Recherche avancée

Sur d’autres sites (7562)

• Understanding a script which uses ffmpeg to send rtmp input to node.js script

  4 juin 2022, par Arpit Shukla

  I was trying to understand this shell script which uses ffmpeg to take an rtmp input stream and send it to a node.js script. But I am having trouble understanding the syntax. What is going on here ?

  


  The script :

  


  while :
do
  echo "Loop start"

  feed_time=$(ffprobe -v error -show_entries format=start_time -of default=noprint_wrappers=1:nokey=1 $RTMP_INPUT)
  printf "feed_time value: ${feed_time}"

  if [ ! -z "${feed_time}" ]
  then
  ffmpeg -i $RTMP_INPUT -tune zerolatency -muxdelay 0 -af "afftdn=nf=-20, highpass=f=200, lowpass=f=3000" -vn -sn -dn -f wav -ar 16000 -ac 1 - 2>/dev/null | node src/transcribe.js $feed_time

  else
  echo "FFprobe returned null as a feed time."
  
  fi

  echo "Loop finish"
  sleep 3
done


  


  

  • What is feed_time here ? What does it represent ?

  


  • What is this portion doing - 2>/dev/null | node src/transcribe.js $feed_time ?

  


  • What is the use of sleep 3 ? Does this mean that we are sending audio stream to node.js in chuncks of 3 seconds ?

  


  


• AWS Rekognition error : Chunk video failed

  14 juin 2022, par Stefano Leone

  I'm using and launching Amazon Rekognition on my videos uploaded into my S3 with python. Every video is converted with FFMPEG with AAC Audio Codec and H264 Video Codec and then given to Rekognition. Videos are always fine, the problem is that only 60-70% of videos are processed successfully, while the rest of videos goes into error. In particular, inside the JSON returned from Rekognition :

  


  {&#x27;JobId&#x27;: &#x27;<id of="of" my="my" job="job">&#x27;, &#x27;Status&#x27;: &#x27;FAILED&#x27;, &#x27;API&#x27;: &#x27;StartFaceDetection&#x27;, &#x27;Message&#x27;: &#x27;Chunk video failed: Only 1 I-frames found in video&#x27;, &#x27;Timestamp&#x27;: 1655118632996, &#x27;Video&#x27;: {&#x27;S3ObjectName&#x27;: &#x27;<my video="video" inside="inside" s3="s3">&#x27;, &#x27;S3Bucket&#x27;: &#x27;<my s3="s3">&#x27;}, &#x27;ErrorCode&#x27;: &#x27;VideoNotDecodable&#x27;}&#xA;</my></my></id>

  


  As you can see, I get an error "Chunk video failed: only 1 I-Frames found in video". Honestly I don't know what is that, but the fact that I convert every video in the same way, but Rekognition fails only with some, makes me crazy. Googling was not helpful, hope you can tell me what's wrong.

  


• Thumbnails from S3 Videos using FFMPEG - "No such file or directory : '/bin/ffmpeg'"

  28 juin 2022, par Nico

  I am trying to generate thumbnails from videos in an S3 bucket every x frames by following this documentation : https://aws.amazon.com/blogs/media/processing-user-generated-content-using-aws-lambda-and-ffmpeg/

  


  I am at the point where I'm testing the Lambda code provided in the documentation, but receive this error in CloudWatch Logs :

  


  

  


  Here is the portion of the Lambda code associated with this error :

  


  

  


  Any help is appreciated. Thanks !