Nes

[aprentice]

yes, though a xor 0x10 is simpler, oh, and remember to do an if addr&3==0.

yes, start rendering on line 1, just feed the render function with (line-1). vblank still ends at scanline 0 (sr=0).

doing the palette mirroring inverts the colors in every game, im probably doing something wrong, this change is in the ppu_write ?

edit: fixed it, overlooked your addr&0x3

thanks dude for your help!

 

[hap]

neat

addr&=0x3fff; // mirrored
if(addr>=0x3F00) {
// 0 4 8 c are the same as 10 14 18 1c
if ((addr&3)==0) vmem.vram[(addr&0x3F1F)^0x10]=data;
vmem.vram[addr&0x3F1F]=data;
}
else {
if (addr>=0x2000) { // <0x2000=patterntable. if you add external vram (like unrom), it should be allowed then
vmem.vrompage[addr>>10&0xf][addr&0x3FF]=data;
}
}


*edit* good, do you get the same results now as in those images ?

 

[aprentice]

neat

addr&=0x3fff; // mirrored
if(addr>=0x3F00) {
// 0 4 8 c are the same as 10 14 18 1c
if ((addr&3)==0) vmem.vram[(addr&0x3F1F)^0x10]=data;
vmem.vram[addr&0x3F1F]=data;
}
else {
if (addr>=0x2000) { // <0x2000=patterntable. if you add external vram (like unrom), it should be allowed then
vmem.vrompage[addr>>10&0xf][addr&0x3FF]=data;
}
}


*edit* good, do you get the same results now as in those images ?

yeah, exact down to the scanline, i went down your list and fixed everything, now i need to figure out sprite 0 hit, controllers and whatever mapper super mario bros 3 uses

 

[aprentice]

implemented mapper4 (mmc3), only game that seems to run is super mario bros 2 though, there seems to be a bug in my code. What confuses me is command 6 and 7 on $8001 and how $8000 works, does $8000 swap banks in and out cause i know the banks suppose to change on $8001..

 

[hap]

Let's say 'command' 6 is prg bank 3, and 7 is prg bank 4, then depending on 0x8000.6, the prg swap mode bit or whatever ;p, the prg banks will swap like this:

0x8000.6 is clear:
prg0=bank 3
prg1=bank 4
prg2=second to last bank
prg3=last bank

0x8000.6 is set:
prg0=second to last bank
prg1=bank 4
prg2=bank 3
prg3=last bank

writing to 0x8000 will change prg banks if bit 6 was different as it was before the write.

 

[aprentice]

added controllers although they dont seem to be working 100%, in nestest every button except select works, select doesnt seem to work in any game, and in nesstress i cant seem to select anything except ppu tests. in mario bros 2 i can run around, jump and pause. wierd stuff

read:
shift = (pad&0x01);
pad>>=1;
return shift;

write:
if(data&0x01) update_padstate();

 

[hap]

Perhaps the bug's not emulation related ?

INES Demo (PD).nes has got a joypad tester, though joypad 2 select and start don't work here with that.
*edit* you need to press the select key in nesstress to do other tests, that's why it won't let you do any other than ppu

 

[smcd]

Also i think it's ines (or tuxnes?) that comes with "cart.nes" that has a joypad test and stuff in it.

 

[aprentice]

i got the select key working in nestest, but its giving me borked info, it keeps saying "42" for every single test which makes no sense whatso ever. nesstress doesnt want to accept my select key, but in the ppu tests this is what i got, dunno what some of these errors even mean

 

[hap]

test 1 gives me all Oks here (and DPjr for the last 1, whatever that means), test 2 all errors (this is probably accurate), test 3 an Ok.

your 2nd error is probably due to memmap mirroring: 0xc000-0xffff=0x8000-0xbfff=0x4000-0x7fff=0x0-0x3fff.
you've implemented scrolling ? if you haven't yet, i think the other errors are due to that.

 

[aprentice]

some sprite work done and fixed a few cpu bugs, some progress
sprite attribute is kinda off though, but still looks cool

*edit* fixed sprite colors, was pointing at the wrong palette before

 

[ShizZy]

Looks great ren. Now if only I made that much progress with my GB emu

 

[hap]

*edit* evil smiley, automatically went to lower case

 

[aprentice]

*edit* evil smiley, automatically went to lower case

ack, looks like you got pokemon running

 

[hap]

I'm kind of on a break again, last thing I added was support for DDraw. Any progress on your side aprentice ?

 

[|\/|-a-\/]

Hi,
my NES emu is fetching the first instructions and emulates them correctly. pacman.nes lands in an infinite loop, which it does with other emus, too. i think the problem is that the PPU isn't emulated completly (means no gfx and the emu doesn't react on the ppu reg's flags).
I have problems continuing my work because i haven't any idea how to emulate the graphics. What is the background? The content of the nametables?

 

[hap]

nestech covers that very well: http://web.textfiles.com/games/nestech.txt
if you don't understand: reread, reread, reread, reread.. that's basically my approach to every complex technical document

 

[aprentice]

I'm kind of on a break again, last thing I added was support for DDraw. Any progress on your side aprentice ?

ive been on break too, but i've got some progress since last time, double dragon 2 works aside from lack of scrolling and a lot of other games work now too, i might resume it sometime soon. i think theres still bugs in my emulation somewhere causing glitches here and there

 

[|\/|-a-\/]

i read a lot of things the last days and now i unterstood a lot after rereading rereading rereading

Question1: pacman is landing in its infinite loop as already mentioned. the first thing it does is clearing the ppu regs ($2000, $2001, $2002). How will it escape from the infinite loop because the NMI on VBlank flag isn't set?

Question2: mirroring - a few docs are talking about flags A10 and A11 to set mirroring. A10 and A11 are pins of the ppu chip as far i know. Can the mirroring direction be changed? i use the mirroring dir bit in rom control byte 1, but Patrick Diskin wrote in his nes doc something about "single screen mirroring" ?!

 

[hap]

1: it probably exits the loop after 2002.7 gets set at vblank
2: yeah, some mappers can change mirroring 'direction' while a game is running, usually vertical/horizontal. single-screen mirroring means that all 4 nametables point to the same memory region.