1 2
6 7 8
15 16
Player (64)
Joined: 3/2/2007
Posts: 123
Location: Lake Havasu, Arizona
http://acmlm.kafuka.org/uploader/get.php?id=3770 is this what you're talking about ?? or not quite exactly ?? Cause this is just plain funny and awesome! :) ( i did a test based on how you described it ) ..and it just cracks me up, not losing any time with the walking back-forth like this. :) It was kind of hard to do it while it was recording. But tried as much as posible for a couple seconds right here hehe. i posted a direct AVI, so that you can see it 60 FPS, no delay/no lag, in windows media player or whatever people use. i really hope i didn't spoil anybody about your description. :(
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Yes, that's what I'm referring to. It looks better at 30 fps because you only see Mega Man facing backwards. But after more testing it turns out it does slow you down after all. It's not noticeable at first, but by the end of the first stretch, the subpixels lost added up to 2 whole frames. I wont be using it in my run after all :( And to answer your screw crusher comment: I can't use the DLC weapons in this run. It's not available from the start of the game. It requires DLC and then beating the respective bosses to even use.
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Chill Man is down. Give it some time to upload. Sitting at 80% for some reason. I've tried uploading it three times so far. It's only 30MB so I don't know what the problem is. Gets to 75% then slows considerably.
Player (64)
Joined: 3/2/2007
Posts: 123
Location: Lake Havasu, Arizona
The vid keeps saying "processing please try again in a few minutes" :-\ ------------------- oh yea i see. bummer that it cant be done. makes sense now. i was looking into the programming on NES mm6 specifically with frame advance and the debugger of fceux. Normal running speed is 1.4C pixels per frame. If i stop or turn around for 1 frame, the speed drops to 1.40, continues to drop by 0.0C pixels every frame. You reach 0.00 much much earlier compared to mm4 or 10. mm4 nes has a entirly different algorithim. It starts at 1.4C, then 1.32, 1.19, 1.01 (the rate becomes smaller per frame, so the sliding lasts longer) 10 is probably the very same as 6 i would think. But like 0.04 or something would be my blind-guess, if only just 2 frames are lost from doing it for the whole entire first hallway from READY, to the ice block.
ALAKTORN
He/Him
Former player
Joined: 10/19/2009
Posts: 2527
Location: Italy
you need to reupload it
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Alright its up. Had to re-encode it for some reason. Alaktorn, could you unquote me to remove the dead link? Here's the right one: Link to video
Skilled player (1327)
Joined: 9/7/2007
Posts: 1354
Location: U.S.
Pretty awesome stuff there. Looking foward to the complete run.
ALAKTORN
He/Him
Former player
Joined: 10/19/2009
Posts: 2527
Location: Italy
can't help but think the strategy is improvable the moonwalking was sweet
Player (64)
Joined: 3/2/2007
Posts: 123
Location: Lake Havasu, Arizona
oh so then you were still allowed to do it a bit, just only for enough amount of time, to not lose any frames before the next "screen scroll" case. I see. nice. (the walk back-forth) anymore longer in the 1st and final hallways would of probably resulted in a lost frame huh ? wow. that's what i'd figured then. the sliding does last longer on 10 compared to the nes games i think, so it's probably 0.02 to 0.04 (hex, not decimal) pixels deducted from mm's walking speed. which is like 80 to 120-something frames (or less) until losing 1 frame. nice. :)
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
ALAKTORN wrote:
can't help but think the strategy is improvable
Not so sure about this one, unless you see something specific. I manually timed it and it's only a second off of my world record time trial for this level. The slower areas are only the spots where having more weapons would be slightly faster. In fact, the first screen is actually 2 frames faster thanks to new tactic that jumps off the last ice block for extra height. @Kuja Killer Yes, before moonwalking, I made sure the subpixels lost never added up to a whole frame on each screen.
ALAKTORN
He/Him
Former player
Joined: 10/19/2009
Posts: 2527
Location: Italy
diggidoyo wrote:
ALAKTORN wrote:
can't help but think the strategy is improvable
Not so sure about this one, unless you see something specific. I manually timed it and it's only a second off of my world record time trial for this level. The slower areas are only the spots where having more weapons would be slightly faster. In fact, the first screen is actually 2 frames faster thanks to new tactic that jumps off the last ice block for extra height.
it's just that all the ice blocks destruction looks slow, not very TAS-like at 0:58 if you called Rush sooner it would've landed on the floor above?
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
No, Rush teleports past all obstacles until he's within one block of your level. If he can't appear on that level (for instance, if its in the middle of a block, or in the air for coil), he will continue and appear as soon as he can. If there are no spots below you available, he will continue to the bottom of the screen and disappear. That screen is actually the slowest deviation from the record route (aside from the boss fight). The problem is the upper ice block. In the time trial, a Solar Blaze easily destroys it, and a wheel cutter takes you to the next level. But without those weapons, I have to use Triple Blade and Rush. The reason why I call Rush there, is because that's the first frame its possible for me to call him after shooting the ice block with T. Blade. I switch to Rush immediately after the ice is hit, and then hop again on the ice below me to put the coil on the floor. The reason for this is because there isn't enough room above the ice block for Mega Man to jump on him. BUT After studying the screen again, I *might* be able to call him on top of the lower ice block, jump out into the space where the upper ice block was, and coil my way to the next the screen. I'll test this out tonight. If it works, it will save several frames for not having coil have to go all the way down to the floor, and then several more for the coil being higher up. Thanks for pointing me in the right direction. The other ice breaking rooms will be harder to find improvements for because of the enemy placements.
ALAKTORN
He/Him
Former player
Joined: 10/19/2009
Posts: 2527
Location: Italy
diggidoyo wrote:
After studying the screen again, I *might* be able to call him on top of the lower ice block, jump out into the space where the upper ice block was, and coil my way to the next the screen.
that's what I was thinking about, hope it works
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Yup, it definitely worked. Saved 21 frames. Then I used the Rush Coil cancel trick to call him again up top and saved 6 more. Link to video
Player (64)
Joined: 3/2/2007
Posts: 123
Location: Lake Havasu, Arizona
oh huh well that's certainly interesting. and i personally had a serious urge, i just couldn't stand it, wanting to try 1 little experiment of my own with rush jet in the first section, about how jumping off the ice cube saves 2 frames...well i was going to see what'd happen if i did this: but it's a failure. :P about 0.08 time-loss compared to the ice cube jump
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Yeah, I tried that when I ran the time trial. The problem is there's no way to get onto the Rush Jet without stopping. Even the highest ground in that screen isn't high enough to not have to wait.
Editor, Skilled player (1440)
Joined: 3/31/2010
Posts: 2109
Very nice work. I'm also very impressed by how you take the effort to investigae actual improvements if someone points them out, and I'm definitely looking forward for more.
ALAKTORN
He/Him
Former player
Joined: 10/19/2009
Posts: 2527
Location: Italy
diggidoyo wrote:
Yup, it definitely worked. Saved 21 frames. Then I used the Rush Coil cancel trick to call him again up top and saved 6 more. Link to video
now that looks better ;) edit: been studying the vid again but I can't think of anything to break the blocks faster, what about calling Rush at 0:54? is it slower?
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Unfortunately, yes. By a single flippin frame. Only. the Jet is faster, but I don't have him yet.
Joined: 12/6/2008
Posts: 1193
At 0:50 might it not be faster to take the downstairs way and the right latter to not have to wait for these 3 second bombs?
Player (64)
Joined: 3/2/2007
Posts: 123
Location: Lake Havasu, Arizona
no far too slow, by about a second a half or more. you gotta travel to the right more than that screen with the 3 bombs, and your far lower to the ground, so even a wheel isn't enough, then the screen with the bomb-wait, still have to travel a bit more compared to the left path, cause your waaay at the far right edge of the screen, while the other path has ya in the middle almost. and the only way to possibly ever get past the 2 eyeball creatures as fast as possible without getting hit is with the pink ball weapon which he doesn't have at this time yet. and then even if you get up there, shoot the ice cube in the way, and jump down ontop of the bomb on the absolute correct frame, you'll land too far to the left on the ground to gain back any time when you jump up to the higher ground, so yet even more time loss, and it also makes the rush jet-wheel-fire trick 100% impossible most likely.
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Kuja is right. I tried this promising path extensively during the time trials. I was hopeful, but even with all the weapons at your disposal, the extra distance Mega Man has to travel completely negates any benefits from avoiding the time bombs. I was able to to coax another 1-up out of an enemy in Chill Man's stage, but I don't think there's too much left on this level.
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Here is the input file for anyone who wants to see my current progress on the actual emulator. http://www.megaupload.com/?d=UNXRZPS3 It includes the first 3 stages and ends at the beginning of Nitro Man's stage. It syncs with r7302 that I recorded it on, as well as the current revision r7435. Use the following settings Please note, the playback will be 20-30fps. While watching the recording, you can only pause, frame advance, and save state. Loading a state stops playback and continues recording from that point, even with read-only mode enabled. ~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~ Also, here is some RNG information that I posted on Speed Demos Archive. I'm reposting it here since it it's more useful to this community. "You mentioned that boss patterns are only influenced by total time elasped, down to the specific frame. That's only half right. Killing an enemy also manipulates boss patterns. I proved this during my TAS of Commando Man, making him do his optimal pattern without waiting. I don't know how many patterns he has, because even after trying 50 times waiting an extra frame, he never repeated his optimal pattern. It's probably around 60. The reason is because all patterns (boss, enemy, crushers, etc) and enemy drops are influenced by the random number generator (RNG). Killing an enemy calls for the RNG to caculate which (if any) power-up to drop. This calculation changes the RNG, and all subsequent drops and patterns in that level appear 1 frame sooner. The RNG is seperate and renewed for each level. Regardless, its 's too hard to manipulate the patterns successfully everytime on the console. Even a single frame could mean the difference between a favorable pattern and a bad one. Still, if you're consitently having trouble, try killing more or less enemies. Also, keep in mind that some patterns are actually dependent on Mega Man's position. Strike Man, Blade Man, and Tsurare Stomps to name a few." ~~~~~~~~~~~~~~ "Jumping and shooting doesn't change the RNG. Also, to clarify, it doesn't matter specifically which enemies you kill. The only thing that matters is the number of enemies. That's why destroying the first and second missile turrets yield the same starting boss. So for instance, assuming never stopping and arriving to the first Wily 1 boss door at the first possible frame: -Killing nothing yields Wood Man -Killing 1 enemy yields Elec Man -Killing 2 enemies yields Elec Man Even if you killed the 2 Tsurare Stomps, it would still yield Elec Man. If you could kill 3 enemies, it's possible it might yield Gemini Man. There might be a few more ways to change the RNG, but I haven't tested them enough to be sure. I've heard some small objects like the actual missiles and footballs effect the RNG. I haven't tested that yet, but I can confirm the the bombs from Arigocks and Bomb Puttons on Commando Man's stage do not effect the RNG. Also, the ice blocks in Chill Man stage seem to do weird things to the RNG. Instead of moving patterns/drops sooner by 1 frame, I've seen it move some forward by 10 or more frames. This might be because of the way the game displays the animation when a ice block is destroyed." ~~~~~~~~~~~~~~ Q. Are you sure the weapon you kill the enemy with isn't relevant? In MM9 I'm pretty sure it was. A.What you might be mistaking for an RNG change is actually a variance in weapon speed. I tested this in Chill Man's time trial by damaging the first snowman enemy, and then delivering the kill shot with different weapons. NOTE: The missing milliseconds (2, 4, 7, & 9) are never shown on the timer. This is how game keeps accurate time even at 60 frames per second. Using Mega Buster: Shoot time - Enemy hit time - Drop 4.56 --------- 4.73 ------------- Nothing 4.58 --------- 4.75 ------------- Large Health Refill 4.60 --------- 4.75 ------------- Large Health Refill 4.61 --------- 4.76 ------------- Small Weapon Refill 4.63 --------- 4.78 ------------- Nothing 4.65 --------- 4.80 ------------- Nothing Using Triple Blade: Shoot time - Enemy hit time - Drop 4.56 --------- 4.71 ------------- Nothing 4.58 --------- 4.73 ------------- Nothing 4.60 --------- 4.73 ------------- Nothing 4.61 --------- 4.75 ------------- Large Health Refill 4.63 --------- 4.76 ------------- Small Weapon Refill 4.65 --------- 4.78 ------------- Nothing So as you can see, looking only at the weapon shoot time and the drop it yields, using a different weapon appears to have changed the RNG. But factoring in the time the shot actually reached the enemy, it's clear the RNG didn't actually change at all. Some weapons are just faster than others.
darkszero
He/Him
Joined: 7/12/2009
Posts: 181
Location: São Paulo, Brazil
Have you tried encoding this yourself? I just gave it a try and when dumping video it desyncs at the shop, and sound desyncs at the main menu. You should've made the movie file with Idle Skipping disabled, since this makes it more deterministic. However, just plain replaying syncs fairly well (met some desyncs at the main menu) and it run at 60 fps (with 30 fps when there's an achievment popup lol)
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Yes, I recorded and encoded the movie with Idle Skipping disabled. I only showed it enabled because it speeds playback. I have it uploaded already to Youtube on my channel www.youtube.com/diggidoyo Also, sound slowly desyncs over time. I don't know why, but I think it plays at a slightly different speed for some reason. Not noticeable at first, but become obvious after a few minutes. I correct for this by advancing the sound by .20 seconds every level or so after it's all dumped.
1 2
6 7 8
15 16