What Makes Maria's Way of Doing Things Better?

[christo930]

I was listening to this talk the other day where the founders of GCC were having a get together on the anniversary of some GCC thing. One of the things they were talking about was how Maria would draw graphics line by line and not through tiles and how this was much better and that this was how arcade machines worked.

 

Is this actually better? if so, why?

 

Also, if a machine is creating graphics 1 scanline at a time, why are there full screen color limits outside of hardware sprites? Shouldn't these limits just apply to a scanline?

 

Alternatively, I could be completely misinterpreting what they said. If so, feel free to correct me....

[+SmittyB]

What they would have been referring to is how a lot of consoles and computers of the time would use a grid of tiles with a number of sprites on top limited to a certain number on screen or number on a row (without clever coding), whereas the 7800 has a number of 'zones' which are arbitrarily tall rows (typically 8 or 16 pixels high) which contain an arbitrary number of objects. These objects are 1 zone high, up to 32 bytes wide (256 pixels in 320A or 128 pixels in 160A modes) and can either read a single chunk of graphics representing one huge object, or read from a list and effectively become a tilemap. The only limit is the amount of RAM/ROM you have to store your 'display-list list' which defines what needs to be drawn, and the time MARIA has to actually read that list and draw it to the screen.

 

With this setup the 7800 can throw around some impressively large visuals on screen as evidenced by the Punch-Out! demo below (thanks for the video Trebor) but the downside is that you spend a lot of your processing time building the lists that define each zone, and this is further hampered by how MARIA has to stop the CPU each time it wants to read any data.

Also, given that a palette is chosen for each object, if you used your objects to simulate a tiled display you would have to have rows of tiles using the same palette to avoid having too many objects. If you were to try having a different palette per tile as other systems allow, you'd have to have 20 objects in a zone before you draw anything else on top and there just isn't the time to do that for the whole screen.

 

To sum it up, it just has different strengths and weaknesses to most other consoles so that it can handle a different kind of game, and providing developers work to its strengths it allows for some pretty impressive imagery.

 

Edited July 10, 2018 by SmittyB