Page 2 of 2
Posted: Wed Jan 31, 2007 1:53 pm
by blargg
Yep, I think the issue is one of VRAM bandwidth. If so, then it would have required faster, more-expensive video RAM, possibly in the cartridge as well. This would probably have required a faster master clock. And having a faster clock might have required faster logic, increasing cost more.
Posted: Wed Jan 31, 2007 2:29 pm
by tepples
tokumaru wrote:Doesn't this limit have to do with time rather than bytes? Isn't it like this because the PPU is not fast enough to fetch more than 8 rows of patterns?
It's like this so that the PPU can reuse its pattern table fetch circuitry for both the sprites and the background. The PPU could have supported up to 16 sprites per scanline with the existing amount of hblank time, but Nintendo and Ricoh chose to have it make garbage nametable fetches instead to save gates.