Before the WIP - Oswald (KoF XI) - Place for Discussion, Opinions and Guidance

[Ashram McDuck]

I haven't been in the MUGEN community long and, really, the only thing I've actually produced with it is just a bunch of videos. In anycase, the character I've spent the most time with is ahuron's Oswald XI, which I am learning more and more is not that accurate to his actual game counterpart. I've been talking with a few people about possibly pulling a remake of this character. I've more or less decided that I'll be using (most of) ahuron's sprites setup as a base, mainly because they are all set and ready to go (I don't think there are any alignment issues) and I'll produce everything else.

Now, here's where the purpose of this thread comes into play - I'm the kind of person who can very easily lose motivation in something when I am unaware of the interest various people may or may not have in my personal projects. (in short, if it's not apart of my job, I've got to be aware of some sort of outward influence/interest)
What I want to know is this - is there any interest in this potential project?

I would like to know what people would be looking for if I decided to go through with this idea. I've got many things in mind, from implementing more of the variety of chaining (and unchaining) options that ahuron's Oswald lacks, reworking the CLSN boxes, and possibly including some of the abilities that came with the Skill Meter.

What I've done so far is nothing much ... I've made some palettes that reflect the actual game palettes and I've been experimenting with the actual game to see just how much reworking of Oswald's chains and such are in order. In closing, I'd like to share the palettes I've created. If anyone is interested in them, I can look into making them available on some file sharing service.



So, that's it. Keep in mind that I have never done MUGEN coding before. I've read tutorials and taken a look at many characters inside Fighter Factory but I haven't taken any steps beyond that. If anyone would like to collaborate on this project, that would make things even better.

LET ME KNOW WHAT YOU GUYS THINK.

-Ash

[Leon Belmont]

Oswald is not exactly the easiest character to code for a beginner, especially when you're looking to an accurate conversion.

It would help if you had KOFXI on PS2 and the PS2 emulator to go along with it, so you can have timings and everything. Then you could play it on your PS2 if the speed isn't convenient enough and taking note of his chains. Tel's FAQ of KOFXI IIRC also have lists of cancellable attacks and attack properties and stuff, and that would help.

Also, I guarantee Ahuron screwed up alignment somewhere. You're going to have to check those things in the actual game.

[Ashram McDuck]

Oswald is not exactly the easiest character to code for a beginner, especially when you're looking to an accurate conversion.

It would help if you had KOFXI on PS2 and the PS2 emulator to go along with it, so you can have timings and everything. Then you could play it on your PS2 if the speed isn't convenient enough and taking note of his chains. Tel's FAQ of KOFXI IIRC also have lists of cancellable attacks and attack properties and stuff, and that would help.

Also, I guarantee Ahuron screwed up alignment somewhere. You're going to have to check those things in the actual game.

This project's got a long list of things that are going to have to be checked, which is why I'm gauging how much interest there is. If I chose to go through with this, just about everything will need to be revamped. (minus the sprites, which are, more or less, ready to go - but you are correct in saying that I'll probably have to tweak the alignments ... but since I'd have to rework the CLSN boxes anyway, it will go alongside that)

I do have the actual game in my possession. It's not where my interest in the character began, but its certainly where my interest has taken me. (the game in and of itself is fantastic ... all except that bastard of a last boss)

[Vans]

As Leon said, Oswald is not a good option for a beginner due to the complexity of the cancels and juggle rules in general, you might wanna try your hand at something easier from KOFXI first.

In any case, I can confirm ahuron guesses all of his alignments and this may compromise possible accuracy with the character due to posadds and the like. If you are still willing to go at it, I can recommend PCSX2 0.9.2, its recording feature can record 60 FPS video that you can use to obtain timing.

Good luck.

[Bastard Mami]

He can "easily" be done the hard way, by using said 60fps record feature.

[Vans]

The feature does not keep the correct aspect ratio of the game, though.

[Ashram McDuck]

As Leon said, Oswald is not a good option for a beginner due to the complexity of the cancels and juggle rules in general, you might wanna try your hand at something easier from KOFXI first.

In any case, I can confirm ahuron guesses all of his alignments and this may compromise possible accuracy with the character due to posadds and the like. If you are still willing to go at it, I can recommend PCSX2 0.9.2, its recording feature can record 60 FPS video that you can use to obtain timing.

Good luck.

I had considered starting with another KoFXI character, but as it stands right now, Oswald is the only character I have interest in creating.

PCSX2 0.9.2, eh? I'll get my hands on that as soon as I can. Also, one quick question, "posadds" ... ?

He can "easily" be done the hard way, by using said 60fps record feature.

How, specifically, will this make the process "easier"? I've been told that I won't be able to get to the hit boxes in XI, so I'm guessing it's in concern to other things, such as velocity and, as L said, timing.

[Bastard Mami]

You can get clsn data, but it will be really painful.

[Vans]

PCSX2 0.9.2, eh? I'll get my hands on that as soon as I can. Also, one quick question, "posadds" ... ?

KOF sometimes offsets the character by a certain amount in each sprite of the current animation in contrast to velocities that offset the character in every animation frame.

This is recreated using the PosAdd controller.

[Bastard Mami]

I don't think he will get what you mean, since I had trouble understanding it as well, but yeah, that's a stepping move (or I actually forgot the actual term I was using before for those kind of moves).

[Cybaster]

How, specifically, will this make the process "easier"? I've been told that I won't be able to get to the hit boxes in XI, so I'm guessing it's in concern to other things, such as velocity and, as L said, timing.

Mugen runs at 60fps, as the games. Recording at 60fps, and going through the video frame by frame, allows you to count how many ticks the characters are in for each sprite, get the hitpauses, etc.  and if the video is at very good quality, to count pixels to get vels and other kind of stuff like that.

[Bastard Mami]

Can't count pixels using the video, because it will be in the wrong resolution but he still can count time and use physics formulas to figure out the rest; provided he uses otehr means to get correct resolution screenshots.

[Ashram McDuck]

KOF sometimes offsets the character by a certain amount in each sprite of the current animation in contrast to velocities that offset the character in every animation frame.

This is recreated using the PosAdd controller.

Let me seek to clarify this a little further. Let me use two Oswald move states to see if I understand what you mean. Is it ...

A) Oswald starts in stand anim. and performs his LK. His alignment doesn't move, but his left (from player's perspective) leg instantly comes to the right in a straight position while he performs the kick.

B) Oswald starts in stand anim. and performs his Heart (SK version) move, which ends up moving him a little backwards. The anim. itself doesn't move, but the PosAdd is what causes the movement backwards.

Mugen runs at 60fps, as the games. Recording at 60fps, and going through the video frame by frame, allows you to count how many ticks the characters are in for each sprite, get the hitpauses, etc.  and if the video is at very good quality, to count pixels to get vels and other kind of stuff like that.

I remember you first telling me about this a little bit. This additional information looks to make things a little simpler. (attemps at finding hitboxes as well I would imagine)

--------------------
I started a list of things that look in need of fixing by playing XI's Oswald alongside his Mugen counterpart. I haven't gotten to his jump attacks and crouch attacks just yet, but once I've collected all that info, (or as much as I can) I'll post it here and see if anyone else has any additions/subtractions.

[Ashram McDuck]

With this list I hope to gather a better understanding of Oswald as a coded, sprited character rather than just a character I've been fighting with in both MUGEN and KoF XI. I figure this little habit will help me along the way of restructuring Oswald ... and there's plenty of that to be had. Anyway, here's what I've found so far that is in need of revamping. (This is all from the sprite and animation set-up, as well as a few notes on the moves themselves ... I haven't gotten to a lot of the technical stuff since I'll be wanting  the emulator before tackling all of that) Any additions or subtractions would be appreciated and welcomed.

Spoiler (click to show)

Standing – Fix CLSN boxes (from 2 to 3) *Let's just apply this one to everything, more or less*
Walk forward – Animation is too fast, change to the same set-up as backward walk
Walk backward – Saw no problems
Crouch – Get rid of constant "breathing" animation, change to one "breath"
Jump – End parts of animation possibly wrong. (1 missed sprite?)
ADDITION – Short jump, (cybaster's velocities) Long jump, Regular jump (need velocities)
Dodge – Omit after image "trail" ; add backwards dodge
Taunts – All seem fine.

Light Punch – Make cancelable by continual LP input
NOTE: At closest range, only 4 consecutive hits before out of range.
Light Kick – Make cancelable by continual LK input
Strong Punch – Make follow up SP input a tad faster
Strong Kick – Seems fine.
Blow Out – Animation is wrong, speed is too fast, follow up BO is too fast
Crouch LP – Make cancelable by continual down + LP input
Crouch LK – Make cancelable by continual down + LK input
Crouch SP – Seems fine.
Crouch SK – Animation is slightly fast (?) Sound for second strike comes out too fast
SUBTRACTION – Crouch SK to 10 (DF + LP) chain
Jump LP – Something seems wrong here, but I don't know what ... HELP!
Jump LK – Seems fine.
Jump SP – Seems fine.
Jump SK – Hit boxes are definitely wrong.
Jump BO – Seems fine.
Close LK – Make two consecutive Close LKs possible
Close SP – Seems fine.
Close SK – Seems fine.

10 – Seems fine. (change chain options)
Jack – Animation is too fast
Spade (LP) – Attack should come out a little sooner when from a distance ; Additional sound ; Follow up Spade's speed is way too slow
Spade (SP) – Additional sound
Queen – Possible problem with Card animation
LP distance – Seems fine.
LK distance – Seems fine.
SP distance – Card's angle should land a little further right
SK distance – Card's angle should land a little further right (should strike lower leg/foot area of opponent on other side of screen)
King – Lots of problems. Will need help here.
Club (LK) – Velocity should be higher (alignment issue?) ; Range issue?
Club (SK) – Same problems
Diamond (LP) – Animation too fast during end parts ; Follow up Diamond's speed is way too slow
Diamond (SP) – Animation speed is closer to correct during end parts
Heart (LK) – Seems fine.
Heart (SK) – Steps back too far.
Ace – Lots of problems. Will need help here.

Four Suites – Wrong sounds (?) ; Check Timing of hits ; Check Animation
Royal Flush – Add "Jackpot!" quote ; ??????
Joker – Full quote before attack ; Wrong blocked Joker Animation

[Cyanide]

Not to rain on anyones parade but to a certain extent accuracy CAN be blurred. Getting "near as possible" settings will be unnoticeable to 99% of people and you certainly won't see it at all if you don't go into the code, find the values yourself and then compare.

Having a velocity of -4.11348 isn't really going to be THAT much different in the short run to a vel of -4. At worst, it will allow you to connect a different move that you can't combo into afterwards during training mode. Over about 20 moves and involving different starting positions.

Do it to the best of your abilities and you'll find that it's got perfectly good to the naked eye accuracy. Code and shit not so much but unless people go inside the character and pick at things they won't notice it. Playing him in game will give you a good feel on what yours is like in terms of accuracy as well.

[LaQuak]

Like I told you before Ashram the 7th and I won't tell you again but I think you need to make this character if you really truly enjoy playing with him and I know for a fact that you do because I seen your ass wooping rep sheet with him.

Countless open sources for you to seek if you stumble on issues and I am sure others can help you. Just don't start of thinking all negatively and seeking others to be like "Hey man nice WIP +1".

Do it and create it because you want to thats all man.

[Bastard Mami]

Not to rain on anyones parade but to a certain extent accuracy CAN be blurred. Getting "near as possible" settings will be unnoticeable to 99% of people and you certainly won't see it at all if you don't go into the code, find the values yourself and then compare.

Having a velocity of -4.11348 isn't really going to be THAT much different in the short run to a vel of -4. At worst, it will allow you to connect a different move that you can't combo into afterwards during training mode. Over about 20 moves and involving different starting positions.

Do it to the best of your abilities and you'll find that it's got perfectly good to the naked eye accuracy. Code and shit not so much but unless people go inside the character and pick at things they won't notice it. Playing him in game will give you a good feel on what yours is like in terms of accuracy as well.

except that getting the 4 using naked eye is about 10 times harder thn getting 4 using the simpler accuracy methods (unless you are actually very experienced in what you are doing).

[Cyanide]

Meh, not that hard to stand in front of someone and hit punch as many times as neccessary till you miss then put in a value that sounds right. Only takes about 3 tries. Did that just recently then compared to someone who'd used the actual value, i was .5 off.

[Ashram McDuck]

Not to rain on anyones parade but to a certain extent accuracy CAN be blurred. Getting "near as possible" settings will be unnoticeable to 99% of people and you certainly won't see it at all if you don't go into the code, find the values yourself and then compare.

Having a velocity of -4.11348 isn't really going to be THAT much different in the short run to a vel of -4. At worst, it will allow you to connect a different move that you can't combo into afterwards during training mode. Over about 20 moves and involving different starting positions.

Do it to the best of your abilities and you'll find that it's got perfectly good to the naked eye accuracy. Code and shit not so much but unless people go inside the character and pick at things they won't notice it. Playing him in game will give you a good feel on what yours is like in terms of accuracy as well.

This is basically what I want to shoot for. I know I'm probably not going to be able to get everything to exact source levels, but I would like to get everything to as close as accurate as I can. I agree with people that say ahuron's Oswald is not very accurate ... but I don't think his is so inaccurate it doesn't feel like the actual Oswald. Yes, there's a whole lot missing ... but a great deal of his essential parts are there and usable. So I'm going to take what was started and make it better.

That's the plan anyway.

(and who knows? I may add a little something to him and release a custom version alongside his "more accurate" version)

[Vans]

Meh, not that hard to stand in front of someone and hit punch as many times as neccessary till you miss then put in a value that sounds right. Only takes about 3 tries. Did that just recently then compared to someone who'd used the actual value, i was .5 off.

A waste of time considering you can get the actual values for lot less trouble. :|