I'm helping with the Seattle Police body camera Youtube channel, see http://ift.tt/1NrQM8g and http://ift.tt/1MdgbPU The department wants to overredact and publish to Youtube each body camera video as soon as an officer uploads the video to the internal server. To do this SPD is going to use AWS. I want to find out if it's easier to use Lambda to process each video through FFMPEG or EC2 instances.
The process likely will be: 1. Officer uploads video to internal server 2. Internal script copies/uploads video to S3 bucket 3. Either lambda or EC2 gets video, overredacts it, uploads it to video, and removes it from S3
The FFMPEG command I use is: ffmpeg -i [[input]] -threads 2 -crf 20 -preset ultrafast -vf format=gray,"boxblur=6:4:cr=2:ar=2" -c:v libx264 -an [[output]]
Aucun commentaire:
Enregistrer un commentaire