I do my best to get accurate values. They’re based on video recordings so it’s entirely possible that I’ve missed hidden start-up frames. I suspect that T. Akiba was more careful and patient than I, but I do see that difference on moves where I took extra steps to be sure about the start-up time - so it may be a difference of methodology.
Based on the numbers and theory about blockstun that I’ve seen, the result of a 1 frame difference is consistent with hitting with the close part of the jab version, and the extension on the strong/fierce version. FWIW up-close, the lariats, or the overhead frame of the jab are the fastest, further away, standing jab… I suppose I should tidy up the tabulated frame count data and make it available as a download…
Thanks for letting me know about the neutral up strong I’ll check it out. The close foward has been corrected.
Am I the only one who didn’t notice until now that you have frame counts for every attack in the game except for Akuma? This whole time, I was not seeing your frame counts to the left of every hitbox image. I was wondering why you weren’t continuing with the hitbox + frame concepts you had originally started out with, but I thought it prudent to keep quiet and let you do your thing. I was wondering what Big.O was talking about, but I figured you’d correct him. Imagine my surprise when I look again and finally see the frame data.
This is huge, it’s a really big deal. I thought the only frame data you had captured was the Remixed changes. Rufus, I suggest you put the frame numbers below each image, or highlight them, or put them to the right of every image, or make the numbers bigger, cuz they’re almost invisible the way they are now. The frame data for fireballs is hard to miss, but I had never checked out the fireball images until now, so I missed the data. I was too busy checking out the other special attacks that YBH doesn’t have pics for.
Some non-remixed attacks have frame counts that are completely off from YBH and Akiba’s data. Don’t take this the wrong way, I think the work you’ve done is incredibly good and you know I greatly appreciate all the contributions you’ve made. I think it’s incredible that you’ve accomplished so much actually.
However, I would be remiss in not pointing out some flaws that I see. I think there are some attacks that should be recaptured and recounted. Some people may not think frame data is important, but I think it’s very important. And I think it’s crucial that frame data be accurate and precise, otherwise it’s not useful. I’m only trying to ensure that all the frame data is as high quality as possible.
For example, you have captured Ryu’s Short and Forward Tatsus as having 6 grounded, recovery frames, and you captured the Roundhouse Tatsu having 0 grounded, recovery frames. But YBH and Akiba’s data shows 0 grounded, recovery frames for all Tatsus. Another example is Ryu’s Shoryuken grounded, recovery frames. For the Jab and Strong you have 1 grounded, recovery frame and for the Fierce you have 0 grounded, recovery frames. But YBH and Akiba show 5 grounded recovery frames for all Shoryukens. And that’s just Ryu. If you’d like I can make a list of attacks that I think should be recaptured.
Some people might think that Remix is a different game, therefore that explains the differences in the frame data. But that’s not true. Other than the remixed changes, all the other attacks are exactly the same as they are in ST. So it doesn’t make sense that the frame data should be different. All I’m saying is, it’s great that you’ve done so much Rufus, and it would be even better if the frame data was high quality. Again, I’m not trying to be a critic or point out flaws just for the sake of pointing out flaws. I’m only trying to ensure that all the frame data is as accurate and precise as possible.
Thanks again for all your contributions bro! :pleased:
Edit: I just wanted to point out that Ryu’s Fake Hadoken has 10 startup frames before he extends his hands forward, compared to 12 startup frames for the real Hadokens. I just knew the startup was slightly faster than a real fireball. Nice to finally have confirmation. :tup:
Edit2: I updated the HDR Wiki to reflect the frame counts.
On that note it would be very interesting if you could somehow get the framedata and hitbox for the walldive fake the actual animation of it before the fall since we might be able to deduce at what times it is faster to simply do the normal walldive pounce to get down as quickly as possible and what times it is better to do the fake.
New Running Bear Grab images (with frame counts) are up.
At least an example would be a good thing to have, I’ll add it to the list.
Most of them are probably wrong, I expect that a couple of them are right. There are a number of plausible reasons for the differences. For example, I don’t think that T.Akiba had access to hitbox mode when he did his frame counting, and then there’s the whole version differences thing. His data is (as far as I know) from the arcade, and HDR is based on the DC version of the game.
It is 5 frames of recovery for the SRKs (corrected), and I expect that the tatsus have cancelable landing animations. I’m not particularly gifted vis-a-vis technical execution so checking some stuff can get difficult.
If you look closely at the red and blue hadoken frame images, you will see that the animations are identical. Ryu has hits hands out for 2 frames before the fireball shows up on real ones.
I’ve changed the frame count stuff on the SRKs and split the double-rows on the Hados - don’t think you’re going crazy. I put the numbers up over a week of evenings, knowing that I’d have to revisit them so I’ll be addressing stuff as it comes up. There are surely bunches of transfer errors, but I’ve been working on other projects and other parts of this one. I figured nobody was complaining or commenting so nobody cared that much.
A list of places where the frame data is suspect would be very useful. (Just as long as it isn’t ‘everywhere’.)
True, but Akiba was a part of the YBH team too. Akiba has captured ST’s frame data twice, cuz he wasn’t happy with the way frame data was captured for the YBH, or something like that.
Edit: Here’s a post by Ganelon that confirms this. There are a couple of other posts that verify that Akiba did work on YBH (I can’t find the links for them right now).
I care a lot. I wasn’t complaining cuz I had no idea you had frame data for the non-remixed attacks until now.
I did some related reading, and apparently there’s ‘japanese’ notation where they count the first hitting frame rather than the number of non-hitting frames at the start.
Well that is strange (and presents a different problem). The issue is that special moves and aerials match yoga book and T. Akiba numbers while the ground normals don’t (for Zangief anyway), so even if they are in the “Japanese” notation you mentioned that would mean all your specials/aerials are off.
As I said above, I can’t speak to how the T.Akiba or YBH numbers were produced. It is much less likely that there are ‘missed hidden startup’ frames on jumping attacks since it’s going from one animation to another. To illustrate how I calculate start-up frames, I’ve put up an extended set of images from Zangief’s diagonal jumping splash in the misc section: http://www.pedantic.org/~nate/HDR/misc/zangiefsplash.html
I regularly find errors on the web site - mislabled moves, frames out of place, mistransferred numbers and so on, but if there’s an issue of all jumping moves are off by X, it’s basically going to be “here’s what I did”.
Jump Frame Counts. Nice, thanks Rufus. I always thought Zangief was the slowest jumper, but looks like Honda takes that cake. = J I noticed you don’t have post-jump frames. They’re probably the same as the pre-jump frames, but it would be nice to have a column for the post-jumps too. Thanks again. = J
I have so much on my plate now, I’m not gonna be able to get on that list you wanted. As soon as I have more time, I’ll get it for you.
I’m just writing up a bit on what would need to be done to balance Akuma (lol) and I was wondering if someone can confirm the actual recovery for each of his ground fireballs. I’ve heard conflicting information. One site said all his blue fireballs have 40 frames of recovery (in relation to Ryu’s 40,41,42 and Ken’s 39,40,41), but the wiki states 53 (I’m guessing this is the length of the fireball animation overall). Then in your post the last number in brackets is 39 for all 3, which would make sense seeing as I heard 40 elsewhere.
Also what is the number in blue. Is that invincible startup?
I’m guessing it is just 39 or 40 frames, but some confirmation would be cool, instead of just assuming.
Edit - so I just reconfirmed:
All strenghts of blue fireballs 14 frame start-up, 39 frame recovery.
The first 4 frames of the start-up are invulnerable.
I don’t actually see a number in the articles. It may be that I’m overcounting the start-up and undercounting the recovery of the fireballs since I’m using the first frame where it appears.
Nice job on the charge time info Rufus. I’m curious as to how you managed to get hose numbers. Programmable controller perhaps?
I’d like to request frame advantages for Zangief’s holds. A per character frame advantage count for each throw is probably out of the question (would be nice to know obviously), but I’d really just like to know which hold has the most frame advantage. I suspect it is his bite hold but I wouldn’t be surprised to be wrong. As far as I know, the only noticeable difference is how far the other guy ends up after they break out (and that bite is somewhat faster than the other two).
Blue fireballs have more startup and recovery. Now same startup as Ryu’s fireball and slightly better recovery than Ryu’s. Akuma’s jab/strong/fierce blue fireball recovery times are 40,40,40 compared to Ryu’s 41,42,43, frames.