I took a look at the ISO and there are five videos. They're longer than that clip though. It appears to be 16 colors, fixed palette for the whole thing (per movie stream), and some ADPCM audio at the end of the frame. Though it doesn't appear to be enough to sync with the video. Maybe it's buffered. The frame is 16x16 cells but only 16x13 cells (8x8 pixel cells) are shown. Whatever they were trying to do, they probably had issues getting it to work. The video frame packet size is 8k. That's enough to do 15fps (huvideo driver maintains a steady 122k per second). Whether it does or not, I dunno.
Given the wasted space per frame packet, there's enough room a palette, palette map, and decent audio for each frame (1.5k room for it). You only need 266 bytes @ 15fps for 8khz ADPCM (there's enough room do to 16khz too). The palette map would be about 104 bytes (you only need nybbles stored). Palette data per frame would be 288byte packed (256 color entries). But for whatever reason, they don't.
The animation in the game is sub-optimal as well. That is to say, the storage method. Most of it is stored as full frames, and read into memory as full frames. They could have simply compressed those screen as redundant tile data in memory. Nothing fancy at all. That would have allowed a lot more animation. I chalk this one up to minimal effort and probably some lack of knowledge.