Homebrew blargSnes -- SNES emulator for the 3DS (WIP)

[bobmcjr]

I know people are asking for a lot since this is still early but will super mario world romhacks be supposed soon? There are so many of those.

Depends on the romhack. Obviously the ones that use the SA-1 aren't supported, but the couple of dumb mods I've tried work flawlessly (including the ones that are expanded to 3MB).

 

[Xenon Hacks]

Is there any reason I cant save in donkey kong country (Great work btw StapleButter)

 

[weatMod]

reporting super ghouls and ghosts not working game starts intro mode 7 works , 1st level cant see anything but it moves and there is sound
i don't think this game used any special chip but i could be wrong

 

[loco365]

Has anyone tried using the SNES test program yet?

I make a point of running this official service testing program on all kinds of SNES emulators, and blargSnes 1.2 trips up on a lot of things.

1. Electronics Test
Fails - Doesn't even display the usual "FAILED ELECTRONICS TEST" message, instead emulator debugger returns "UNSUPPORTED VRAM MODE 84"

2. Character Test
Most things work, but there are failures.
- After Star sprite rows, debug message "!! SMALL SPRITES"
- Does not display the two halves of Service Mario graphic
- SMB3 promo art test fails to do the mosaic pixelation effect
- Cannot display the "Look for stars" portion at all, or the multiple Service Marios frame.
- Mario rotating test is slow (performance issue)
- Unable to flip the SMB3 promo art, despite being able to flip the Princess Peach sprite earlier

3. Controller Test
Each button press is responded to with a sound effect, but there are no visuals.

4. Sound Test
Sound output is pretty clear, but it's monaural, including speaker or headphone jack output.

5. Color Test
No issues at all, pixel-perfect on 3DS display.
There is some overscan that bleeds out of the top and bottom of the boarder.

6. Accessory Test
Detects that no super scope is inserted, and that no mouse is inserted. Menus function fine.

There's another one that I have, and when I get the chance, I'll run it and see if it works. Previously, it didn't boot, so I'll try this new version on both Ninjhax and as a CIA and see how it goes.

Edit: The Burn-In Test rom still fails to boot, and Final Fantasy III completely locks up blargSnes, so heads up on that. Final Fantasy II runs decently, it no longer sounds like a crashing airplane, although cutins and cutouts seem to not display properly.

Edit 2: F-Zero also fails to boot, and locks up blargSnes.

Edit 3: Undake 30 Same Game Daisakusen - Mario Version still crashes when you press Start. Nothing is dumped, it just seems to enter some kind of endless loop that locks up the emulator.

Edit 4: Sutte Hakkun also still crashes when selecting Password or Start Game from the main screen.

 

[Strife89]

Is there an official place to report which games do and do not work? I briefly skimmed the thread and the website, but found nothing.

I tested Chrono Trigger (U), Super Mario All Stars (E), and Mega Man X (U). All three caused the emulator (v 1.2) to hang on the file picker.

Super Mario World (U) runs, natch.

Tested with blargSNES v 1.2.
Launch method: Ninjhax/Homebrew Launcher

 

[piratesephiroth]

These are the classic games I use for tests that don't even show anything when run:

Axelay
Metal Warriors
Donkey Kong Country 2
Rendering Ranger R2

 

[piratesephiroth]

I tested Chrono Trigger (U), Super Mario All Stars (E), and Mega Man X (U). All three caused the emulator (v 1.2) to hang on the file picker.

Tested with blargSNES v 1.2.
Launch method: Ninjhax/Homebrew Launcher

Megaman X (U) (v1.1) runs fine (at least on the CIA version, but they should have the same compatibility I think)

 

[sj33]

Is there an official place to report which games do and do not work? I briefly skimmed the thread and the website, but found nothing.

I tested Chrono Trigger (U), Super Mario All Stars (E), and Mega Man X (U). All three caused the emulator (v 1.2) to hang on the file picker.

Super Mario World (U) runs, natch.

Tested with blargSNES v 1.2.
Launch method: Ninjhax/Homebrew Launcher

Super Mario All-Stars + Super Mario World (U) works for me, so maybe try that.

I have the same issue with Chrono Trigger, however.

 

[9thSage]

I've tried mostly Soul Blazer, Breath of Fire II (the retranslation project...the streaming audio in the intro works surprisingly well on the emulator), and Terranigma. All of them work pretty well! One game I found that didn't is Tenchi Muyo. I'd wager it's a sound bug. The first screen you see is the Banpresto logo, and it's meant to say Banpresto. It's not playing the sound and just hanging there. This game seems really sensitive to bugs in sound emulation...back in the days of snesDS the game was super unstable there because of it (sometimes getting past the Banpresto screen, then crashing later when some other sound refused to play).

I'm sure it'll work a lot better in blargSNES eventually.

 

[dronesplitter]

I came to report that v1.2 is a big leap in progress. Mega Man X now runs the text boxes as you can see and I don't really experience any slowdown for MMX now, which is awesome.

Super Metroid runs so well now and I've been able to view the ending with no glitches (Tourian actually runs full speed even when stuff is blowing up, crazy)

And Super Caltevania IV is also running almost perfectly other than some spots that slow the action down. Level 9 in Super Castlevania IV is really hard on the emu, as it has this animation of ghosts flying on the screen which the emu doesn't seem to like at all. But it does run full speed for the room that rotates around you and the area right after it where it looks like a tunnel effect with the background rolling.

 

[the_randomizer]

I've tried mostly Soul Blazer, Breath of Fire II (the retranslation project...the streaming audio in the intro works surprisingly well on the emulator), and Terranigma. All of them work pretty well! One game I found that didn't is Tenchi Muyo. I'd wager it's a sound bug. The first screen you see is the Banpresto logo, and it's meant to say Banpresto. It's not playing the sound and just hanging there. This game seems really sensitive to bugs in sound emulation...back in the days of snesDS the game was super unstable there because of it (sometimes getting past the Banpresto screen, then crashing later when some other sound refused to play).

I'm sure it'll work a lot better in blargSNES eventually.

Dang, Breath of Fire 2 Retrans working is great news

 

[DiscostewSM]

I came to report that v1.2 is a big leap in progress. Mega Man X now runs the text boxes as you can see and I don't really experience any slowdown for MMX now, which is awesome.

It does display correctly in that spot, but the first instance of it with Zero has the black box extending all the way to the right. If I understand correctly, windowing is split into sections for each row, and in this particular case, it's supposed to be 2 clear sections surrounding 1 black section. Instead, it's just 1 clear section on the left of the box, and the remaining is set with the black box. It is evident of that when reducing the number of vertices to 3 to make 1 tri polygon (rather than 6 to make 2 that form a rectangular box for the section).

 

[davhuit]

So, I was right after all Gateway users have sound with the cia file

About the compatiblity, except if the in-game log say to report it to StappleButter, you better wait before reporting crashs. I'm not sure but I would say that reporting graphical glitches, missing gfx layers might be more interested.

Lots of games will keep crashing until most of the specifications of the SNES are emulated (and I think some games can crash before a desynch of the spc/sound).

 

[reprep]

Is there any reason I cant save in donkey kong country (Great work btw StapleButter)

You can't save till you reach a save point in DKC. and that is pretty far in the beginning.

 

[c4655362]

So I downloaded some roms for random testing, I recorded all my impressions but when I tried to run Pokemon Stadium I got an error with a bunch info and something telling me to tell StapleButter
Here's some of the stuff I assume blargSNES left on my SD card after it

 

[the_randomizer]

So I downloaded some roms for random testing, I recorded all my impressions but when I tried to run Pokemon Stadium I got an error with a bunch info and something telling me to tell StapleButter
Here's some of the stuff I assume blargSNES left on my SD card after it

Pokemon on the Snes?? Wrong thread?

 

[c4655362]

Pokemon on the Snes?? Wrong thread?

Bootleg game
https://encrypted.google.com/search?hl=en&q=pokemon stadium snes
I don't remember if it was good or not, but it was a popular SNES game from where I got it so I figured why not

 

[sj33]

Would it be worth making a separate thread for game compatibility issues? It seems like it would be easier to maintain without flooding the main thread.

 

[gamesquest1]

So I downloaded some roms for random testing, I recorded all my impressions but when I tried to run Pokemon Stadium I got an error with a bunch info and something telling me to tell StapleButter
Here's some of the stuff I assume blargSNES left on my SD card after it

Pirate and bootleg roms are often quite fickle, they usually have some sort of copy protection built in, or are just a hacky code, I think stuff like that may work in future, but i doubt crappy bootleg games are high up on compatibility fixing.........your not missing out on much, most of the downloads will most likely be people getting it to see what it's like and never playing it again

 

[DiscostewSM]

Just did another test involving windowing, and it seems the problematic areas I mentioned in prior posts revolve around non-single-window cases. when I forced the PPU_ComputeSingleWindow function to ignore the "if" case, the window effect shown by @dronesplitter in the MMX screenshot was not there, but the problematic windowing I mentioned was still there.

edit:

My mistake. It wasn't so much the non-single-window cases, but that it defaulted to the non-single-window case when it shouldn't have. I made a few changes in ppu.c that corrected this (which fixes both hardware and software rendering of the effect). Forgive me, StapleButter, if you were planning to use the WinLogic array for something else, as I ended up using it for this purpose (because I didn't see anything else using it).

(in PPU_Write8)

case 0x23:
PPU.BG[0].WindowMask = val & 0x0F;
PPU.BG[1].WindowMask = val >> 4;
PPU.WinLogic[0] = (val?1:0);
break;
case 0x24:
PPU.BG[2].WindowMask = val & 0x0F;
PPU.BG[3].WindowMask = val >> 4;
PPU.WinLogic[1] = (val?1:0);
break;

(in PPU_ComputeWindows)

PPU_WindowSegment* first_s = s;
u8 Win12 = PPU.WinLogic[0] & (PPU.WinX[0]<PPU.WinX[1]);
u8 Win34 = PPU.WinLogic[1] & (PPU.WinX[2]<PPU.WinX[3]);

// check for cases that would disable windows fully
if (((!((PPU.MainScreen|PPU.SubScreen) & 0x1F00)) &&
(((PPU.ColorMath1 & 0x30) == 0x00) || ((PPU.ColorMath1 & 0x30) == 0x30))) ||
(!Win12 && !Win34))
{
s->EndOffset = 256;
s->WindowMask = 0x0F;
s->ColorMath = 0x10;
return;
}

// first, check single-window cases

if (!Win34)
{
PPU_ComputeSingleWindow(s, PPU.WinX[0], PPU.WinX[1], WINMASK_1);
}
else if (!Win12)
{
PPU_ComputeSingleWindow(s, PPU.WinX[2], PPU.WinX[3], WINMASK_2);
}
else
{
// okay, we have two windows
if (PPU.WinX[0] < PPU.WinX[2])