let me put this from the perspective of encoding movies.
to make a movie, you will need to show at least 24 photos in one second. which means 24fps.
now consider this, if you have an uncompressed movie; say from a video camera, it would be 24 bitmap files put in a sequence,
A mpg video is made by compressing each bmp image frame to a jpg image frame. decent quality but requires some cpu/gpu processing power and some loss of quality.
when you talk about the newer movie codecs; they do more than converting each frame to a jpg. there would be some key frames(p frames) and some delta frames (b frames) each delta frame would only have the changed pixels from w.r.t p frames converted to jpg images. (this is why you will see strange images until the scene change on some badly encoded movies when you seek to random location)
the latest H264 codecs do even more than that to compress the movie further, they apply lot of matrix transformations to decrease movie size and the lost information is guessed using matrix equations (some thing like franhaufer transformations which I never understood.)
So when someone says that the movie is encoded to 800kbps, it means that the codec uses a max of 800kb per scond to encode the frames, means lots of artifacts compared to higher bitrate movies at that particular resolution.
Now coming to your query:
128kbps audio for 2hr movie takes in the order of ~100mb. a dvd quality of sound (ac3 6 ch at 448kbps) takes around ~600mb. DTS takes around 1.3GB. so rest is the data for movie. Ideally a 2-3GB 720p looks as good as ~6GB 1080p movie with ac3 6ch sound. if you go below that, you will find artifacts at fast moving scenes, if you dont have powerful hardware, the VLC will skip frames to compensate the lack of decoding power hence jerky movie at car chases.
So... if you want to play movies like avatar at 12GB, you will need good graphics card to render and a fast hdd to read atleast 2-3mBps speed constantly without lags. (if you try to play from a thumb drive, it will be jerky)
on the other hand, if you have less powerful hardware, just stick with 720.
lastly, compressibilty for anime and 2d cartoons is very high(obviously) so their size is much smaller. PS: highly encoded ****ed rips= yify good encodes with good audio, public china hd, nlupper