i want to optimize the speed of the ffmpeg video creation.
At the moment: Creating small videos and concatenating them with xfade and using HLS after all video chunks are made. So it does take time to make those video chunks, if i have 50 video chunks of 6 seconds each, the waiting time is high.
What i want: Create a big command and use HLS on it, maybe it will start faster so i avoid the chunk creation.
Question: What is better? Is it worth creating that big command? I ask because the refactor is big, i will need to do dynamic input placement and dynamic complex filter with input and output dynamic naming/mapping.
Are there other faster ways? The one with multiple ffmpeg instances to create chunks and concatenate them after they are done takes some time, and i want to change to something faster.