The film encoded on your typical DVD is stored mostly progressively (as frames). The MPEG2 decoder produces interlaced fields from these frames. Then for progressive players, the deinterlacer grabs those interlaced fields and reconstructs the original progressive frames. I have been told that if you play DVDs on an X-Box, it is a native progressive device. Thus it will spit out the progressive frames directly, rather than interlacing then de-interlacing them.
4:2:2 vs 4:2:0 has nothing to do with progressive. What 4:2:0 does is discard chroma information from entire lines to save space. The chroma upsampling process is then used to take 1 line of chroma and reconstruct the original 2 lines. But the way you are supposed to do this depends on whether the encoder was using interlaced or progressive chroma encoding mode. DVD players with the "chroma bug" treat all 4:2:0 content as field data, even though some of it is frame data. 4:2:2, whether used on fields or frames, does not discard lines and thus does not require the same sort of chroma upsampling process.
The point still stands. If they encoded chroma info in at least 4:2:2, not only the bug would go away, the colors would be displayed more accurately (much less and simpler interpolation).
There would be more color information but the bitrate required would also increase. The whole point of discarding chroma information is that the human eye is much less sensitive to chroma than it is to luma. The point is to maintain maximum quality for a GIVEN bitrate. Discarding chroma allows the encoder to turn down the compression ratio while using the same space.
They can't encode 4:2:2 on a DVD. The MPEG-2 standard is divided into different profiles and levels. DVD and most other MPEG-2 based products use Main Profile which is strictly 4:2:0. There is a higher profile for professional use which is 4:2:2. For lack of a more suitable name they just named it "4:2:2 Profile".
Nevertheless, the DVD standard requires what is called Main Profile, Main Level (MP@ML). Furthermore, it uses only a subset of MP@ML. For example, DVD can't use progressive sequences, although the required interlaced sequences can have progressively encoded frames. (There's a difference between a progressive sequence and an interlaced sequence of progressively encoded frames. The latter includes field repeat flags to create 3:2 pulldown.)
Even HDTV is 4:2:0, because like DVD it is Main Profile, although instead of Main Level it uses High Level (MP@HL) which basically means higher resolutions than MP@ML.
Yes, if DVD used studio profile, we would not have this problem. It all comes down to econimics, at the time they started DVD many many years ago, memory was expensive.
The goal is to build players as cheap as possible. I am sure Ian will correct me if I am wrong, but some players coming out of China cut the RAM in half and double the processing speed just to save money.
Yeah, Chinese manufacturers are especially nutty about saving memory. One Chinese manufacturer decided that they wanted to transcode PAL->NTSC, but they didn't want to allocate the extra memory (PAL has more lines of video than NTSC, so normally it takes more memory to hold a PAL image for transcoding). So they switched the transcoder from 16 bit per pixel color to 8 bit per pixel. (Normal YUV data works out to be 16 bits per pixel).