Research & Development Quick Research/Development Thread

[Midnight Bear]

Okay, I know this isn't exactly what this thread is for, but whatever.

What I've been working on for the past few hours is a rewrite of the stone evolution checking routine. It, essentially, makes it much easier to add custom (stone) evolution methods. If this is unclear, essentially you could write a code to check just for gender (or really, whatever you wanted), put it somewhere, and put that pointer into a table, and it would allow you to evolve using a particular stone on a particular gender of Pokemon. It uses the evolution types above 'F' (as in, all the unused ones).

Right now, I'm doing stone evolutions, though I plan to work on doing the same for levelup and trade evolutions eventually.

What I was wondering, though, is - would it even be worth it to even release it once I finish working out the bugs? Sure, it makes it easier to implement custom evolution methods, but you still need to know ASM to even use it, and, so, the only people that would be able to even use it could do the exact same thing.

[tajaros]

Ok, I found this over at the pokahacking forum and it seems to be useful for legendary events! So I did not discovered this credits to Steven and Casta for some Info about this.

So ever wanted to make your game like a movie or something like that. xD Well you can you can now make something look like this.



So how to do it huh? Pretty easy just put this in your script.

Ruby:

writebytetooffset 0x6C 0x4000044
writebytetooffset 0x23 0x4000045

Fire red:

writebytetooffset 0x80 0x4000044
writebytetooffset 0x20 0x4000045

Emerald:

writebytetooffset 0x7F 0x4000044
writebytetooffset 0x20 0x4000045

Just include them in every script that you want it only goes away after a warp to another place. That's all. :D

Credits:
Steven
Casta

[PokemonMasters]

You can also use this for trailers and stuff. :D

[tajaros]

Originally Posted by PokemonMasters

You can also use this for trailers and stuff. :D

Yes and cutscenes as well including flashbacks and it would be awesome! :D

[PokemonMasters]

btw, does this have bugs?

[tajaros]

Originally Posted by PokemonMasters

btw, does this have bugs?

As far as I know none. It only goes away after you warp.

[Team Fail]

So today I discovered how to convert MIDI files to .SSEQ using official SDK tools. This was a test for modulation, however it seems the music file is broken. I presume it may be from when the music file was exported, though.



Attached is the MIDI, the text version of it, and the resulting SSEQ from the video above. Just like GBA Music hacking, ladies and gentlemen.

I also ripped this from the executable to convert this wall of text into the SSEQ. IT'S ALL THE COMMANDS! :D

Edit: After about a day of work, I made my own totally custom .SDAT.

[shiny quagsire]

Originally Posted by tajaros

Ok, I found this over at the pokahacking forum and it seems to be useful for legendary events! So I did not discovered this credits to Steven and Casta for some Info about this.

So ever wanted to make your game like a movie or something like that. xD Well you can you can now make something look like this.



...

I actually found the other half of this that changes things to greyscale without doing any division or other crazy ASM. It's one RAM location:

0x02036E28

So just wrote a 0x1 to that offset for greyscale, and 0x2 for sepia (I think...). To reset, set it to 0x0.

[Darthatron]

hinkage messaged me about possibly removing the introduction part of the New Game function in FireRed. It was surprisingly easy. Pretty much entailed nulling some function calls and replacing one pointer.

0812ED1E: 00 00 00 00
0812ED2E: 00 00 00 00
0812EE3E: 00 00 00 00
0812EE48: 00 00 00 00
0812EE4E: 00 00 00 00
0812EE54: 00 00 00 00
0812EE60: 00 00 00 00
0812EE8C: D1 08 13 08

There's a small delay, though. I'm currently unsure about how to fix it.

EDIT: Nevermind. It only works if there is already a save file, for some reason.

[Iacobus]

Originally Posted by Darthatron

EDIT: Nevermind. It only works if there is already a save file, for some reason.

You have to write a function that writes a temporary player and rival name to the right location. Then it will work, trust me The game tries to load those.

[InMooseWeTrust]

Is there a way to hack an unused ability (like Cacophony) into Snow Warning? Drought, Drizzle and Sand Stream are already in 3rd gen but I'd like to have the option in Fire Red for a Hail team in the same way.

[Agastya]

em swarms @0x58D094
species - 2
movepool - 8
level - 1 (i think)
map - 1 (i think, based on first map bank)

assuming this is correct, these are the swarms in this data block
Seedot Lv3, Route 102: Bide, Harden, Leech Seed
Nuzleaf Lv15, Route 115: Harden, Growth, Nature Power, Leech Seed
Seedot Lv13, Route 117: Harden, Growth, Nature Power, Leech Seed
Seedot Lv25, Route 120: Giga Drain, Frustration, Solarbeam, Leech Seed
Skitty Lv13, Route 116: Growl, Tackle, Tail Whip, Attract

they primarily have egg moves and they're primarily just seedot so idk
they really dont seem that special and theyre completely off from what other sites say so maybe theres another block of swarm data somewhere

[shiny quagsire]

The address at 0x030008c9 is a step counter that increases by three for each step. Just something I found while searching the RAM.

[SSJ4 Furanki]

Is there a way to delete your rival from the naming process at the beginning of the game in FireRed?

[Quickster]

Originally Posted by SSJ4 Furanki

Is there a way to delete your rival from the naming process at the beginning of the game in FireRed?

Hey, I figured Id link you to this page.

http://www.thepkmnworld.co.uk/v5/for...ce-centre.396/

If you look at the naming rival from overworld section, Im sure you could learn something as Jambo removes it, Im pretty sure at least.

[robin22gongon]

Originally Posted by Quickster

Hey, I figured Id link you to this page.

http://www.thepkmnworld.co.uk/v5/for...ce-centre.396/

If you look at the naming rival from overworld section, Im sure you could learn something as Jambo removes it, Im pretty sure at least.

Uh... Could you just copy the thing written on your link and paste it here cause that site isn't available in my country?

[tajaros]

Originally Posted by robin22gongon

Uh... Could you just copy the thing written on your link and paste it here cause that site isn't available in my country?

Here..

Naming the Rival from the Overworld:

Spoiler:

The Code:

.text
.align 2
.thumb
.thumb_func
.global rivalnamingingame
main:
push {r0-r4,lr}
ldr r0, place
str r0, [sp, #0x4]
ldr r1, ramlocation
ldr r1, [r1, #0x0]
ldr r0, standard
add r1, r1, r0
mov r0, #0x4
mov r2, #0x0
mov r3, #0x0
bl place2
ldr r1, ramlocation
ldr r1, [r1, #0x0]
ldr r0, standard
add r1, r1, r0
ldrb r0, [r1, #0x0]
cmp r0, #0xFF
beq failsafe
cmp r0, #0x0
beq failsafe
return: pop {r0-r4}
pop {pc}
place2: ldr r4, actualroutine
bx r4
failsafe: ldr r0, rivalname
loop: ldrb r2, [r0, #0x0]
strb r2, [r1, #0x0]
cmp r2, #0xFF
beq return
add r0, #0x1
add r1, #0x1
b loop
.align
place: .word 0x080568E1
ramlocation: .word 0x03005008
standard: .word 0x00003A4C
actualroutine: .word 0x0809D955
rivalname: .word 0x08FFFFFF

Where to insert: Anywhere! This routine is called from a script in the game, and as such, doesn't need to be carefully linked into existing routines. All you need to do is remember the rule of THUMB, add 1 to the offset when calling it!

Explanation: It's what it says on the tin, it lets you name your rival from the in game overworld, allowing for Silver like naming, or indeed, any other possibility you can come up with. The rivalname: .word 0x08FFFFFF is for if the player leaves the name box blank. It will fill it with a standard name instead. This standard name is up to you to create. Simply compile a name with XSE, and point that pointer to it.

You also need to put this routine in the game, or else it won't be able to load past the intro without naming the rival. All it does is temporarily write a holding name to the rival buffer so that the game doesn't crash.

.text
.align 2
.thumb
.thumb_func
.global rivalnamingskip
main:
ldr r1, [r6, #0x0]
ldr r0, number
mov r8, r0
add r1, r8
push {r1}
ldr r0, rivalname
loop: ldrb r2, [r0, #0x0]
cmp r2, #0xFF
beq end
strb r2, [r1, #0x0]
add r0, #0x1
add r1, #0x1
b loop
end: strb r2, [r1, #0x0]
pop {r1}
mov r0, sp
ldr r3, return
bx r3
.align
number: .word 0x00003A4C
rivalname: .word 0x08FFFFFF
return: .word 0x08054A75

Where to insert: Navigate to 0x54A6A, and change the following bytes to this:
01 49 08 47 00 00 XX XX XX 08.
Where the XX's stand for the pointer to the new routine plus 1.

Explanation: It simply writes a placeholder name into the [rival] buffer so the game doesn't crash. Obviously, you shouldn't be using the [rival] buffer until the player has named the rival, so it shouldn't be a problem.

To make this work, navigate to 0x103690, and change the pointer there to 0x1307D1.

[wwmb]

Originally Posted by Midnight Bear

Okay, I know this isn't exactly what this thread is for, but whatever.

What I've been working on for the past few hours is a rewrite of the stone evolution checking routine. It, essentially, makes it much easier to add custom (stone) evolution methods. If this is unclear, essentially you could write a code to check just for gender (or really, whatever you wanted), put it somewhere, and put that pointer into a table, and it would allow you to evolve using a particular stone on a particular gender of Pokemon. It uses the evolution types above 'F' (as in, all the unused ones).

Right now, I'm doing stone evolutions, though I plan to work on doing the same for levelup and trade evolutions eventually.

What I was wondering, though, is - would it even be worth it to even release it once I finish working out the bugs? Sure, it makes it easier to implement custom evolution methods, but you still need to know ASM to even use it, and, so, the only people that would be able to even use it could do the exact same thing.

I think that you should. And just because someone can do something... doesn't mean they would want to go through all the work of designing such a system, ironing out bugs, building it and so on.

[shiny quagsire]

Originally Posted by tajaros

To make this work, navigate to 0x103690, and change the pointer there to 0x1307D1.

Actually, you wrote that down wrong. it's 0x130690, not 0x103690.

[tajaros]

Originally Posted by shiny quagsire

Actually, you wrote that down wrong. it's 0x130690, not 0x103690.

It's like that in Jambo's post so I'm not sure if that is the pointer.

[interdpth]

Are camera addresses known? If so is there already hack that presents the 3D Pokemon games in First Person mode?

[wwmb]

POKEMON EMERALD
I found something kind of wacky when I was messing with strings:

//---------------
#org @start
msgbox @msg 0x2
end

//---------
// Strings
//---------
#org @msg
= \hF9

This seems to:
Repeat last speech from last triggered script/person
It is also, always, missing the first 2 characters of the repeated text. (ie, talking to the person who says "PROF. BIRCH spends days in his LAB [blablabla]" will repeat "OF. BIRCH spends days in his LAB [blablabla]")


Probably useless. but still funny.

[FBI agent]

Originally Posted by Jambo51

So, last night, I worked out how type effectiveness is done in FireRed. You'll be pleased to know that it's relatively simple. Firstly, the method of storage:

[Attack type - 1 byte][Target Type - 1 Byte][Effectiveness value - 1 Byte]

This data is stored at 0x24F050.

There are several of these, kept in an array which ends with a 0xFE value. The game loops over each entry in the array, checking if the attack type matches the used attack's type, and if it does, then checks the Target Type against the 2 types of the targeted Pokémon.

If either type matches, it loads the effectiveness byte and branches off to another routine which takes care of physically making the attack super effective/not effective. The type effectiveness values are as follows:

Note, these are in hex
00 - Has No effect (Eg, Ghost on Normal)
05 - 50% effectiveness
14 - 200% Effectiveness

I don't know why they used these values, but that's what they are nonetheless. What you may gleam from these values, however, is that, unless explicitly told otherwise, the game treats all moves as normal effectiveness on any type.

If neither type matches, it jumps to the loop counter, increments it and continues checking until it hits that magical 0xFE value.

However, the way it is implemented leaves something to be desired as it is very slow (looping over several entries), and taking up more space than needed, both in terms of code and data. It's obviously wasteful in so far as it continues to check even after having already found both effectiveness values, as well as the wasteful nature of the loop in and of itself.

Related thread:
http://www.pokecommunity.com/showthread.php?t=83674

[Chaos Rush]

Here are some offsets for the Pokedex HUD:
0x451F55 = X position of text list
0x451F56 = Y position of text list
0x451F5D = X position of Habitat picture
0x451F5E = Y position of Habitat picture
0x451F65 = X position of "Seen:__ Owned:__" box
0x451F66 = Y position of "Seen:__ Owned:__" box
0x451F61 = BG Palette number that habitat picture uses
0x4520C8 = # of list entries on main pokedex page (National Pokedex enabled only)

Mess around with the bytes at those offsets, allowing you to do stuff like this:

[karatekid552]

Just some research on high #flags.

I have always used flags in the 1000's for my hacking, until tajaros showed me this post: http://www.pokecommunity.com/showthr...56#post6829256

In it, it states that flags above 900 overwrite variables from 4000 up. I have never had a problem with flags in the 1000's so I did some investigating.

I wrote a script that would set all of the flags from 900 to 11FF and then did some variable comparing. I checked variables 4000 to 4380 (the var that flag 1000 should affect, if this was true, since flags are 1 byte while vars are 2) and I found that only vars 4000-408F were affected. Nothing higher. I even checked a few variables like 5000, 4400, and 4500, all to receive that same result. None higher than 408F were affected.

This brings me to my question. Why? I know that flags 900 to A22 are completely off limits by the above research, but what about the rest? Where are they stored? Either vars 4090+ are stored in a non-DMA location or the flags above A22 aren't stored with the rest. Does anyone know anything about these high number flags?