I am trying to encode a RGB24 image sequence into a mathematically (not merely visually) lossless video. huffyuv was suggested on many online forums so I tried the following.
ffmpeg -i frames\%06d.png -vcodec huffyuv test.avi
The resulting video was then decoded into frames again using ffmpeg
ffmpeg -i test.avi outframes\%06d.png
However, the input and output frames are not bit-by-bit identical as promised by huffyuv here. Any idea how I can accomplish this? My eventual goal is to read the video file using OpenCV but I am willing to cross that bridge later once I obtain a losslessly encoded video file.
This SO question mentions an attempt to obtain a lossless h264 avi and the summary of responses seems to indicate h264 cannot completely accomplish lossless encoding.
Once again, to emphasize, I am interested in bit-by-bit identical encoding, not just visually similar. Large file sizes are acceptable as is large compression/decompression time.