; ++k)
{
ret[k] = func(s, arg);
}
return 0;
}
int thread_execute2(AVCodecContext* s, int (*func)(AVCodecContext* c2, void* arg2, int, int), void* arg, int* ret, int count)
{
// Do it all serially for now
std::cout << "thread_execute2" << std::endl;
for (int k = 0; k < count; ++k)
{
ret[k] = func(s, arg, k, count);
}
return 0;
}
static void log_packet(const AVFormatContext *fmt_ctx, const AVPacket *pkt)
{
char s[AV_TS_MAX_STRING_SIZE];
AVRational *time_base = &fmt_ctx->streams[pkt->stream_index]->time_base;
printf("pts:%s pts_time:%s dts:%s dts_time:%s duration:%s duration_time:%s stream_index:%d\n",
cb_av_ts2str(s,pkt->pts), cb_av_ts2timestr(s,pkt->pts, time_base),
cb_av_ts2str(s,pkt->dts), cb_av_ts2timestr(s,pkt->dts, time_base),
cb_av_ts2str(s,pkt->duration), cb_av_ts2timestr(s,pkt->duration, time_base),
pkt->stream_index);
}
static int write_frame(AVFormatContext *fmt_ctx, const AVRational *time_base, AVStream *st, AVPacket *pkt)
{
/* rescale output packet timestamp values from codec to stream timebase */
av_packet_rescale_ts(pkt, *time_base, st->time_base);
pkt->stream_index = st->index;
/* Write the compressed frame to the media file. */
log_packet(fmt_ctx, pkt);
return av_interleaved_write_frame(fmt_ctx, pkt);
}
/* Add an output stream. */
static void add_stream(OutputStream *ost, AVFormatContext *oc,
AVCodec **codec,
enum AVCodecID codec_id)
{
AVCodecContext *c;
int i;
/* find the encoder */
*codec = avcodec_find_encoder(codec_id);
if (!(*codec)) {
fprintf(stderr, "Could not find encoder for '%s'\n",
avcodec_get_name(codec_id));
exit(1);
}
ost->st = avformat_new_stream(oc, NULL);
if (!ost->st) {
fprintf(stderr, "Could not allocate stream\n");
exit(1);
}
ost->st->id = oc->nb_streams-1;
c = avcodec_alloc_context3(*codec);
if (!c) {
fprintf(stderr, "Could not alloc an encoding context\n");
exit(1);
}
ost->enc = c;
switch ((*codec)->type)
{
case AVMEDIA_TYPE_AUDIO:
c->sample_fmt = (*codec)->sample_fmts ?
(*codec)->sample_fmts[0] : AV_SAMPLE_FMT_FLTP;
c->bit_rate = 64000;
c->sample_rate = 44100;
if ((*codec)->supported_samplerates) {
c->sample_rate = (*codec)->supported_samplerates[0];
for (i = 0; (*codec)->supported_samplerates[i]; i++) {
if ((*codec)->supported_samplerates[i] == 44100)
c->sample_rate = 44100;
}
}
c->channels = av_get_channel_layout_nb_channels(c->channel_layout);
c->channel_layout = AV_CH_LAYOUT_STEREO;
if ((*codec)->channel_layouts) {
c->channel_layout = (*codec)->channel_layouts[0];
for (i = 0; (*codec)->channel_layouts[i]; i++) {
if ((*codec)->channel_layouts[i] == AV_CH_LAYOUT_STEREO)
c->channel_layout = AV_CH_LAYOUT_STEREO;
}
}
c->channels = av_get_channel_layout_nb_channels(c->channel_layout);
ost->st->time_base = (AVRational){ 1, c->sample_rate };
break;
case AVMEDIA_TYPE_VIDEO:
c->codec_id = codec_id;
c->bit_rate = 400000;
/* Resolution must be a multiple of two. */
c->width = 352;
c->height = 288;
/* timebase: This is the fundamental unit of time (in seconds) in terms
* of which frame timestamps are represented. For fixed-fps content,
* timebase should be 1/framerate and timestamp increments should be
* identical to 1. */
ost->st->time_base = (AVRational){ 1, STREAM_FRAME_RATE };
c->time_base = ost->st->time_base;
c->gop_size = 12; /* emit one intra frame every twelve frames at most */
c->pix_fmt = STREAM_PIX_FMT;
if (c->codec_id == AV_CODEC_ID_MPEG2VIDEO) {
/* just for testing, we also add B-frames */
c->max_b_frames = 2;
}
if (c->codec_id == AV_CODEC_ID_MPEG1VIDEO) {
/* Needed to avoid using macroblocks in which some coeffs overflow.
* This does not happen with normal video, it just happens here as
* the motion of the chroma plane does not match the luma plane. */
c->mb_decision = 2;
}
break;
default:
break;
}
if (c->codec->capabilities & AV_CODEC_CAP_FRAME_THREADS ||
c->codec->capabilities & AV_CODEC_CAP_SLICE_THREADS)
{
if (c->codec->capabilities & AV_CODEC_CAP_FRAME_THREADS)
{
c->thread_type = FF_THREAD_FRAME;
}
if (c->codec->capabilities & AV_CODEC_CAP_SLICE_THREADS)
{
c->thread_type = FF_THREAD_SLICE;
}
c->execute = &thread_execute;
c->execute2 = &thread_execute2;
c->thread_count = 4;
// NOTE: Testing opaque.
c->opaque = (void*)0xff;
}
/* Some formats want stream headers to be separate. */
if (oc->oformat->flags & AVFMT_GLOBALHEADER)
c->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
}
/**************************************************************/
/* video output */
static AVFrame *alloc_picture(enum AVPixelFormat pix_fmt, int width, int height)
{
AVFrame *picture;
int ret;
picture = av_frame_alloc();
if (!picture)
return NULL;
picture->format = pix_fmt;
picture->width = width;
picture->height = height;
/* allocate the buffers for the frame data */
ret = av_frame_get_buffer(picture, 32);
if (ret < 0) {
fprintf(stderr, "Could not allocate frame data.\n");
exit(1);
}
return picture;
}
static void open_video(AVFormatContext *oc, AVCodec *codec, OutputStream *ost, AVDictionary *opt_arg)
{
int ret;
AVCodecContext *c = ost->enc;
//AVDictionary *opt = NULL;
//av_dict_copy(&opt, opt_arg, 0);
/* open the codec */
ret = avcodec_open2(c, codec, NULL);
//av_dict_free(&opt);
if (ret < 0) {
char s[AV_ERROR_MAX_STRING_SIZE];
fprintf(stderr, "Could not open video codec: %s\n", cb_av_err2str(s,AV_ERROR_MAX_STRING_SIZE,ret));
exit(1);
}
/* allocate and init a re-usable frame */
ost->frame = alloc_picture(c->pix_fmt, c->width, c->height);
if (!ost->frame) {
fprintf(stderr, "Could not allocate video frame\n");
exit(1);
}
/* If the output format is not YUV420P, then a temporary YUV420P
* picture is needed too. It is then converted to the required
* output format. */
ost->tmp_frame = NULL;
if (c->pix_fmt != AV_PIX_FMT_YUV420P) {
ost->tmp_frame = alloc_picture(AV_PIX_FMT_YUV420P, c->width, c->height);
if (!ost->tmp_frame) {
fprintf(stderr, "Could not allocate temporary picture\n");
exit(1);
}
}
/* copy the stream parameters to the muxer */
ret = avcodec_parameters_from_context(ost->st->codecpar, c);
if (ret < 0) {
fprintf(stderr, "Could not copy the stream parameters\n");
exit(1);
}
}
/* Prepare a dummy image. */
static void fill_yuv_image(AVFrame *pict, int frame_index,
int width, int height)
{
int x, y, i;
i = frame_index;
/* Y */
for (y = 0; y < height; y++)
for (x = 0; x < width; x++)
pict->data[0][y * pict->linesize[0] + x] = x + y + i * 3;
/* Cb and Cr */
for (y = 0; y < height / 2; y++) {
for (x = 0; x < width / 2; x++) {
pict->data[1][y * pict->linesize[1] + x] = 128 + y + i * 2;
pict->data[2][y * pict->linesize[2] + x] = 64 + x + i * 5;
}
}
}
static AVFrame *get_video_frame(OutputStream *ost)
{
AVCodecContext *c = ost->enc;
/* check if we want to generate more frames */
if (av_compare_ts(ost->next_pts, c->time_base,
STREAM_DURATION, (AVRational){ 1, 1 }) >= 0)
return NULL;
/* when we pass a frame to the encoder, it may keep a reference to it
* internally; make sure we do not overwrite it here */
if (av_frame_make_writable(ost->frame) < 0)
exit(1);
if (c->pix_fmt != AV_PIX_FMT_YUV420P) {
/* as we only generate a YUV420P picture, we must convert it
* to the codec pixel format if needed */
/*if (!ost->sws_ctx) {
ost->sws_ctx = sws_getContext(c->width, c->height,
AV_PIX_FMT_YUV420P,
c->width, c->height,
c->pix_fmt,
SCALE_FLAGS, NULL, NULL, NULL);
if (!ost->sws_ctx) {
fprintf(stderr,
"Could not initialize the conversion context\n");
exit(1);
}
}
fill_yuv_image(ost->tmp_frame, ost->next_pts, c->width, c->height);
sws_scale(ost->sws_ctx,
(const uint8_t * const *)ost->tmp_frame->data, ost->tmp_frame->linesize,
0, c->height, ost->frame->data, ost->frame->linesize);*/
} else {
fill_yuv_image(ost->frame, ost->next_pts, c->width, c->height);
}
ost->frame->pts = ost->next_pts++;
return ost->frame;
}
/*
* encode one video frame and send it to the muxer
* return 1 when encoding is finished, 0 otherwise
*/
static int write_video_frame(AVFormatContext *oc, OutputStream *ost)
{
int ret;
AVCodecContext *c;
AVFrame *frame;
int got_packet = 0;
AVPacket pkt = { 0 };
c = ost->enc;
frame = get_video_frame(ost);
if (frame)
{
ret = avcodec_send_frame(ost->enc, frame);
if (ret < 0)
{
char s[AV_ERROR_MAX_STRING_SIZE];
fprintf(stderr, "Error encoding video frame: %s\n", cb_av_err2str(s, AV_ERROR_MAX_STRING_SIZE, ret));
exit(1);
}
}
av_init_packet(&pkt);
ret = avcodec_receive_packet(ost->enc,&pkt);
if (ret < 0)
{
if (ret == AVERROR(EAGAIN)) { ret = 0; }
else
{
char s[AV_ERROR_MAX_STRING_SIZE];
fprintf(stderr, "Error receiving packet: %s\n", cb_av_err2str(s,AV_ERROR_MAX_STRING_SIZE,ret));
exit(1);
}
}
else
{
got_packet = 1;
ret = write_frame(oc, &c->time_base, ost->st, &pkt);
}
if (ret < 0) {
char s[AV_ERROR_MAX_STRING_SIZE];
fprintf(stderr, "Error while writing video frame: %s\n", cb_av_err2str(s,AV_ERROR_MAX_STRING_SIZE,ret));
exit(1);
}
return (frame || got_packet) ? 0 : 1;
}
static void close_stream(AVFormatContext *oc, OutputStream *ost)
{
avcodec_free_context(&ost->enc);
av_frame_free(&ost->frame);
av_frame_free(&ost->tmp_frame);
//sws_freeContext(ost->sws_ctx);
//swr_free(&ost->swr_ctx);
}
/**************************************************************/
/* media file output */
int main(int argc, char **argv)
{
OutputStream video_st = { 0 }, audio_st = { 0 };
const char *filename;
AVOutputFormat *fmt;
AVFormatContext *oc;
AVCodec /**audio_codec,*/ *video_codec;
int ret;
int have_video = 0, have_audio = 0;
int encode_video = 0, encode_audio = 0;
AVDictionary *opt = NULL;
int i;
/* Initialize libavcodec, and register all codecs and formats. */
av_register_all();
avformat_network_init();
if (argc < 2) {
printf("usage: %s output_file\n"
"API example program to output a media file with libavformat.\n"
"This program generates a synthetic audio and video stream, encodes and\n"
"muxes them into a file named output_file.\n"
"The output format is automatically guessed according to the file extension.\n"
"Raw images can also be output by using '%%d' in the filename.\n"
"\n", argv[0]);
return 1;
}
filename = argv[1];
for (i = 2; i+1 < argc; i+=2) {
if (!strcmp(argv[i], "-flags") || !strcmp(argv[i], "-fflags"))
av_dict_set(&opt, argv[i]+1, argv[i+1], 0);
}
const char *pfilename = filename;
/* allocate the output media context */
avformat_alloc_output_context2(&oc, NULL, "mpegts", pfilename);
if (!oc) {
printf("Could not deduce output format from file extension: using MPEG.\n");
avformat_alloc_output_context2(&oc, NULL, "mpeg", pfilename);
}
if (!oc)
return 1;
fmt = oc->oformat;
/* Add the audio and video streams using the default format codecs
* and initialize the codecs. */
if (fmt->video_codec != AV_CODEC_ID_NONE) {
add_stream(&video_st, oc, &video_codec, fmt->video_codec);
have_video = 1;
encode_video = 1;
}
/*if (fmt->audio_codec != AV_CODEC_ID_NONE) {
add_stream(&audio_st, oc, &audio_codec, fmt->audio_codec);
have_audio = 1;
encode_audio = 1;
}*/
/* Now that all the parameters are set, we can open the audio and
* video codecs and allocate the necessary encode buffers. */
if (have_video)
open_video(oc, video_codec, &video_st, opt);
//if (have_audio)
// open_audio(oc, audio_codec, &audio_st, opt);
av_dump_format(oc, 0, pfilename, 1);
/* open the output file, if needed */
if (!(fmt->flags & AVFMT_NOFILE)) {
ret = avio_open(&oc->pb, pfilename, AVIO_FLAG_WRITE);
if (ret < 0) {
char s[AV_ERROR_MAX_STRING_SIZE];
fprintf(stderr, "Could not open '%s': %s\n", pfilename,
cb_av_err2str(s,AV_ERROR_MAX_STRING_SIZE,ret));
return 1;
}
}
/* Write the stream header, if any. */
ret = avformat_write_header(oc, &opt);
if (ret < 0) {
char s[AV_ERROR_MAX_STRING_SIZE];
fprintf(stderr, "Error occurred when opening output file: %s\n",
cb_av_err2str(s,AV_ERROR_MAX_STRING_SIZE,ret));
return 1;
}
while (encode_video || encode_audio) {
/* select the stream to encode */
if (encode_video &&
(!encode_audio || av_compare_ts(video_st.next_pts, video_st.enc->time_base,
audio_st.next_pts, audio_st.enc->time_base) <= 0)) {
encode_video = !write_video_frame(oc, &video_st);
} else {
//encode_audio = !write_audio_frame(oc, &audio_st);
}
//std::this_thread::sleep_for(std::chrono::milliseconds(35));
}
/* Write the trailer, if any. The trailer must be written before you
* close the CodecContexts open when you wrote the header; otherwise
* av_write_trailer() may try to use memory that was freed on
* av_codec_close(). */
av_write_trailer(oc);
/* Close each codec. */
if (have_video)
close_stream(oc, &video_st);
if (have_audio)
close_stream(oc, &audio_st);
if (!(fmt->flags & AVFMT_NOFILE))
/* Close the output file. */
avio_closep(&oc->pb);
/* free the stream */
avformat_free_context(oc);
return 0;
}
//
Environment:
Ubuntu Zesty (17.04)
FFmpeg version 3.2.4 (via package manager)
gcc 6.3 (C++)
You have to do following:
call avcodec_alloc_context3(...). This call will set default execute and execute2 functions in new context
set c->thread_count = number_of_threads_in_your_thread_pool()
call avcodec_open2(...).
set c->execute and c->execute2 to point to your functions
call ff_thread_free(c). This function isnt exposed in libavcodec headers but you can add following line:
extern "C" void ff_thread_free(AVCodecContext *s);
Drawback is that libavcodec will create internal thread pool after avcodec_open2(...) call, and that pool will be deleted in ff_thread_free() call.
Internal thread pool is very efficient, but its not good if you plan to do parallel encoding of multiple video feeds. In that case libavcodec will create separate thread pool for each encoding video feed.
I have a question regarding buffer read with ffmpeg.
Idea is as it follows: an outside module (can not change it) is providing me video stream in chunks of data and it is giving me input data and it's size in bytes ("framfunction" function input parameters). I have to copy input data to a buffer and read it with ffmpeg (Zeranoe) and extract video frames. Each time I receive new data, my function "framfunction" will be called. All unprocessed data from a first run will be moved at the beginning of the buffer followed by a new data on the second run and so on. It is essentially based on source and Dranger tutorials. My current attempt is like this and just look at comments (I've left only ones regarding current buffer function) in code to get a picture what I want to do(I know it is messy and it works - sort of; skipping some frames. Any suggestions around ffmpeg code and buffer design are welcome):
#include <iostream>
#include <string>
extern "C"
{
#include <libavformat/avformat.h>
#include <libavcodec/avcodec.h>
#include <libswscale/swscale.h>
#include <libavformat/avio.h>
#include <libavutil/file.h>
}
struct buffer_data {
uint8_t *ptr;
size_t size;
};
static int read_packet(void *opaque, uint8_t *buf, int buf_size)
{
struct buffer_data *bd = (struct buffer_data *)opaque;
buf_size = FFMIN(buf_size, bd->size);
memcpy(buf, bd->ptr, buf_size);
bd->ptr += buf_size;
bd->size -= buf_size;
return buf_size;
}
class videoclass
{
private:
uint8_t* inputdatabuffer;
size_t offset;
public:
videoclass();
~videoclass();
int framfunction(uint8_t* inputbytes, int inputbytessize);
};
videoclass::videoclass()
: inputdatabuffer(nullptr)
, offset(0)
{
inputdatabuffer = new uint8_t[8388608]; //buffer where the input data will be stored
}
videoclass::~videoclass()
{
delete[] inputdatabuffer;
}
int videoclass::framfunction(uint8_t* inputbytes, int inputbytessize)
{
int i, videoStream, numBytes, frameFinished;
AVFormatContext *pFormatCtx = NULL;
AVCodecContext *pCodecCtx = NULL;
AVIOContext *avio_ctx = NULL;
AVCodec *pCodec = NULL;
AVFrame *pFrame = NULL;
AVFrame *pFrameRGB = NULL;
AVPacket packet;
uint8_t *buffer = NULL;
uint8_t *avio_ctx_buffer = NULL;
size_t avio_ctx_buffer_size = 4096;
size_t bytes_processed = 0;
struct buffer_data bd = { 0 };
//if (av_file_map("sample.ts", &inputbytes, &inputbytessize, 0, NULL) < 0)//
// return -1;
memcpy(inputdatabuffer + offset, inputbytes, inputbytessize);//copy new data to buffer inputdatabuffer with offset calculated at the end of previous function run. In other words - cope new data after unprocessed data from a previous call
offset += inputbytessize; //total number of bytes in buffer. Size of an unprocessed data from the last run + size of new data (inputbytessize)
bd.ptr = inputdatabuffer;
bd.size = offset;
if (!(pFormatCtx = avformat_alloc_context()))
return -1;
avio_ctx_buffer = (uint8_t *)av_malloc(avio_ctx_buffer_size);
avio_ctx = avio_alloc_context(avio_ctx_buffer, avio_ctx_buffer_size,0, &bd, &read_packet, NULL, NULL);
pFormatCtx->pb = avio_ctx;
av_register_all();
avcodec_register_all();
pFrame = av_frame_alloc();
pFrameRGB = av_frame_alloc();
if (avformat_open_input(&pFormatCtx, NULL, NULL, NULL) != 0)
return -2;
if (avformat_find_stream_info(pFormatCtx, NULL) < 0)
return -3;
videoStream = -1;
for (i = 0; i < pFormatCtx->nb_streams; i++)
if (pFormatCtx->streams[i]->codec->codec_type == AVMEDIA_TYPE_VIDEO) {
videoStream = i;
break;
}
if (videoStream == -1)
return -4;
pCodecCtx = pFormatCtx->streams[videoStream]->codec;
pCodec = avcodec_find_decoder(pCodecCtx->codec_id);
if (pCodec == NULL){
std::cout << "Unsupported codec" << std::endl;
return -5;
}
if (avcodec_open2(pCodecCtx, pCodec, NULL) < 0)
return -6;
numBytes = avpicture_get_size(PIX_FMT_BGR24, pCodecCtx->width, pCodecCtx->height);
buffer = (uint8_t *)av_malloc(numBytes*sizeof(uint8_t));
avpicture_fill((AVPicture *)pFrameRGB, buffer, PIX_FMT_BGR24, pCodecCtx->width, pCodecCtx->height);
while (av_read_frame(pFormatCtx, &packet) >= 0){
if (packet.stream_index == videoStream){
avcodec_decode_video2(pCodecCtx, pFrame, &frameFinished, &packet);
if (frameFinished){
std::cout << "Yaay, frame found" << std::endl;
}
}
av_free_packet(&packet);
bytes_processed = (size_t)pFormatCtx->pb->pos; //data which is processed so far (x bytes out of inputbytessize)??????????????????????????????
}
offset -= bytes_processed; //size of unprocessed data
av_free(buffer);
av_free(pFrameRGB);
av_free(pFrame);
avcodec_close(pCodecCtx);
av_freep(&avio_ctx->buffer);
av_freep(&avio_ctx);
avformat_close_input(&pFormatCtx);
memmove(inputdatabuffer, inputdatabuffer + bytes_processed, offset);//move unprocessed data to begining of the main buffer
return 0;
}
Call of my function would be something like this
WHILE(VIDEO_INPUT)
{
READ VIDEO DATA FROM INPUT BUFFER
STORE DATA FROM INPUT BUFFER AND SIZE OP THAT DATA TO VARIABLES NEW_DATA AND NEW_DATA_SIZE
CALL FUNCTION FRAMMUNCTION AND PASS NEW_DATA AND NEW_DATA_FUNCTION
DO OTHER THINGS
}
What I would like to know is what is exact size of unprocessed data. Comments in code are showing my attempt but I think it is not good enough so I need some help with that issue.
EDIT: magic question is how to get correct "bytes_processed" size. I've also made an pdf with explanation how my buffer should work pdf file
Thanks
I have a C++ DLL for use from C#. I have a function which takes a string passed to it, and I have those set on the C++ function parameters as const char * like so:
int __stdcall extract_all_frames(const char* szDestination, float scaleFactor)
The main body of this function is copied directly from a working FFmpeg example function so I'm almost certain the problem isn't there. I feel like the problem is in this modification I made to it:
//Open file
char szFilename[32];
sprintf_s(szFilename, sizeof(szFilename), "frame%d.ppm\0", iFrame);
// JOIN szFILENAME AND szDESTINATION
std::string buffer(szDestination, sizeof(szDestination));
buffer.append("\\");
buffer.append(szDestination);
Which is supposed to be a concatenated path and directory. I then pass buffer.c_str() into fopen_s(), which takes const char * not std::string. Whenever calling this function from C#, I get the following exception:
A first chance exception of type 'System.AccessViolationException' occurred in XRF FFmpeg Invoke Test.exe
Additional information: Attempted to read or write protected memory. This is often an indication that other memory is corrupt.
This is the complete code:
#include "stdafx.h"
#pragma comment (lib, "avcodec.lib")
#pragma comment (lib, "avformat.lib")
#pragma comment (lib, "avutil.lib")
#pragma comment (lib, "swscale.lib")
extern "C"
{
#include <libavcodec\avcodec.h>
#include <libavformat\avformat.h>
#include <libavutil\avutil.h>
#include <libswscale\swscale.h>
}
#include <string>
#include "Xrf.FFmpeg.hpp"
void save_frame(AVFrame* pFrame, int iFrame, const char* szDestination)
{
//Open file
char szFilename[32];
sprintf_s(szFilename, sizeof(szFilename), "frame%d.ppm\0", iFrame);
// JOIN szFILENAME AND szDESTINATION
std::string buffer(szDestination, sizeof(szDestination));
buffer.append("\\");
buffer.append(szDestination);
FILE* pFile;
errno_t openError = fopen_s(&pFile, buffer.c_str(), "wb");
if (pFile == NULL)
{
return;
}
//Write header
int width = pFrame->width;
int height = pFrame->height;
fprintf(pFile, "P6\n%d %d\n255\n", width, height);
//Write pixel data
for (int y = 0; y < height; y++)
{
fwrite(pFrame->data[0] + y * pFrame->linesize[0], 1, width * 3, pFile);
}
// Close file
fclose(pFile);
}
int __stdcall extract_all_frames(const char* szPath, const char* szDestination, float scaleFactor)
{
// Check if scaleFactor is valid
if ((scaleFactor != 0.f) &&
(scaleFactor > 3.f))
{
fprintf(stderr, "Xrf: Scale factor '%f' out of bounds!\nMust be greater than 0, and less then or equal to 3.0.\n", scaleFactor);
return -1;
}
// Register all formats and codecs
av_register_all();
AVFormatContext* pFormatCtx;
if (avformat_open_input(&pFormatCtx, szPath, nullptr, nullptr) != 0)
{
fprintf(stderr, "libavformat: Couldn't open file '%s'!\n", szPath);
return -1;
}
// Retrieve stream information
if (avformat_find_stream_info(pFormatCtx, nullptr) < 0)
{
fprintf(stderr, "libavformat: Unable to find stream information!\n");
return -1;
}
// Dump information about file onto standard error
av_dump_format(pFormatCtx, 0, szPath, 0);
// Find the first video stream
size_t i;
int videoStream = -1;
for (i = 0; i < pFormatCtx->nb_streams; i++)
{
if (pFormatCtx->streams[i]->codec->codec_type == AVMEDIA_TYPE_VIDEO)
{
videoStream = i;
break;
}
}
if (videoStream == -1)
{
fprintf(stderr, "libavformat: No video stream found!\n");
return -1;
}
// Get a pointer to the codec context for the video stream
AVCodecContext* pCodecCtx = pFormatCtx->streams[videoStream]->codec;
// Scale the frame
int scaleHeight = static_cast<int>(floor(pCodecCtx->height * scaleFactor));
int scaleWidth = static_cast<int>(floor(pCodecCtx->width * scaleFactor));
//Check if frame sizes are valid (not 0, because that's dumb)
if (scaleHeight == 0 || scaleWidth == 0)
{
fprintf(stderr, "Xrf: Scale factor caused a zero value in either width or height!\n");
return -1;
}
// Find the decoder for the video stream
AVCodec* pCodec = avcodec_find_decoder(pCodecCtx->codec_id);
if (pCodec == NULL)
{
fprintf(stderr, "libavcodec: Unsupported codec!\n");
return -1; // Codec not found
}
// Open codec
AVDictionary* optionsDict = nullptr;
if (avcodec_open2(pCodecCtx, pCodec, &optionsDict) < 0)
{
fprintf(stderr, "libavcodec: Couldn't open codec '%s'!\n", pCodec->long_name);
return -1;
}
// Allocate video frame
AVFrame* pFrame = av_frame_alloc();
// Allocate an AVFrame structure
AVFrame* pFrameRGB = av_frame_alloc();
if (pFrameRGB == NULL)
{
fprintf(stderr, "libavformat: Unable to allocate a YUV->RGB resampling AVFrame!\n");
return -1;
}
// Determine required buffer size and allocate buffer
int numBytes = avpicture_get_size(PIX_FMT_RGB24, scaleWidth, scaleHeight);
uint8_t* buffer = static_cast <uint8_t *> (av_malloc(numBytes * sizeof(uint8_t)));
struct SwsContext* sws_ctx = sws_getContext(pCodecCtx->width,
pCodecCtx->height,
pCodecCtx->pix_fmt,
scaleWidth,
scaleHeight,
PIX_FMT_RGB24,
SWS_BILINEAR,
nullptr, nullptr, nullptr);
// Assign appropriate parts of buffer to image planes in pFrameRGB
// Note that pFrameRGB is an AVFrame, but AVFrame is a superset
// of AVPicture
avpicture_fill(reinterpret_cast <AVPicture *> (pFrameRGB),
buffer,
PIX_FMT_RGB24,
scaleWidth,
scaleHeight);
// Read frames and save first five frames to disk
AVPacket packet;
int frameFinished;
while (av_read_frame(pFormatCtx, &packet) >= 0)
{
// Is this a packet from the video stream?
if (packet.stream_index == videoStream)
{
// Decode video frame
avcodec_decode_video2(pCodecCtx, pFrame, &frameFinished, &packet);
// Did we get a video frame?
if (frameFinished)
{
// Convert the image from its native format to RGB
sws_scale(sws_ctx,
static_cast <uint8_t const * const *> (pFrame->data),
pFrame->linesize,
0,
pCodecCtx->height,
pFrameRGB->data,
pFrameRGB->linesize);
// Save the frame to disk
if (++i <= 5)
{
save_frame(pFrameRGB, i, szDestination);
}
}
}
// Free the packet that was allocated by av_read_frame
av_free_packet(&packet);
}
av_free(buffer); // Free the RGB image
av_free(pFrameRGB);
av_free(pFrame); // Free the YUV frame
avcodec_close(pCodecCtx); // Close the codec
avformat_close_input(&pFormatCtx); // Close the video file
return 0;
}
I don't know if the error is in my modification (most likely, I'm extremely new to C++), or the other code, as the exception only throws on the invocation line in C#, not the actual C++ line causing the problem.
This is wrong:
std::string buffer(szDestination, sizeof(szDestination));
szDestination is a pointer, thus sizeof(szDestination) will return the pointer size, in bytes, not the number of characters.
If szDestination is a null terminated string, use strlen or similar function to determine the number of characters. If it isn't null terminated, then you need to pass the number of bytes to copy as a parameter.
The better thing to do is when your DLL function is called:
int __stdcall extract_all_frames(const char* szPath, const char* szDestination, float scaleFactor)
take those pointers and immediately assign them to std::string. Then drop all usage of char* or const char* from there. There is no need for your helper functions to deal with "dumb" character pointers.
Example:
int __stdcall extract_all_frames(const char* szPath, const char* szDestination, float scaleFactor)
{
std::string sPath = szPath;
std::string sDestination = sDestination;
// From here, use sPath and sDestination
//...
}
// redefinition of save_frame
//...
void save_frame(AVFrame* pFrame, int iFrame, const std::string& szDestination)
{
//Open file
std::string buffer = "frame" + to_string(iFrame) + ".ppm\0";
buffer.append("\\");
buffer.append(szDestination);
//...
}
I'm making a program that records and encodes this recording to aac.I made a function for linux and libfaac that does this job. Now I need to make this program for windows.I know i need to use libvo-aacenc and i don't know what to change in my code.Could you tell me what should I do? Here's my code.
static void encodeAac( const char *infilename,const char *filename)
{
AVCodec *codec;
AVCodecContext *c = NULL;
int frame_size, i, j, out_size, outbuf_size;
FILE *f,*fin;
SAMPLE *samples;
float t, tincr;
uint8_t *outbuf;
avcodec_register_all(); //Load all codecs
av_register_all();
codec = avcodec_find_encoder(AV_CODEC_ID_AAC); //Search for AAC codec
if (!codec) {
error("Codec not found");
}
c = avcodec_alloc_context();
c->bit_rate = 64000;
c->sample_fmt = AV_SAMPLE_FMT_S16;
c->sample_rate = SAMPLE_RATE;
c->channels = NUM_CHANNELS;
c->time_base.num= 1;
c->time_base.den= SAMPLE_RATE;
c->profile= FF_PROFILE_AAC_MAIN;
if (avcodec_open(c, codec) < 0) {
error(add("","",avcodec_open(c, codec)).c_str());
exit(1);
}
f = fopen(filename, "wb");
fin=fopen(infilename,"rb");
if (!fin) {
error("could not open temporary file");
}
if (!f) {
error("could not open output file");
}
std::cout << c->frame_size*c->channels << std::endl;
samples = new SAMPLE[c->frame_size*c->channels];
outbuf = new uint8_t[FRAMES_PER_BUFFER * NUM_CHANNELS];
while(fread(samples,sizeof(SAMPLE),c->frame_size*c->channels,fin)){
out_size=avcodec_encode_audio(c,outbuf,FRAMES_PER_BUFFER * NUM_CHANNELS,samples);
fwrite(outbuf,sizeof(uint8_t),out_size,f);
}
for(int i=1;i<=4;i++){ //For buffer flushing
out_size=avcodec_encode_audio(c,outbuf,FRAMES_PER_BUFFER * NUM_CHANNELS,NULL);
fwrite(outbuf,sizeof(uint8_t),out_size,f);
}
fclose(f);
delete outbuf;
delete samples;
avcodec_close(c);
av_free(c);
}
There is a usage example at the zlib website: http://www.zlib.net/zlib_how.html
However in the example they are compressing a file. I would like to compress a binary data stored in a buffer in memory. I don't want to save the compressed buffer to disk either.
Basically here is my buffer:
fIplImageHeader->imageData = (char*)imageIn->getFrame();
How can I compress it with zlib?
I would appreciate some code example of how to do that.
zlib.h has all the functions you need: compress (or compress2) and uncompress. See the source code of zlib for an answer.
ZEXTERN int ZEXPORT compress OF((Bytef *dest, uLongf *destLen, const Bytef *source, uLong sourceLen));
/*
Compresses the source buffer into the destination buffer. sourceLen is
the byte length of the source buffer. Upon entry, destLen is the total size
of the destination buffer, which must be at least the value returned by
compressBound(sourceLen). Upon exit, destLen is the actual size of the
compressed buffer.
compress returns Z_OK if success, Z_MEM_ERROR if there was not
enough memory, Z_BUF_ERROR if there was not enough room in the output
buffer.
*/
ZEXTERN int ZEXPORT uncompress OF((Bytef *dest, uLongf *destLen, const Bytef *source, uLong sourceLen));
/*
Decompresses the source buffer into the destination buffer. sourceLen is
the byte length of the source buffer. Upon entry, destLen is the total size
of the destination buffer, which must be large enough to hold the entire
uncompressed data. (The size of the uncompressed data must have been saved
previously by the compressor and transmitted to the decompressor by some
mechanism outside the scope of this compression library.) Upon exit, destLen
is the actual size of the uncompressed buffer.
uncompress returns Z_OK if success, Z_MEM_ERROR if there was not
enough memory, Z_BUF_ERROR if there was not enough room in the output
buffer, or Z_DATA_ERROR if the input data was corrupted or incomplete. In
the case where there is not enough room, uncompress() will fill the output
buffer with the uncompressed data up to that point.
*/
This is an example to pack a buffer with zlib and save the compressed contents in a vector.
void compress_memory(void *in_data, size_t in_data_size, std::vector<uint8_t> &out_data)
{
std::vector<uint8_t> buffer;
const size_t BUFSIZE = 128 * 1024;
uint8_t temp_buffer[BUFSIZE];
z_stream strm;
strm.zalloc = 0;
strm.zfree = 0;
strm.next_in = reinterpret_cast<uint8_t *>(in_data);
strm.avail_in = in_data_size;
strm.next_out = temp_buffer;
strm.avail_out = BUFSIZE;
deflateInit(&strm, Z_BEST_COMPRESSION);
while (strm.avail_in != 0)
{
int res = deflate(&strm, Z_NO_FLUSH);
assert(res == Z_OK);
if (strm.avail_out == 0)
{
buffer.insert(buffer.end(), temp_buffer, temp_buffer + BUFSIZE);
strm.next_out = temp_buffer;
strm.avail_out = BUFSIZE;
}
}
int deflate_res = Z_OK;
while (deflate_res == Z_OK)
{
if (strm.avail_out == 0)
{
buffer.insert(buffer.end(), temp_buffer, temp_buffer + BUFSIZE);
strm.next_out = temp_buffer;
strm.avail_out = BUFSIZE;
}
deflate_res = deflate(&strm, Z_FINISH);
}
assert(deflate_res == Z_STREAM_END);
buffer.insert(buffer.end(), temp_buffer, temp_buffer + BUFSIZE - strm.avail_out);
deflateEnd(&strm);
out_data.swap(buffer);
}
You can easily adapt the example by replacing fread() and fwrite() calls with direct pointers to your data. For zlib compression (referred to as deflate as you "take out all the air of your data") you allocate z_stream structure, call deflateInit() and then:
fill next_in with the next chunk of data you want to compress
set avail_in to the number of bytes available in next_in
set next_out to where the compressed data should be written which should usually be a pointer inside your buffer that advances as you go along
set avail_out to the number of bytes available in next_out
call deflate
repeat steps 3-5 until avail_out is non-zero (i.e. there's more room in the output buffer than zlib needs - no more data to write)
repeat steps 1-6 while you have data to compress
Eventually you call deflateEnd() and you're done.
You're basically feeding it chunks of input and output until you're out of input and it is out of output.
The classic way more convenient with C++ features
Here's a full example which demonstrates compression and decompression using C++ std::vector objects:
#include <cstdio>
#include <iosfwd>
#include <iostream>
#include <vector>
#include <zconf.h>
#include <zlib.h>
#include <iomanip>
#include <cassert>
void add_buffer_to_vector(std::vector<char> &vector, const char *buffer, uLongf length) {
for (int character_index = 0; character_index < length; character_index++) {
char current_character = buffer[character_index];
vector.push_back(current_character);
}
}
int compress_vector(std::vector<char> source, std::vector<char> &destination) {
unsigned long source_length = source.size();
uLongf destination_length = compressBound(source_length);
char *destination_data = (char *) malloc(destination_length);
if (destination_data == nullptr) {
return Z_MEM_ERROR;
}
Bytef *source_data = (Bytef *) source.data();
int return_value = compress2((Bytef *) destination_data, &destination_length, source_data, source_length,
Z_BEST_COMPRESSION);
add_buffer_to_vector(destination, destination_data, destination_length);
free(destination_data);
return return_value;
}
int decompress_vector(std::vector<char> source, std::vector<char> &destination) {
unsigned long source_length = source.size();
uLongf destination_length = compressBound(source_length);
char *destination_data = (char *) malloc(destination_length);
if (destination_data == nullptr) {
return Z_MEM_ERROR;
}
Bytef *source_data = (Bytef *) source.data();
int return_value = uncompress((Bytef *) destination_data, &destination_length, source_data, source.size());
add_buffer_to_vector(destination, destination_data, destination_length);
free(destination_data);
return return_value;
}
void add_string_to_vector(std::vector<char> &uncompressed_data,
const char *my_string) {
int character_index = 0;
while (true) {
char current_character = my_string[character_index];
uncompressed_data.push_back(current_character);
if (current_character == '\00') {
break;
}
character_index++;
}
}
// https://stackoverflow.com/a/27173017/3764804
void print_bytes(std::ostream &stream, const unsigned char *data, size_t data_length, bool format = true) {
stream << std::setfill('0');
for (size_t data_index = 0; data_index < data_length; ++data_index) {
stream << std::hex << std::setw(2) << (int) data[data_index];
if (format) {
stream << (((data_index + 1) % 16 == 0) ? "\n" : " ");
}
}
stream << std::endl;
}
void test_compression() {
std::vector<char> uncompressed(0);
auto *my_string = (char *) "Hello, world!";
add_string_to_vector(uncompressed, my_string);
std::vector<char> compressed(0);
int compression_result = compress_vector(uncompressed, compressed);
assert(compression_result == F_OK);
std::vector<char> decompressed(0);
int decompression_result = decompress_vector(compressed, decompressed);
assert(decompression_result == F_OK);
printf("Uncompressed: %s\n", uncompressed.data());
printf("Compressed: ");
std::ostream &standard_output = std::cout;
print_bytes(standard_output, (const unsigned char *) compressed.data(), compressed.size(), false);
printf("Decompressed: %s\n", decompressed.data());
}
In your main.cpp simply call:
int main(int argc, char *argv[]) {
test_compression();
return EXIT_SUCCESS;
}
The output produced:
Uncompressed: Hello, world!
Compressed: 78daf348cdc9c9d75128cf2fca495164000024e8048a
Decompressed: Hello, world!
The Boost way
#include <iostream>
#include <boost/iostreams/filtering_streambuf.hpp>
#include <boost/iostreams/copy.hpp>
#include <boost/iostreams/filter/zlib.hpp>
std::string compress(const std::string &data) {
boost::iostreams::filtering_streambuf<boost::iostreams::output> output_stream;
output_stream.push(boost::iostreams::zlib_compressor());
std::stringstream string_stream;
output_stream.push(string_stream);
boost::iostreams::copy(boost::iostreams::basic_array_source<char>(data.c_str(),
data.size()), output_stream);
return string_stream.str();
}
std::string decompress(const std::string &cipher_text) {
std::stringstream string_stream;
string_stream << cipher_text;
boost::iostreams::filtering_streambuf<boost::iostreams::input> input_stream;
input_stream.push(boost::iostreams::zlib_decompressor());
input_stream.push(string_stream);
std::stringstream unpacked_text;
boost::iostreams::copy(input_stream, unpacked_text);
return unpacked_text.str();
}
TEST_CASE("zlib") {
std::string plain_text = "Hello, world!";
const auto cipher_text = compress(plain_text);
const auto decompressed_plain_text = decompress(cipher_text);
REQUIRE(plain_text == decompressed_plain_text);
}
This is not a direct answer on your question about the zlib API, but you may be interested in boost::iostreams library paired with zlib.
This allows to use zlib-driven packing algorithms using the basic "stream" operations notation and then your data could be easily compressed by opening some memory stream and doing the << data operation on it.
In case of boost::iostreams this would automatically invoke the corresponding packing filter for every data that passes through the stream.