Recherche avancée
Autres articles (92)
-
Contribute to a better visual interface
13 avril 2011MediaSPIP is based on a system of themes and templates. Templates define the placement of information on the page, and can be adapted to a wide range of uses. Themes define the overall graphic appearance of the site.
Anyone can submit a new graphic theme or template and make it available to the MediaSPIP community. -
Automated installation script of MediaSPIP
25 avril 2011, parTo overcome the difficulties mainly due to the installation of server side software dependencies, an "all-in-one" installation script written in bash was created to facilitate this step on a server with a compatible Linux distribution.
You must have access to your server via SSH and a root account to use it, which will install the dependencies. Contact your provider if you do not have that.
The documentation of the use of this installation script is available here.
The code of this (...) -
Les autorisations surchargées par les plugins
27 avril 2010, parMediaspip core
autoriser_auteur_modifier() afin que les visiteurs soient capables de modifier leurs informations sur la page d’auteurs
Sur d’autres sites (8362)
-
save a serial of images(cv::Mat) to a mp4 file in Variable Frame Rate mode by using ffmpeg library, how to set the pts ?
2 mai 2023, par ollydbg23In C++ code, I can correctly save a serial of images(opencv's
cv::Mat) to a mp4 file by usingffmpeglibrary, see the question and answer here : avformat_write_header() function call crashed when I try to save several RGB data to a output.mp4 file

Now here comes another question :


Rotem' answer in that question can have the
output.mp4saved correctly. When play the mp4 file, I see the frames(OpenCV'scv::Matimage) shows in a const rate.

What I can do if I have got the frames not in a const frequency, for example, the I got the first frame at the
0ms, and the second frame at the50msthe third frame at75ms, so that the each frame as associated time stamp, for example, those time stamp array are something like below :

int timestamp[100] = {0, 50, 75, ...};


What is the method to modify the Rotem's answer to reflect this ? It looks like I have to change the
ptsfield of eachframe. Because I just test the code, if I change this :

yuvFrame->pts = av_rescale_q(frame_count*frame_count, outCodecCtx->time_base, outStream->time_base); //Set PTS timestamp

// note I change from frame_count to frame_count*frame_count


Then the
output.mp4plays slower and slower, because the later frame has largeptsvalues.

Thanks.


EDIT


This is the code I'm currently used :


#include <iostream>
#include <vector>
#include <cstring>
#include <fstream>
#include <sstream>
#include <stdexcept>
#include <opencv2></opencv2>opencv.hpp>
extern "C" {
#include <libavutil></libavutil>imgutils.h>
#include <libavcodec></libavcodec>avcodec.h>
#include <libavformat></libavformat>avformat.h>
#include <libavutil></libavutil>opt.h>
}

#include<cstdlib> // to generate time stamps

using namespace std;
using namespace cv;

int main()
{
 // Set up input frames as BGR byte arrays
 vector<mat> frames;

 int width = 640;
 int height = 480;
 int num_frames = 100;
 Scalar black(0, 0, 0);
 Scalar white(255, 255, 255);
 int font = FONT_HERSHEY_SIMPLEX;
 double font_scale = 1.0;
 int thickness = 2;

 for (int i = 0; i < num_frames; i++) {
 Mat frame = Mat::zeros(height, width, CV_8UC3);
 putText(frame, std::to_string(i), Point(width / 2 - 50, height / 2), font, font_scale, white, thickness);
 frames.push_back(frame);
 }

 // generate a serial of time stamps which is used to set the PTS value
 // suppose they are in ms unit, the time interval is between 30ms to 59ms
 vector<int> timestamps;

 for (int i = 0; i < num_frames; i++) {
 int timestamp;
 if (i == 0)
 timestamp = 0;
 else
 {
 int random = 30 + (rand() % 30);
 timestamp = timestamps[i-0] + random;
 }

 timestamps.push_back(timestamp);
 }

 // Populate frames with BGR byte arrays

 // Initialize FFmpeg
 //av_register_all();

 // Set up output file
 AVFormatContext* outFormatCtx = nullptr;
 //AVCodec* outCodec = nullptr;
 AVCodecContext* outCodecCtx = nullptr;
 //AVStream* outStream = nullptr;
 //AVPacket outPacket;

 const char* outFile = "output.mp4";
 int outWidth = frames[0].cols;
 int outHeight = frames[0].rows;
 int fps = 25;

 // Open the output file context
 avformat_alloc_output_context2(&outFormatCtx, nullptr, nullptr, outFile);
 if (!outFormatCtx) {
 cerr << "Error: Could not allocate output format context" << endl;
 return -1;
 }

 // Open the output file
 if (avio_open(&outFormatCtx->pb, outFile, AVIO_FLAG_WRITE) < 0) {
 cerr << "Error opening output file" << std::endl;
 return -1;
 }

 // Set up output codec
 const AVCodec* outCodec = avcodec_find_encoder(AV_CODEC_ID_H264);
 if (!outCodec) {
 cerr << "Error: Could not find H.264 codec" << endl;
 return -1;
 }

 outCodecCtx = avcodec_alloc_context3(outCodec);
 if (!outCodecCtx) {
 cerr << "Error: Could not allocate output codec context" << endl;
 return -1;
 }
 outCodecCtx->codec_id = AV_CODEC_ID_H264;
 outCodecCtx->codec_type = AVMEDIA_TYPE_VIDEO;
 outCodecCtx->pix_fmt = AV_PIX_FMT_YUV420P;
 outCodecCtx->width = outWidth;
 outCodecCtx->height = outHeight;
 outCodecCtx->time_base = { 1, fps*1000 }; // 25000
 outCodecCtx->framerate = {fps, 1}; // 25
 outCodecCtx->bit_rate = 4000000;

 //https://github.com/leandromoreira/ffmpeg-libav-tutorial
 //We set the flag AV_CODEC_FLAG_GLOBAL_HEADER which tells the encoder that it can use the global headers.
 if (outFormatCtx->oformat->flags & AVFMT_GLOBALHEADER)
 {
 outCodecCtx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER; //
 }

 // Open output codec
 if (avcodec_open2(outCodecCtx, outCodec, nullptr) < 0) {
 cerr << "Error: Could not open output codec" << endl;
 return -1;
 }

 // Create output stream
 AVStream* outStream = avformat_new_stream(outFormatCtx, outCodec);
 if (!outStream) {
 cerr << "Error: Could not allocate output stream" << endl;
 return -1;
 }

 // Configure output stream parameters (e.g., time base, codec parameters, etc.)
 // ...

 // Connect output stream to format context
 outStream->codecpar->codec_id = outCodecCtx->codec_id;
 outStream->codecpar->codec_type = AVMEDIA_TYPE_VIDEO;
 outStream->codecpar->width = outCodecCtx->width;
 outStream->codecpar->height = outCodecCtx->height;
 outStream->codecpar->format = outCodecCtx->pix_fmt;
 outStream->time_base = outCodecCtx->time_base;

 int ret = avcodec_parameters_from_context(outStream->codecpar, outCodecCtx);
 if (ret < 0) {
 cerr << "Error: Could not copy codec parameters to output stream" << endl;
 return -1;
 }

 outStream->avg_frame_rate = outCodecCtx->framerate;
 //outStream->id = outFormatCtx->nb_streams++; <--- We shouldn't modify outStream->id

 ret = avformat_write_header(outFormatCtx, nullptr);
 if (ret < 0) {
 cerr << "Error: Could not write output header" << endl;
 return -1;
 }

 // Convert frames to YUV format and write to output file
 int frame_count = -1;
 for (const auto& frame : frames) {
 frame_count++;
 AVFrame* yuvFrame = av_frame_alloc();
 if (!yuvFrame) {
 cerr << "Error: Could not allocate YUV frame" << endl;
 return -1;
 }
 av_image_alloc(yuvFrame->data, yuvFrame->linesize, outWidth, outHeight, AV_PIX_FMT_YUV420P, 32);

 yuvFrame->width = outWidth;
 yuvFrame->height = outHeight;
 yuvFrame->format = AV_PIX_FMT_YUV420P;

 // Convert BGR frame to YUV format
 Mat yuvMat;
 cvtColor(frame, yuvMat, COLOR_BGR2YUV_I420);
 memcpy(yuvFrame->data[0], yuvMat.data, outWidth * outHeight);
 memcpy(yuvFrame->data[1], yuvMat.data + outWidth * outHeight, outWidth * outHeight / 4);
 memcpy(yuvFrame->data[2], yuvMat.data + outWidth * outHeight * 5 / 4, outWidth * outHeight / 4);

 // Set up output packet
 //av_init_packet(&outPacket); //error C4996: 'av_init_packet': was declared deprecated
 AVPacket* outPacket = av_packet_alloc();
 memset(outPacket, 0, sizeof(outPacket)); //Use memset instead of av_init_packet (probably unnecessary).
 //outPacket->data = nullptr;
 //outPacket->size = 0;

 yuvFrame->pts = av_rescale_q(timestamps[frame_count], outCodecCtx->time_base, outStream->time_base); //Set PTS timestamp

 // Encode frame and write to output file
 int ret = avcodec_send_frame(outCodecCtx, yuvFrame);
 if (ret < 0) {
 cerr << "Error: Could not send frame to output codec" << endl;
 return -1;
 }
 while (ret >= 0) {
 ret = avcodec_receive_packet(outCodecCtx, outPacket);
 if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF) {
 break;
 } else if (ret < 0) {
 cerr << "Error: Could not receive packet from output codec" << endl;
 return -1;
 }

 //av_packet_rescale_ts(&outPacket, outCodecCtx->time_base, outStream->time_base);

 outPacket->stream_index = outStream->index;

 outPacket->duration = av_rescale_q(1, outCodecCtx->time_base, outStream->time_base); // Set packet duration

 ret = av_interleaved_write_frame(outFormatCtx, outPacket);
 av_packet_unref(outPacket);
 if (ret < 0) {
 cerr << "Error: Could not write packet to output file" << endl;
 return -1;
 }
 }

 av_frame_free(&yuvFrame);
 }

 // Flush the encoder
 ret = avcodec_send_frame(outCodecCtx, nullptr);
 if (ret < 0) {
 std::cerr << "Error flushing encoder: " << std::endl;
 return -1;
 }

 while (ret >= 0) {
 AVPacket* pkt = av_packet_alloc();
 if (!pkt) {
 std::cerr << "Error allocating packet" << std::endl;
 return -1;
 }
 ret = avcodec_receive_packet(outCodecCtx, pkt);

 // Write the packet to the output file
 if (ret == 0)
 {
 pkt->stream_index = outStream->index;
 pkt->duration = av_rescale_q(1, outCodecCtx->time_base, outStream->time_base); // <---- Set packet duration
 ret = av_interleaved_write_frame(outFormatCtx, pkt);
 av_packet_unref(pkt);
 if (ret < 0) {
 std::cerr << "Error writing packet to output file: " << std::endl;
 return -1;
 }
 }
 }


 // Write output trailer
 av_write_trailer(outFormatCtx);

 // Clean up
 avcodec_close(outCodecCtx);
 avcodec_free_context(&outCodecCtx);
 avformat_free_context(outFormatCtx);

 return 0;
}
</int></mat></cstdlib></stdexcept></sstream></fstream></cstring></vector></iostream>

Especially, I have those changes to the original Rotem's answer :


Fist, I have some code to generate a time stamp array :


// generate a serial of time stamps which is used to set the PTS value
 // suppose they are in ms unit, the time interval is between 30ms to 59ms
 vector<int> timestamps;

 for (int i = 0; i < num_frames; i++) {
 int timestamp;
 if (i == 0)
 timestamp = 0;
 else
 {
 int random = 30 + (rand() % 30);
 timestamp = timestamps[i-0] + random;
 }

 timestamps.push_back(timestamp);
 }
</int>

Second, I just set the PTS by those values :


yuvFrame->pts = av_rescale_q(timestamps[frame_count], outCodecCtx->time_base, outStream->time_base); //Set PTS timestamp


Note that I have set the fps like below :


outCodecCtx->time_base = { 1, fps*1000 }; // 25000
 outCodecCtx->framerate = {fps, 1}; // 25


Now, when I run the program, I got a lot of warnings in the console :


[libx264 @ 0000022e7fa621c0] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2
[libx264 @ 0000022e7fa621c0] profile High, level 3.0, 4:2:0, 8-bit
[libx264 @ 0000022e7fa621c0] 264 - core 164 r3094M bfc87b7 - H.264/MPEG-4 AVC codec - Copyleft 2003-2022 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=6 lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=25 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=abr mbtree=1 bitrate=4000 ratetol=1.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] non-strictly-monotonic PTS
[libx264 @ 0000022e7fa621c0] invalid DTS: PTS is less than DTS
[mp4 @ 0000022e090b2300] pts (592) < dts (1129348497) in stream 0
Error: Could not write packet to output file


Any ideas ?
Thanks.


-
How to make cropping when doing motion-tracking look super smooth ? (using cv2 and yolo v8)
4 décembre 2024, par Thomas LancerI'm using yolo to track a face in a video on a frame by frame basis, and then I'm using cv2 to crop the video (with some padding) around the speakers face so it creates a motion tracking type effect. (short example here : https://www.instagram.com/reel/DCIFkFEObkE/?hl=en)


The problem I'm running into is my final crop looks "shaky". Like it's flickering a bit. I think this is because every frame yolo is updating the coordinates, and so there's a slight change in x, y and because of that it creates this shakiness/flickering effect. Here's what mine looks like compared to the example : https://drive.google.com/file/d/1MKHWK-5EH5abSq6i32r71GWld76tN1GP/view?usp=sharing


You can see there's this "shakiness" type look to it. it's especially apparent when the speaker is relatively still.


I've tried doing an exponential moving average and a few other averaging techniques but none have fixed it.


Is my fundamental approach of doing a frame by frame crop wrong ? or is there just some smoothing algorithm or some dumb mistake that you think I may be making ? Any advice is mega appreciated !


-
Extracting thumbnails with FFMPEG is super slow on large video files ? [duplicate]
5 octobre 2015, par vaidThis question already has an answer here :
I extract thumbnails from a
.MOVfile using FFMPEG on Linus (Debian 64bit).The file I extract the thumbnail from is about 430 Megabytes large.
I use the following command to do so :
ffmpeg -i 'largeVideoFile.mov' -ss 00:14:37 -vframes 1 'thumbnail.jpg'It takes well over 3 minutes for a single frame to be extracted.
How can I speed it up ?
Syndiquer tout le site
