Sur d’autres sites (13416)

• ffmpeg doesn't sometimes start reading RTSP stream from IP camera without error

  8 décembre 2022, par Petr Dub

  I've an IP camera, which provide RTSP stream. I want to read this stream, rotate it 90° and provide data for displaying video via IIS server. I have a working solution via ffmpeg, but it is not reliable. It ends after several minutes without error and it is not starting sometimes.

  


  I use this command :
ffmpeg.exe -thread_queue_size 128 -i "rtsp://<login>:<pwd>@192.168.0.201:554/live/ch0" -vf "transpose=1" -y -c:a aac -b:a 160000 -ac 1 -s 432x768 -g 50 -hls_time 2 -hls_list_size 1 -start_number 1 -hls_flags delete_segments m:\playlist.m3u8</pwd></login>

  


  It works, but after several minutes (40 to 100) it stops without error (I have no idea why). When I restart the same ffmeg command, it sometimes doesn’t start reading the stream, but it doesn’t produce any error, the ffmpeg process is still running. Error output from ffmeg shows :

  


  ffmpeg version 2022-11-03-git-5ccd4d3060-full_build-www.gyan.dev Copyright (c) 2000-2022 the FFmpeg developers
  built with gcc 12.1.0 (Rev2, Built by MSYS2 project)
  configuration: --enable-gpl --enable-version3 --enable-static --disable-w32threads --disable-autodetect --enable-fontconfig --enable-iconv --enable-gnutls --enable-libxml2 --enable-gmp --enable-bzlib --enable-lzma --enable-libsnappy --enable-zlib --enable-librist --enable-libsrt --enable-libssh --enable-libzmq --enable-avisynth --enable-libbluray --enable-libcaca --enable-sdl2 --enable-libaribb24 --enable-libdav1d --enable-libdavs2 --enable-libuavs3d --enable-libzvbi --enable-librav1e --enable-libsvtav1 --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxavs2 --enable-libxvid --enable-libaom --enable-libjxl --enable-libopenjpeg --enable-libvpx --enable-mediafoundation --enable-libass --enable-frei0r --enable-libfreetype --enable-libfribidi --enable-liblensfun --enable-libvidstab --enable-libvmaf --enable-libzimg --enable-amf --enable-cuda-llvm --enable-cuvid --enable-ffnvcodec --enable-nvdec --enable-nvenc --enable-d3d11va --enable-dxva2 --enable-libvpl --enable-libshaderc --enable-vulkan --enable-libplacebo --enable-opencl --enable-libcdio --enable-libgme --enable-libmodplug --enable-libopenmpt --enable-libopencore-amrwb --enable-libmp3lame --enable-libshine --enable-libtheora --enable-libtwolame --enable-libvo-amrwbenc --enable-libilbc --enable-libgsm --enable-libopencore-amrnb --enable-libopus --enable-libspeex --enable-libvorbis --enable-ladspa --enable-libbs2b --enable-libflite --enable-libmysofa --enable-librubberband --enable-libsoxr --enable-chromaprint
  libavutil      57. 40.100 / 57. 40.100
  libavcodec     59. 51.101 / 59. 51.101
  libavformat    59. 34.101 / 59. 34.101
  libavdevice    59.  8.101 / 59.  8.101
  libavfilter     8. 49.101 /  8. 49.101
  libswscale      6.  8.112 /  6.  8.112
  libswresample   4.  9.100 /  4.  9.100
  libpostproc    56.  7.100 / 56.  7.100


  


  The next row should be :

  


  Input #0, rtsp, from &#x27;rtsp://<login>:<pwd>@192.168.0.201:554/live/ch0&#x27;</pwd></login>

  


  Any ideas, what should I do ? At least to force ffmpeg end with error and not stays “running”…

  


  Edit : I finally wrote a windows service, which starts the ffmpeg and monitors its error output. When nothing comes to this output for several seconds, it kills the ffmpeg process and starts a new one.

  


• av::av_encode_video produces blurred video from crisp files

  9 janvier 2023, par Abdirizak

  I am using the av::av_encode_video() function to produce a .mp4 video from png files, as in this answer. I pay special attention for the png files to be perfectly crisp, i.e. I plot matrices to png with exactly the same dimensions, so that every matrix cell maps to one cell in the png. Here is code to illustrate what I mean :

  


  library(av)

n.times <- 24*10
n.row <- 50
n.col <- 26

out.dir <- "/Users/Abdirizak/Documents/some/directory/"

for (i in 1:n.times) {
   
   current.mat <- matrix(data = rnorm(n.row*n.col), nrow = n.row, ncol = n.col)
   
   setwd(out.dir)
   png(filename = paste0(i, ".png"), width = n.row, height = n.col)
   
   par(mar = rep(0,4))
   image(current.mat)
   
   dev.off()
   
}

setwd(out.dir)
png.files <- paste0(out.dir, list.files(pattern = ".png$"))

av::av_encode_video(input = png.files, output = "000.mp4", framerate = 24)


  


  Now, even though my input .png files are perfectly crisp, the output .mp4 video is blurry. How can I prevent that from happening ? I.e. how can I get a perfectly crisp video ? (I am aware that this will likely inflate the file size of the .mp4 file)

  


  I already digged a bit into the vfilter argument of av_encode_video() and the underlying ffmpeg filter graphs on here, but could not yet really get the hang on it.

  


  I am looking at the .mp4 file through QuickTime Player on MacOS Ventura 13.0.1, and using R version 4.2.2 through RStudio 2022.12.0.353

  


  Edit 9.1.2023 :

  


  To further clarify, here is an example screenshot of a png image (not the original png file because that one appears too small here) :

  


  

  


  As I said, a perfectly crisp image.

  


  On the other hand, a screenshot of the .mp4 movie looks like this :

  


  

  


  Somewhat unsatisfactory. Any suggestions are welcome.

  


• ffmpeg video duration is not equal to duration after hls process

  9 décembre 2022, par Timur Khazhiev

  Problem

  


  For videofile input I preprocess/encode it to libx264. Then I'm trying to cut video into 2 second chunks except last one (last one being < 2 seconds) using hls processing

  &#xA;

  ffmpeg -v error -i video.mp4 -x264-params "keyint=60:min-keyint=60:scenecut=0" 
-hls_time 2 -hls_list_size 0 -hls_segment_filename chunk%03d.ts index.m3u8


  &#xA;

  which follows this problem

  &#xA;

  then looking into index.m3u8 file reveals that sum of all chunks' durations does not add up to original video duration :

  &#xA;

  % ffprobe -loglevel quiet -print_format flat -show_entries format=duration index.m3u8     

format.duration="204.880000"

% ffprobe -loglevel quiet -print_format flat -show_entries format=duration video.mp4 

format.duration="204.892813"


  &#xA;

  Moreover duration of a chunk in index.m3u8 is not the same as in chunk file.

  &#xA;

  % ffprobe -loglevel quiet -print_format flat -show_entries format=duration chunk000.ts

format.duration="2.023222"

% cat index.m3u8
#EXTM3U
#EXT-X-VERSION:3
#EXT-X-TARGETDURATION:2
#EXT-X-MEDIA-SEQUENCE:0
#EXTINF:2.000000,
chunk000.ts
#EXTINF:2.000000,
chunk001.ts
#EXTINF:2.000000,
chunk002.ts
...


  &#xA;

  Misc

  &#xA;

  Used these videos : Me at the zoo and I finally found a useful monorail.

  &#xA;

  OS : macos Monterey 12.3.1 (21E258)

  &#xA;

  ffmpeg version 5.1.2 Copyright (c) 2000-2022 the FFmpeg developers
built with Apple clang version 14.0.0 (clang-1400.0.29.102)
configuration: --prefix=/usr/local/Cellar/ffmpeg/5.1.2 --enable-shared --enable-pthreads --enable-version3 --cc=clang --host-cflags= --host-ldflags= --enable-ffplay --enable-gnutls --enable-gpl --enable-libaom --enable-libbluray --enable-libdav1d --enable-libmp3lame --enable-libopus --enable-librav1e --enable-librist --enable-librubberband --enable-libsnappy --enable-libsrt --enable-libtesseract --enable-libtheora --enable-libvidstab --enable-libvmaf --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxml2 --enable-libxvid --enable-lzma --enable-libfontconfig --enable-libfreetype --enable-frei0r --enable-libass --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libspeex --enable-libsoxr --enable-libzmq --enable-libzimg --disable-libjack --disable-indev=jack --enable-videotoolbox
libavutil      57. 28.100 / 57. 28.100
libavcodec     59. 37.100 / 59. 37.100
libavformat    59. 27.100 / 59. 27.100
libavdevice    59.  7.100 / 59.  7.100
libavfilter     8. 44.100 /  8. 44.100
libswscale      6.  7.100 /  6.  7.100
libswresample   4.  7.100 /  4.  7.100
libpostproc    56.  6.100 / 56.  6.100


  &#xA;

  What I tried

  &#xA;

  &#xA;
  • Preprocess/transcode video before proceeding with hls

  &#xA;

  • Force keyframes before proceeding with hls

  &#xA;

  • Use 1, 3, 4 seconds chunks

  &#xA;

  • Use different GOP sizes

  &#xA;

  • -hls_flags split_by_time

  &#xA;

  &#xA;

  but it leads to same result or very unequal chunks

  &#xA;

  Questions

  &#xA;

  &#xA;
  1. Is it ok or I'm missing something ?

  &#xA;

  2. If not, what parameters/options should I use to force chunks to be precisely 2 seconds ?

  &#xA;

  &#xA;