Reputation: 39913
I'd like to analyze some video. I'd love to be able to view a video as a series of image frames, but it seems that for compression sake this is hard to do out of the box.
Which file format is most suited for computer-vision type analysis of video data? Basically, what video format is easiest to dissect with a program? I don't necessarily care about file size or the nature of the format as long as it is easy to crack open.
Upvotes: 0
Views: 499
Reputation: 7148
I don't know what your platform is, but on most platforms you have an API that abstracts away most of the complexity of video decoding for you. See for example OpenCV VideoCapture for some documentation on how to do this in OpenCV. VideoForWindows and/or some directx or windows media API's make opening most any video format trivial on windows, the gstreamer framework does the same on linux. So on most platforms no video format is inherently harder to process than any other format. That said, there are of course considerations of file-size and video quality.
You say you don't really care about either, so I'd go for a high quality video codec that still offers good compression. If you go for uncompressed video or even separate images you get all kinds of data handling issues that you will not enjoy, so some compression is usually preferable. I'd suggest ogg theora if you have control over your codec. It is open source and available for most platforms.
Upvotes: 1