Recherche avancée

Sur d’autres sites (5097)

• Encode youtube live stream to UDP out using youtube-dl and ffmpeg

  1er mars 2023, par Darshan Patel

  I am trying to encode youtube live stream to UDP destination using youtube-dl and ffmpeg with below command

  


  youtube-dl -f best --buffer-size 2M -o - "https://www.youtube.com/watch?v=tkUvWJiTf9A" | ffmpeg -re -f mp4 -i pipe:0 -codec copy -f mpegts udp://192.168.1.107:1234?pkt_size=1316


  


  But its not working, its just downloading ts segments of that live stream.
  
When I am trying with video of youtube its working fine with below commands

  


  youtube-dl -f best --buffer-size 2M -o - "https://www.youtube.com/watch?v=snDI6AaL04g" | ffmpeg -re -f mp4 -i pipe:0 -codec copy -f mpegts udp://192.168.1.107:1234?pkt_size=1316


  


  Any help or suggestion appreciated.

  


• Merging videos. multiple overlays and audio in one step

  3 février 2019, par user14567

  I have a process that I run hourly to update a 60 sec weather video.

  I combine a background video with a generated m4a audio track, scrolling text and separate top and bottom overlays.

  I need all of this in an mp4 and in a TLS (m3u8) format.

  Right now I do this 3 ffmpeg consecutive sessions and end up with a rendered 720p mp4.

  Then I run a 4th ffpmeg session to convert the the mp4 to a TLS version (I need both mp4 and m3u8.)

  I am doing this in 4 ffmpeg runs because I haven’t figured out how to combine the filters and do it in less.

  When I try an combine this, I lose either the overlays or the text. I can’t get it all to come out in one session. I’ve tried comma separating within the filters but I haven’t found the right combination.

  Any assistance in showing me how to combine this work and reduce the number of runs would be greatly appreciated.

  This runs on a Centos 7 box with ffmpeg version 3.4.1.

  #
  
  # Step 1: Combine background video with audio and scrolling text
  
  #
  
  ffmpeg -i /video/weather/media/sunnybeach.mp4 \
  
  -i /video/weather/prod/currentweather.m4a \
  
  -i /video/weather/media/weather-lower-third1.png \
  
  -map 2:v:0 -map 1:a:0 -strict -2 \
  
  -filter_complex  "[0]split[txt][orig];[txt]drawtext=fontfile=/var/www/vhosts/30a.tv/httpdocs/openx/lib/pear/Image/Canvas/Fonts/arial.ttf:fontsize=80:fontcolor=white:x=(w-text_w)/2+20:y=h-70*t:textfile=/video/weather/prod/currentweather.txt:bordercolor=black:borderw=2[txt];[orig]crop=iw:50:0:0[orig];[txt][orig]overlay” \
  
  -t 60 -c:v libx264 -y -s hd720  -loglevel quiet /video/weather/prod/currentweather.mp4 2>&1 >> /dev/null
  
  
  
  if [ $? != 0 ]; then echo fail step 1; exit 1;fi
  
  
  
  #
  
  # Step 2: now overlay current lowerthird
  
  #
  
  ffmpeg -i /video/weather/prod/currentweather.mp4 \
  
  -i /video/weather/media/weather-lower-third1.png \
  
  -strict -2 -filter_complex  "[0:v][1:v]overlay" -c:a copy -y  \
  
  -loglevel quiet $FKWI/currentweather1.mp4 2>&1 >> /dev/null
  
  
  
  if [ $? != 0 ]; then echo fail step 2; exit 1;fi
  
  
  
  #
  
  # Step 3: now overlay current topbanner
  
  #
  
  ffmpeg -i $FKWI/currentweather1.mp4 \
  
  -i /video/weather/media/weather-top-banner.png \
  
  -strict -2 -filter_complex  "[0:v][1:v]overlay" -c:a copy -y  \
  
  -loglevel quiet $FKWI/currentweather.mp4 2>&1 >> /dev/null
  
  
  
  if [ $? != 0 ]; then echo fail step 3; exit 1;fi
  
  
  
  #
  
  # Step 4: Now make TLS version
  
  #
  
  
  
  cd $FKWI/httpdocs/weather
  
  RES1=`/usr/bin/ffmpeg -i $FKWI/httpdocs/currentweather.mp4 \
  
  -profile:v baseline -level 3.0 -s 1280x720 \
  
  -c:a aac -ar 48000 -g 60 -start_number 0 \
  
  -hls_time 6 -hls_list_size 0 -f hls \
  
  -hls_base_url "http://$URL.com/weather/" \
  
  -hls_segment_filename "video%04d.ts" \
  
  -strict -2   -loglevel quiet "currentweather.m3u8" 2>&1>>/dev/null `
  
  
  
  if [ $? != 0 ]; then echo fail step 4; exit 1;fi

  This code yields both videos fine, but it takes 4 runs and I need it to happen in less runs.

• ffmpeg is adding an extra new line into a box

  4 août 2017, par newbie123

  I’m overlaying weather data over my webcam stream. I put a background box filter, but the padding between the text and the box is uneven. There is "a new line" of extended box bellow the text.

  

  Why is that ? There are no empty lines in a text file I’m providing to ffmpeg. Here is my code for the ffmpeg :

  ffmpeg \
  
  -f lavfi -i anullsrc \
  
  -rtsp_transport tcp \
  
  -i "$SOURCE" \
  
  -vcodec libx264 -pix_fmt yuv420p -preset ultrafast -g 20 -b:v 1000k \
  
  -vf "drawtext="fontfile=${FONT}":textfile=${textfile}:x=5:y=55:reload=1: \
  
  fontcolor=white:fontsize=${FONTSIZE}:box=1:boxborderw=5:boxcolor=black@0.5" \
  
  -threads $THREADS -bufsize 512k \
  
  -f flv "$YOUTUBE_URL/$KEY"

  EDIT :

  I found out what the problem was. I was using printf "%s\n" "$variable1" "$variable2" to create a text file. Printf %s\n automatically prints each variable into a new line. The solution was to print the last variable without the new line. Code example :

  #!/bin/sh
  
  LC_CTYPE=en_US.utf8
  
  
  
  # Get APRS weather data from aprs.fi
  
  
  
  wxstation="S55MA-10"
  
  name="Juršče, Pivka"
  
  
  
  
  
  # Basic weather data
  
  temp="$(wget -q https://aprs.fi/weather/a/${wxstation} -O - | grep Temperature | egrep '[-+]?([0-9]*\.[0-9]+|[0-9]+)' -o)"
  
  humidity="$(wget -q https://aprs.fi/weather/a/${wxstation} -O - | grep Humidity | egrep '[-+]?([0-9]*\.[0-9]+|[0-9]+)' -o)"
  
  wind="$(wget -q https://aprs.fi/weather/a/${wxstation} -O - | grep Wind | egrep '[-+]?([0-9]*\.[0-9]+|[0-9]+)' -o | sed -n -e 2p)"
  
  rain="$(wget -q https://aprs.fi/weather/a/${wxstation} -O - | grep Rain | egrep '[-+]?([0-9]*\.[0-9]+|[0-9]+)' -o | sed -n -e 1p)"
  
  
  
  # Telemetry
  
  radioactivity="$(wget -q https://aprs.fi/telemetry/a/${wxstation} -O - | grep Radioactivity | egrep '[-+]?([0-9]*\.[0-9]+|[0-9]+)' -o | sed -n -e 5p)"
  
  
  
  printf "%s\n" "$name" "Temperature: ${temp}\°C" "Humidity: ${humidity}\%" "Wind: ${wind} m/s" "Rain: ${rain} mm/h"
  
  printf "Radioactivity: ${radioactivity} uSv/h"