there are a few times when we are essentially transcoding video, which is fine to do if the video is uncompressed, we shouldn't get too much transcoding loss (though we should test that), but for that to work we need to make sure that when we transcode we keep the same encoding in source & target.
making assumptions is not good enough - currently the VideoWriter hardcodes a handful of encoding settings, but we absolutely not assume that every video we handle is raw avi grayscale video, and should have some option like encode_like: Path in the VideoWriter that first ffprobe's the encoding args from the given video path, and then merges any given explicit output kwargs on top of that.
there are a few times when we are essentially transcoding video, which is fine to do if the video is uncompressed, we shouldn't get too much transcoding loss (though we should test that), but for that to work we need to make sure that when we transcode we keep the same encoding in source & target.
making assumptions is not good enough - currently the
VideoWriterhardcodes a handful of encoding settings, but we absolutely not assume that every video we handle is raw avi grayscale video, and should have some option likeencode_like: Pathin the VideoWriter that first ffprobe's the encoding args from the given video path, and then merges any given explicit output kwargs on top of that.