Actually it is possible to see color fringe effects also on msx with NTSC output.
Having a control on them would be very hard when you pass from a machine to another
i've read the explanation. this confirm my initial suspicious that the VDP (expecially the TMS) is some kind of very limited not , only in what it has to offer out of the box, also in what it is, it is architected in a way that is extremely limited and hard to trick:
Lets take some examples:
- the speccy, the ULA is rather limited than VDP, but they achieved things like more than 2 color per 8x8 cell. On msx you cannot immagine this on msx in screen 1 due to the relatively complex vram I/O access, you have the stupid limit of 8 chars sharing the same background/foreground color, you need to resort to screen 2 with a big memory trade off.
- the cga: we have seen how a relatively (on the specs) limited piece of hw could be tricked like a hell to achieve stunning things
- the C64. We have seen every sort of manipulations of the VIC-II, like sprites on border, 9 on scanline (!), every kind of split screen, etc.....
what has the vdp to offer ? nothing. It has hybrid modes, that could be useful, but: you get sprite artifacts.
Maybe, but the GAMES from the machines capable do do those "tricks" sucked 35 years ago and still suck today.
My observations are only related to the tech specs, game quality is not strictly related to hw features.
I can program a sucking game even on a Gfx Card used on nowadays pc.
the vdp it's incredibly poorly designed.
For example:
take the screen 1 mode, could be a great mode without having 12KB of memory to manage, but what in the hell one could design a color attribute table of ONLY 32 Bytes? 256 bytes was a so huge waste of memory?
take the screen 3 mode: multicolor... almost useless.
take the screen 2 mode: the multiple banks could have been more useful with a minimum effort... let's try to play with it and you get sprite artifacts, based on chip temperature!!!
.... the list could be more more more lengthy ...
