• Welcome to Smashboards, the world's largest Super Smash Brothers community! Over 250,000 Smash Bros. fans from around the world have come to discuss these great games in over 19 million posts!

    You are currently viewing our boards as a visitor. Click here to sign up right now and start on your path in the Smash community!

metaknight framerate data

JCav

Smash Lord
Joined
Feb 2, 2008
Messages
1,217
Location
Michigan
let me just start this with with a quote:
if yours don't match mine then yours are wrong, make sure all yours match mine before proposing new ones
... i know this, i understand who m2k is and i have the upmost respect for him as a smash player. with that being said i must say I DID NOT CONTRADICT ANYTHING HE SAID i am not a fool but more importantly, my information matched his so no need to worry so without further ado here ya go.

and if i missed anything, anything at all you want tested LET ME KNOW

Jab (10% jab)
6 startup
7-11 (5) 1st hit
12-13 (2) frames in between hits
14-18 (5) 2nd hit
19-24 (6)
25-29 (5) 3rd hit
30-35 (6)
36-39 (4) 4th hit
40
41-45 (5) 5th hit
46-58 (13) cooldown
58 total

Down tilt
2 startup
3-7 (5) frames the hitbox is out (how many frames the move last)
8-17 (10) cooldown (number of frames in cooldown)
17 total

Forward tilt

2
3-8 (6) 1st hit
9-15 (7)
16-20 (5) 2nd hit
21-25 (5)
26-30 (5) 3rd hit
31-56 (26)
56 total

Up tilt

7
8-13 (6)
14-34 (21)
34 total

Forward smash

20
21-29 (9)
30-43 (14)
43 total

Down smash (front stroke only)

4
5-12 (8)
13-36 (24)
36 total

Down smash (hits in back only)

8
9-17 (8)
18-38 (21)
38 total

Up smash
7
8-12 (5) 1st hit
13-14
15-20 (6) 2nd hit
21-22
23-28 (7) 3rd hit
29-53
53 total

Up air

1
2-7 (6)
8-17 (10)
17 total

Forward air

5
6-10 (5) 1st hit
11-13 (3)
14-18 (5)
19-20 2nd hit
21-26 (6)
27-41 (15)
41 total

Back air
6
7-12 (6) 1st hit
13-16 (4)
17-22 (6) 2nd hit
23-27 (5)
28-32 (5) 3rd hit
33-49 (17)
49 total

Down air
3
4-8 (5)
9-26 (18)
26 total

Neutral air

2
3-6 (4) 1st hit
7-11 (5)
12-17 (6) 2nd hit
18-35 (18)

Dash attack

4
5-10 (6)
11-32 (22)
32 total

Tornado
9
10 hitbox starts
23 frames of cooldown

Dimensional cape

Invisible on frame 13
Attack 4 frames after re-appearance
10 frames of hitbox
41 frames of cooldown

Side b

21
22 hitbox starts
22 frames of cooldown when done into the ground
19 frames for falling animation (meaning you hit the ground on frame 1 and can do an action on frame 20)

Shuttle loop
4
5-13 (9)


Glide attack
4
5-13 (9)

Grab 6
Dash Grab 9


Missed dash grab 28 frames
Missed standing grab 25 frames


Spot dodge

Invincibility starts on frame 6
Last till frame 18
19-21 vulnerable

Roll
Invincibility starts on frame 7
Last till frame 11
27 total
 

K 2

Smash Lord
Joined
Jul 17, 2008
Messages
1,920
Location
Tennessee
Wow...MK's smashes are faster than my tilts...not fair! MK's tilts are faster than my jabs!
 

A2ZOMG

Smash Legend
Joined
Oct 13, 2007
Messages
12,542
Location
RPV, California
NNID
A2ZOMG
Switch FC
SW 8400 1713 9427
ZSS's Jab is faster than MK's U-air. ZERO startup, and hits on frame 1.
 

hotgarbage

Smash Lord
Joined
Oct 15, 2007
Messages
1,028
Location
PA
well first, i use a capture card and record every move

then i use a program called virtualdub to scroll frame by frame




i dont do it the way m2k does, sorry...
Ok, that's the same way I do it :p. I was wondering at the time because I mistook "startup" for "hits on frame ____" and was wondering how the heck you got such messed up data xP. I'm guessing you use vdub's own bob deinterlacer? I use avisynth myself for that, but it shouldn't make a difference.


pretty much doesnt exist, sh tripple u air makes me slightly angry
Landing lag for his aerials does exist if you land them before the auto-cancel point. He has 15 landing lag for improperly landing fair and dair.... I haven't checked the rest.



Spot dodge

Invincibility starts on frame 6
Last till frame 18
19-21 vulnerable

Roll
Invincibility starts on frame 7
Last till frame 11
27 total
I'm guessing you based the invincibility startup of his spotdodge and roll off his animation? Specifically when he "lights up"? Unfortunately you can't do that because the animation doesn't coincide with the actual invincibility frames :dizzy:. Also it sounds like you based the time the hitboxes were out on his animation as well. You may want to test that, as the animation doesn't always correspond with the actual hitbox. Though in MK's case I wouldn't be surprised if it did.



Anyways, I've been working on mk data myself, and some of my own stuff contradicts yours. Not that I'm saying you're wrong; it could be either of us. I'm going to double check some of the stuff and then post some gifs to show what I'm seeing. Oh yeah, and props for doing all this. Getting frame data is a royal pain xP.


Contradictions:

3GOD said:
Meta Knight:
Spot Dodge lasts 25 frames
Invincible Frames 2 – 20

Roll Backward lasts 33 frames
Invincible Frames 4 – 12

Roll Forward lasts 23 frames
Invincible Frames 5 – 12

Air Dodge lasts 39 frames
Invincible Frames 3 – 28
I got the same results for his spotdodge, so I'm pretty confident it's correct.


dtilt:
-15 total instead of 17

ftilt:
-2nd hit on 12 instead of 16
-3rd hit on 21 instead of 26
-52 total instead of 56

fsmash:
-hits on 24 instead of 21
-41 total instead of 43

dsmash
-34 total instead of 36
-2nd hit on 10 instead of 9

usmash:
-2nd hit on 12 instead of 15
-3rd hit on 17 instead of 23
-49 total instead of 53


I wonder what's going on here ?________?. I'll get to double checking this stuff, and once again I'll post some gifs to show what I'm seeing if that's needed.
I guess I should mention how I personally get all this. I capture at 30 fps via virtualdub. And then using virtualdubmod/avisynth I use the leakkernelbob deinterlacer to get 60fps. Next I look through the junk frame by frame.

It looks like this is going to be the official mk frame data thread, so I'll add some stuff to it later (advantage, shield stun, autocancel points).
 

JCav

Smash Lord
Joined
Feb 2, 2008
Messages
1,217
Location
Michigan
ok, if you want to do whatever with this data go ahead. you can totally disregard this thread and post your own. i dont care. and since you wanna do shield lag and whatnot, go ahead
 

hotgarbage

Smash Lord
Joined
Oct 15, 2007
Messages
1,028
Location
PA
Eh, I don't want to post my own if I don't know for sure that the data is correct. I'm more interested in figuring out why we're getting different results if we're using the same method. It's odd that our startup times agree (save for fsmash) while much of the other stuff doesn't. Are you deinterlacing to get 60fps? Are you getting any dropped frames?
 

hotgarbage

Smash Lord
Joined
Oct 15, 2007
Messages
1,028
Location
PA
what?


i set it to 60 fps if that helps...
Where did you set it to 60fps? (er, under what tab was the option to change it?)
What kind of capture card are you using?

Sorry for all the questions; I'm trying to figure out what the heck is going on here :p.
 

Tenki

Smash Hero
Joined
Apr 3, 2008
Messages
6,966
Location
GA
or you could do it in training mode at half speed, so if you record at 30fps you get everything...
I don't know if that's a good idea.

I don't know about 1/2 speed, but 1/4 speed's frames give you about 5/3 more frames than you would get if you recorded at normal speed. I've compared frames that people got from recording at 1/4 speed and normal and it's stayed a pretty constant factor in converting their 'wrong' data with the added frames to the 'correct' amount that you'd get if you recorded at normal speed.
 

hotgarbage

Smash Lord
Joined
Oct 15, 2007
Messages
1,028
Location
PA
ive done this too and people may say its not, but it is accurate
I've messed with this, and for the most part it is accurate. Startup times in particular seem to be correct when doing this. But the fact is that the video is still interlaced at 30fps, which is going to cause problems such as this:




Looking at this frame by frame you can tell that his dtilt comes out on frame 3 and that his shield comes up on frame 17.... meaning that his dtilt lasts 16 frames.


But take a closer look at frame 16:


On frame 16 he is in the middle of his attack AND in neutral position! So which is it??? This is why to get accurate data you really need to bob deinterlace to get 60fps progressive. In addition, I found that training mode was inconsistent... often times giving different results for the same move. I'm not sure why that was happening.... perhaps the 1/2 speed feature isn't exactly frame perfect 1/2 speed? Knowing the developers that really wouldn't surprise me :dizzy:.


Oh, and for comparison here is mk's dtilt in 60fps progressive: (lol IASA frames)



<3




hotgarbage said:
deinterlacing
what?


i set it to 60 fps if that helps...
Oh yeah, I never answered this :p. I'll try and make this brief as possible.

Interlaced video runs at 30 frames per a second. However, each frame is made up of two fields. Each field is half an image, and when combined with another field it makes a frame. The fields combined aren't the same image though, which will give you shenanigans like those seen in that pic I posted above. So in interlaced video there are 60 fields per a second, and 30 frames per a second.
Now bob deinterlacing works by separating the fields and turning them into frames(in other words it completes the half-picture of the field), turning the 60 fields (30 frames) into 60 true frames per a second.


I hope that made sense ?____?. (it probably didn't xP).



One last thing; all pinnacle usb capture cards (and nearly all capture cards for that matter) capture video at 30fps interlaced. You said you set it to 60fps.... I messed with every fps setting on virtualdub and none of them really did anything. Some slowed the video down but it didn't change the video itself. Others really didn't do anything, which makes since as virtualdub can't change what the capture card is capable of doing.
 

napZzz

Smash Hero
Joined
Mar 4, 2008
Messages
5,294
Location
cg, MN
is this even necessary? lol....all hism oves are fast and lagless...its all you need to know ._.
 

Tenki

Smash Hero
Joined
Apr 3, 2008
Messages
6,966
Location
GA
Yes, because it will statistically disproves idiots who think that he has no lag on his moves... or it can disprove idiots who think that there are lag points and openings in his moves.

With raw statistical data instead of "bawww mk is top tier no wei to breek hiz spacingmoveslol unpunishable laulwuh"
 

Crizthakidd

Smash Champion
Joined
Oct 1, 2007
Messages
2,619
Location
NJ
do u guys take into accouny frames dropped by ur capture card or your visual recordings from the game > to the comp or does that not matter.

on teh question of frame hitboxes, whats the fastest thing u can do before u can put ur shield up including downtime and air frames where u dont do naything if its like up air >land>shield
 

Tenki

Smash Hero
Joined
Apr 3, 2008
Messages
6,966
Location
GA
on teh question of frame hitboxes, whats the fastest thing u can do before u can put ur shield up including downtime and air frames where u dont do naything if its like up air >land>shield
Isn't shield the fastest thing you can do?
 

JCav

Smash Lord
Joined
Feb 2, 2008
Messages
1,217
Location
Michigan
ok hotgarbage, i got my data by doing the following

yes you are correct, my virtualdub only runs at 30 fps so i did get my data at 1/2 speed

and i see what you are saying and wow, you know your stuff
 

hotgarbage

Smash Lord
Joined
Oct 15, 2007
Messages
1,028
Location
PA
do u guys take into accouny frames dropped by ur capture card or your visual recordings from the game > to the comp or does that not matter.
Definitely. Dropped frames are rare, but they do happen so it's important to test a move multiple times to eliminate that possibility.

ok hotgarbage, i got my data by doing the following

yes you are correct, my virtualdub only runs at 30 fps so i did get my data at 1/2 speed

and i see what you are saying and wow, you know your stuff
Just learned it from reading stuff online :p.


Anyways I forgot to add to my last post how to bob deinterlace:

- capture your video as you normally would. Make sure you're not getting any dropped or inserted frames. If you are disabling audio capture may be good (as you don't need audio to look at frame data). Also going to "capture" -> "timing" and unchecking the top two boxes is probably a good idea.

- exit out of capture mode and open up your video

- go to "video" -> "filters"

- click "add"

- select the filter "bob doubler" and click "ok"

- check the "top field first" box and click "preview"

- that's it! Now you can watch the vid at 60fps progressive 8].


The quality will be pretty lousy (it'll look like the video is vibrating) but it'll get the job done as far as getting frame data is concerned.
 

JCav

Smash Lord
Joined
Feb 2, 2008
Messages
1,217
Location
Michigan
uh oh



i have "field bob" but no "bob doubler"

solution?



and thanks for all of your help by the way, i really appreciate it :)
 

hotgarbage

Smash Lord
Joined
Oct 15, 2007
Messages
1,028
Location
PA
Sure thing!

What version of virtualdub are you using? Version 1.80 onward should have it. If you don't see it maybe you should try downloading it again.
 

beenman

Smash Rookie
Joined
Jun 18, 2008
Messages
17
i was wondering if you could put air dodge speed, not sure if its uniform for all cha's but either way it would be nice to know
 

hotgarbage

Smash Lord
Joined
Oct 15, 2007
Messages
1,028
Location
PA
meta Knight:
Spot Dodge Lasts 25 Frames
Invincible Frames 2 – 20

Roll Backward Lasts 33 Frames
Invincible Frames 4 – 12

Roll Forward Lasts 23 Frames
Invincible Frames 5 – 12

Air Dodge Lasts 39 Frames
Invincible Frames 3 – 28
1010101010
 
Top Bottom