How to record and retrieve video with depth data?

I don’t want to build any obj model, but I want to save the video data into a file on Windows for every frame. It seems like Skanect does not provide such a function.

Also, how could I read the information from the file?