ComboVid.com

I don't know a lot about Soul Calibur IV, but when I was playing around with Yoshimitsu in traning mode yesterday, I fount this. Is this an infinite? I used forward+A+K during flea to ground Nightmare, then A+B during Flea to keep hitting him. http://s230.photobucket.com/albums/ee20 ... V00010.flv

The combo meter was att 99 hits. I also got it to do over 1000 dmg when I activated it with an attack counter (Which appearently made the damage scaling stop at 79% instead of 17% or whatever it is in that recording)

^I'm pretty sure that if the numbers are blue, it means it's not techable. Looks like an infinite to me, but you might want to double check at 8wayrun.

Bah, it's incredibley hard to set up anyways, because you have to be spaced very precisely sideways for him to not slide out of it.

There's no code to pick Huitzil or Pyron in the original Darkstalkers?

I'm not aware of any unlock codes but if you use the cheat codes, it's interesting that 1) the game doesn't crash and 2) they have their specials working. Shows you there was never any doubt they would be playable.

It freezes up if pyron wins but they gave huitzil a partial ending:


AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

thats kinda like what they did with the bosses in SFEX.
SFEX (no plus, no alpha) has bison and garuda as unplayable bosses (I think, I can't translate the all about SFEX book so maybe they are, not the all about book didn't even give command lists for them, just frame data). But using mame/zinc code you can play as them.

EX2+, same with Bison 2
EX3, same with Shin Bison.

Except Cycloid Beta code for EX3 and maybe EX2+ since he doesn't have any specials and virtually none of his moves hit.

Maj wrote:12) In SFEX2 if you watch the very first attract/intro sequence, Guile does j.HP, c.MP xx Sonic Boom xx Guard Break xx Somersault Justice against Vega. Is that legit? I know that canceling Sonic Boom into Guard Break works as a combo, but there's definitely not enough time to recharge afterwards, cuz Vega is supposed to fall immediately.

thanks to framemame I've almost got this to combo lol.

ur.ur.ur.ur. _d_l. w25. 3. w22. 2. . ^d^l r. _d_l w9. 3. w10. 1 4. w38. ^d^l dr.dl.ur6. w60!

prior to releasing the down and up for the 2nd time, if you change the wait from 38 to something lower guile will perform his super sonic boom. However it only combos from 24-34
so
soonest that the super sonic boom can come out (from what I've tried)
ur.ur.ur.ur. _d_l. w25. 3. w22. 2. . ^d^l r. _d_l w9. 3. w10. 1 4. w24. ^d^l dr.dl.ur6.w60!
latest that it'll still combo
ur.ur.ur.ur. _d_l. w25. 3. w22. 2. . ^d^l r. _d_l w9. 3. w10. 1 4. w34. ^d^l dr.dl.ur6.w60!
soonest that the flashkick super comes out
ur.ur.ur.ur. _d_l. w25. 3. w22. 2. . ^d^l r. _d_l w9. 3. w10. 1 4. w38. ^d^l dr.dl.ur6. w60!

http://www.youtube.com/watch?v=J9MwfK2UGek

How do you get the start colors in 3S? (arcade ver?)
I read somewhere you have to finish singleplayer with the character first. Doesn't seem to work.

All the Start colors are console-only. The arcade version only has the 6 button colors plus the 7th hidden LP+MK+HP color.

Dammit wrote:How do you get the start colors in 3S? (arcade ver?)
I read somewhere you have to finish singleplayer with the character first. Doesn't seem to work.

ugh, I've been wondering this myself

edit: there's this thread on SRK about making custom colours. Haven't tried it out yet, might be useful tho.

can anyone tell me about input delay? A guy asked about it on the tas forums, convinced that there shouldn't be any and I don't see how it could be avoided. Figure someone here might know something about it.

Sorry, this isn't really a TAS-related question, but I was just curious...
Does one exist? To be specific, PCSX-RR with all the default plugins seems to add 2 extra frames of delay (same results with the keyboard and my arcade stick w/ USB converter), and sadly that's the best I've had any emulator perform thus far.
I'd like to stream/record a single-segment run of Street Fighter EX's Expert Mode, but you just can't afford to have any delay with some of the combos there!
Thanks in advance. =)

errror1
why do you think the input delay is in the emulator and the game?

It's extra delay that shouldn't be in the game.

For example, Ryu's medium kick takes 9 frames to hit after pressing the button, but:

1. Published data says it should only take 7 frames.

2. The animation of the move is, in fact, 7 frames.

Hence my point that the emulator is adding 2 extra frames of delay.

That is actually quite a bit for fighting games, which is why I'm trying to find some sort of alternative. No luck yet, though. =/

If you pause the game, then hold down LP, then advance one frame, then Ryu will begin his standing LP when the game moves forward.

That means the game was at neutral and time-stopped. Time is still. You then decided the next frame from right now would have a LP for an action. And that frame did in fact have LP as an action once you pressed unpause. So: "Walking forward - Pause - Press a button - Unpause - Button/attack occurs"

An emulator with input delay means that action occurs later. So with the above example: you pause, then hold down LP, then advance one frame aaand nothing happens (on that frame or the next frame or the frame after). If you pause, hold LP, then advance three frames, then you begin the LP on the next frame.

Now, on a non-frame level what delay basically means is like if you had a 3 frame cancel window on Ryu c.MK. So let's say you want to cancel the c.MK, you look for when it hits. Now without delay you might possibly (not really but bare with me) be able to press the button when you see it hit. So with no delay, theoretically you see c.MK start, c.MK hits and you see that, then you press the button before that 3F window ends. Now with a 3F delay what happens is you see that c.MK hit, and so you cancel immediately afterwards, but with a 3F delay that cancel doesn't occur within the cancel window, you saw it hit and you pushed a button but that button doesn't do anything until three frames later.

So basically like netplay.. laggy.

In retrospect lets say I push c.MK on frame 1
The hit occurs on frame 4
I see it on frame 4
The cancel window is frame 5-8
I see it hit on frame 4 (theoretically) so the next four frames are reaction time, I can press the next attack within that time.

With delay I press attack on frame 1
It starts on frame 3
it hits on frame 7 cuz 4 frame startup
I see the attack occur on frame 7
I push a button on frame 8 (cuz im gdlk 1f reaction)
The attack occurs on frame 11 cuz of the 3f delay. Even tho I pushed the button when I visually saw the cancel window, doesn't matter, the input occurred after the cancel window because my actions occur later in time (3f later).

I understand the concept but I don't see how any game could run at 0 frame lag. It seems like it would be impractical for a game to process inputs that fast. Is there any way to see if a console game really runs with 0 lag and emulators add some to make it easier to compute or if input lag is universal.

Well actually, that's precisely why I laugh at framedata nerds that get confused about the first active frame occurring at the last frame of startup, or the frame after. Dohohohoho.

Also no it's not possible, you push a button, the action is detected after you push it, then the game reacts to that new information.

Button press -> game receives -> game acts

With delay, after the game act then you see what the game does and react. So if a HK hits on frame 3 you don't know about it until frame 6. If you see an attack hit on frame 5 and push a button of frame 6 then that button doesn't do anything till frame 9 and that frame ain't cancelable.

The published data could have been derived by observing the animation. i.e., They captured the video on a VCR, watched it frame by frame, saw that the hit happens 7 frames after he starts moving, so they write 7 in the book. The VCR never knew when the button was pressed so it was never reflected in the published frame data. The delay was there all along but only noticed when we have frame advance.

That's one theory. Another is the emu really is drawing each frame a couple frames too late. The newest Gens-rr has a "latency compensation" feature, where you get to choose how many frames ahead/behind you want to be shown. That this feature even exists implies that the emus can show you what's going to happen earlier than they do. (But still, the delay may be accurate behavior, idunno.)

I don't think so. I think it's within the same vein as TAS/TACV authors wanting it a specific way, because they are use to it.

It's 1 frame. One.

Hi framedata, hi being use to program pads, hi we count frames a certain way. Oh my god one frame difference in notation whaaaaaaaaaaaat /world ends.

It's written the way it is because the question isn't about how long it takes the move to come out after you press a button. The only important questions answered by frame data are: what links from what and what punishes what.

Input delay doesn't matter for either of those because if you're discussing theoretical possibilities, then all that matters is what works mathematically. Execution is a whole different issue because your reflexes and your timing will adapt subconsciously if you play the game enough.

I mean, at the end of the day, 2-frame input delay will carry over throughout an entire combo so once you start the combo the rest of it is normal timing. That's why you can use laggy converters with program pads without any problems as long as the converter isn't also dropping inputs.

As for whether or not input delay exists, i believe it always has because you'd have to give the hardware a little bit of time to process signals. We're just used to it and we've conditioned our reversal/tech/punish timing around it. But nobody talks about it because it probably varies from console to console and probably also from game to game. And it's difficult to test because emulators aren't always 100% precise either, so you'd have to test it on a physical arcade/console and nobody actually cares enough to go through all that trouble.

Nice! Good to know.

But honestly i'm not even sure if i trust those zero's to be accurate representations of the actual hardware. That's the kind of thing people "fix" even if it's not supposed to be fixed. I think maybe the best way to test it would be to hook up a light to the buttons on an arcade board and film with a high-speed camera as you press it, then replay it frame by frame to see if there was lag. I kinda remember somebody doing this to compare different versions of ST. Maybe it was NKI? This was a while ago.

Yeah that's how NKI did it, there's a video here: http://zachd.com/nki/NKI-Testing.CCC2.For.Input.Lag.rar

Maj wrote:Nice! Good to know.

But honestly i'm not even sure if i trust those zero's to be accurate representations of the actual hardware. That's the kind of thing people "fix" even if it's not supposed to be fixed. I think maybe the best way to test it would be to hook up a light to the buttons on an arcade board and film with a high-speed camera as you press it, then replay it frame by frame to see if there was lag. I kinda remember somebody doing this to compare different versions of ST. Maybe it was NKI? This was a while ago.

yeah, that was definitely NKI when he was trying to prove that AE added input lag compared to the DC and arcade versions.

Ihave a question.


Why with Akumas ultra can you LP . LP . -> + LK + HP but you can't LP . LP + -> + LK + HP

if that makes sense,i've always been curious about this one

I don't even think you're supposed to be able to do the first one. Most games force you to do each of the five inputs on separate frames. SF4 is just ghetto when it comes to command shortcuts and sometimes they don't make sense at all.

you cant do the first one.There must be a frame apart forward and LK + HP.

^^You Sure?^^

I think?Don't kill me please.

IS lol kill you?

I'm confused.

Was Samsho Tenka (SS6) ever emulated? I'm a bit too lazy to go research this, so if anyone happens to already know without having to research it'd be appreciated.

Have they emulated Atomiswave games?

There is this one: http://demul.emulation64.com/

I haven't looked into this stuff in ages tho.