1 2
18 19 20 21
Sir VG
He/Him
Experienced Forum User, Player (35)
Joined: 10/9/2004
Posts: 1910
Location: Floating Tower
Damn, Northwall Act II was amazing. Great job. Can't wait to see the final product! Too bad none of my ideas for manipulating Kalia worked. Oh well.
Taking over the world, one game at a time. Currently TASing: Nothing
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
OK, I finished Death Heim. http://www.mediafire.com/?dyn1wdgd6a8izwn Take a look and see what you think.
Sir VG
He/Him
Experienced Forum User, Player (35)
Joined: 10/9/2004
Posts: 1910
Location: Floating Tower
Overall, very good. Couple of things: 1. On Flame Wheel during Death heim, I swear you delay the last hit a bit and take some damage. Intentional? 2. Kalia seems a bit slower then I expected. Is that the best RNG you could get? Also felt like an extra hit was taken here too. 3. Arctic Wyvern - I know you couldn't take any extra damage, but maybe changing how the prior 2 fights go would lead to more swooping attacks (which are faster)? 4. Satan = You should be able to hit all 4 spheres in 2 hits It could cut out a little bit of lag from the 2nd part of the fight, which I know has tons of it as it is. 5, Probably not possible, but any luck trying to get Zeppelin Wolf to get on the first level instead of the 2nd? Though maybe it wouldn't make a difference. It feels really close, but I just have a little nagging doubts about mainly the last few boss refights.
Taking over the world, one game at a time. Currently TASing: Nothing
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
1,2,3:Yes, that delay on Flame Wheel was intentional in order to get a different pattern on Kalia. But I've gone ahead and tried the other pattern and I ended up with 10 less frames overall by the end of the Kalia fight, and time saved in the Arctic Wyvern fight. 4. You're right, it is possible to get all 4 spheres in 2 hits. I had been experimenting but couldn't manage to get the top two spheres in one hit. I must've forgotten to jump boost. Now that I've remembered to jump boost, I can hit both. 5. No luck trying to get Zeppelin Wolf to show up on the 1st level. When I mess around, I just end up with it either showing up on the floor twice in a row before showing up elsewhere, or showing up on the 2nd level. Here's my revised version, 178 frames saved: http://www.mediafire.com/?5yo1egc2sgaxved
Sir VG
He/Him
Experienced Forum User, Player (35)
Joined: 10/9/2004
Posts: 1910
Location: Floating Tower
Damn, got though 3 cycles of Arctic Wyvern and he's left with ONE health left? That sucks. :P No way to get him 6x in the first round?
Taking over the world, one game at a time. Currently TASing: Nothing
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
Sorry, but I think that given that particular pattern that's the best I'll be able to do. Hitting the Arctic Wyvern 5 times when it swoops in is difficult enough. As for trying to hit the Arctic Wyvern 10 times when it drops straight down, I couldn't quite manage it. After some comparison with the US version, I found that the Arctic Wyvern starts flying away about 10 frames sooner in the Japanese version, which is enough to prevent getting in a 10th hit. As for trying to get a different pattern out of the Arctic Wyvern, Messing around with when I killed Kalia didn't make a difference.
Sir VG
He/Him
Experienced Forum User, Player (35)
Joined: 10/9/2004
Posts: 1910
Location: Floating Tower
Damn. The victory seems a bit hollow when there's one stinking health point left and no way to change things up. Between that and the first Kalia fight...makes me wish there was just someway to improve those, but I'm out of ideas. BTW, does the jumping during Satan 2nd half reduce lag? Cause I know that part of the fight is really laggy and I wonder if standing still would be better.
Taking over the world, one game at a time. Currently TASing: Nothing
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
I tried standing still more often, but it actually made things worse. So I tried jumping more often and shaved off 75 more frames. http://www.mediafire.com/?8yqzzz1yoe1y9z3
Sir VG
He/Him
Experienced Forum User, Player (35)
Joined: 10/9/2004
Posts: 1910
Location: Floating Tower
Wow, that seems backwards of normal logic. Usually moving less avoids the lag, but hey, who am I to argue with results? I'm out of ideas for improvement. Any spots that I've questioned have either been improved or aren't possible to manipulate the way the need to be. Great job man. :)
Taking over the world, one game at a time. Currently TASing: Nothing
Active player, Experienced Forum User (474)
Joined: 6/9/2010
Posts: 12
Location: Japan
Hi!I'm Japanese. This is my TAS I made last year. http://dehacked.2y.net/microstorage.php/info/817394996/actraiser%20%28J%29_sp_no%20glitch.smv and,I know a short cut. http://dehacked.2y.net/microstorage.php/info/549037876/actraiser_BP2.smv I think that these are useful for you. Thank you for having me over.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
Hello Hotarubi. Thank you very much for posting the Bloodpool Act 2 trick. I was trying to do the same thing when I was making my TAS but for some reason I didn't quite manage it. After watching your SMV I was able to do it easily just by watching your TAS, without even watching your memory or inputs! I will also compare the rest of my TAS to your TAS to look for other timesavers for my TAS. Since you are a Japanese TASer, I would like to ask if you also made the following TAS: http://www.nicovideo.jp/watch/sm3944699. If you didn't, then do you have a way to contact the person who did? I would like an SMV of their TAS if possible because that makes comparing TASes easier.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
OK, so I've finished comparing my TAS to Hotarubi's TAS. I went ahead and made a table with differences between my TAS and his for segments of each stage. It was... a humbling experience. Don't be fooled by the total number of frames in Hotarubi's SMV. The number of frames to final strike in his TAS is actually 77942. That's only a 383 frame difference with mine, in spite of him not using the walljump glitch! Sure, some of this lack of difference is due to extra lag in 1.53, but Hotarubi showed numerous ways of improving time in stages. Common themes included finding more ways to use the fast fall-forward trick to save time and better management of when to get hit and when not to get hit (getting hit can save time in some places, but you also have to consider the time spent refilling the health bar after beating the boss). There were lots of other things that add up. There were many places where I had managed the same time or even a faster way than Hotarubi, but nevertheless I was impressed with how many stage segments (and sometimes boss fights) he had done more quickly than me. I had surpassed the Nicovideo TAS's times in just about all of the stages, but Hotarubi's is even better in many segments. So now I am almost certain that Hotarubi was not the author of the Nicovideo TAS. Of course, I will reanalyze stuff to see if I can find ways faster than both my first version and Hotarubi's TAS. Hotarubi did a great job, so that will be difficult but I tend to surprise myself by somehow extracting at least one more timesaver. Expect a fair bit of time savings, barring bad luck in the boss fights.
Sir VG
He/Him
Experienced Forum User, Player (35)
Joined: 10/9/2004
Posts: 1910
Location: Floating Tower
Finally got to watch Hotaruubi's TAS. What I thought about it: Good spots compared to zidanax's TAS Bloodpool Act I boss fight Casandra Act I area transition Getting past the alien heads in Marana Act II w/o damage Northwall Act I boss fight (ending) Major thing I noticed: The lag emulation in 1.53 is definitely much more accurate then 1.51. Satan fight in the game on a real console is laggy as hell. 1.53 from what I saw of zidanax's TAS is pretty accurate. The 1.51 TAS was not by a mile. There was very little lag compared to what exists on the real game on a SFC. Can't wait to see a combined product of both pieces of art.
Taking over the world, one game at a time. Currently TASing: Nothing
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
Hmm, I tried playing back my TAS on SNES9x 1.52 and it actually didn't desync until the beginning of Filmoa Act 2. I inserted one extra frame of going right and that was enough to keep it in sync until the Bloodpool Act 1 boss. Those results indicate to me that the amount of lag in 1.52 is much closer to 1.53 than to 1.51. The thing about 1.52 is that unlike 1.53 it has a lag counter and lua scripting. What matters most to me here is the lag counter. I therefore think I will make this next version in 1.52 given the importance of minimizing the effect of increased lag from 1.51, which will be far easier with a lag counter. As for my cited reason of improved emulation in using 1.53, 1.52 still has the "big one" of incorporating much of blargg's research on the SPC700/DSP.
Active player, Experienced Forum User (375)
Joined: 9/7/2007
Posts: 329
Computer TASer of 2015Computer TAS of 2015
Apparently, the email notifications for this thread stopped working for me. I came back from SGDQ to see the run in the workbench. I didn't realize there were posts in this thread until the run was cancelled. I will make sure I don't miss any posts so I can help out. I would like to take a look at how to manipulate the bosses better. The actions the bosses choose seems to change less frequently than stardust, but I hope there is a pattern to it. Good job, Hotarubi, with those improvements. I take it that means that our run is now improvable. I will have to start thinking about possible improvements to the simulations so we can improve our run in the future after we are done with our other TASes.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
If we go ahead and improve our run, we will very likely want to continue using 1.51 even though I think 1.52/1.53's extra lag is likely truer to the SNES. The reason is because 1.52's lua support is very limited (and 1.53's is nonexistent). This isn't a big deal for a Special Mode run, but for a normal mode run it would be bad because you have your script that you use in Simulation Mode and I have my Stardust manipulation script. Of course, by the time we finish our other TASes 1.52 and/or 1.53 may have better lua support in which case nevermind what I just said.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
OK, here's Filmoa Act 1 completed. http://www.mediafire.com/?0qjp4iw58e0hwpo This time play it back with Snes9x 1.52. I managed to improve over my previous run by 17 frames, some of it thanks to Hotarubi and some of it from my own discoveries. Because of extra lag during transitions and the point up to where you can select Special Mode, this SMV is still 44 frames more than Hotarubi's SMV counting up to the beginning of Filmoa Act 2, but rest assured that the actual in-stage gameplay is faster. Dunnius, if you feel like investigating something apart from changing boss behavior, see if you can find a way to reduce lag during transitions, like when you go from one room to another or one stage to another. I experimented but didn't find anything. EDIT: Strangely, some door transitions don't have extra lag in 1.52. The door transition in Filmoa Act 2 just after the miniboss is the same length (or almost the same length) as in 1.51
Active player, Experienced Forum User (375)
Joined: 9/7/2007
Posts: 329
Computer TASer of 2015Computer TAS of 2015
Since 1.52 is supposed to lag more than 1.51 (and is more correct), you shouldn't worry too much about being slower. But I will see what I can do to reduce lag.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
Hmm, the lag counter is already coming in handy. It caught some lag in Filmoa Act 2 that I otherwise wouldn't have noticed, and was able to avoid some of it. Speaking of, here's Filmoa Act 2: http://www.mediafire.com/?9t2z3fzhaonm57m 16 frames improvement in this stage, for a total of 33 frames so far. Again a mix of Hotarubi's ideas and my own discoveries/improvements.
Post subject: actraiser wip: youtube
Experienced Forum User
Joined: 3/18/2006
Posts: 971
Location: Great Britain
youtube http://www.youtube.com/watch?v=b53VvvieZMU
zidanax wrote:
Hmm, the lag counter is already coming in handy. It caught some lag in Filmoa Act 2 that I otherwise wouldn't have noticed, and was able to avoid some of it. Speaking of, here's Filmoa Act 2: http://www.mediafire.com/?9t2z3fzhaonm57m 16 frames improvement in this stage, for a total of 33 frames so far. Again a mix of Hotarubi's ideas and my own discoveries/improvements.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
Bloodpool Act 1: http://www.mediafire.com/?ildicdh2a0pbe72 A 56 frame improvement, mostly from using Hotarubi's method for the boss fight. EDIT: Bloodpool Act 2: http://www.mediafire.com/?t4s426fr52f7508 Good lord that was a pain. Once again, I got not-so-great positioning for the boss, with it positioning itself at the top once, where you're inevitably going to get hit if you want to dispose of it quickly. The problem is that I arrive at the boss with no hits to spare. With Hotarubi's new way of getting past the room with spikes I saw that I could get rid of the goblin after those spikes relatively quickly (lost just 16 frames), sparing me one hit to use against the boss. Also, Hotarubi had demonstrated a way to act against the Bloodpool Act 2 boss when it's at the top while taking only one hit, although I had to change it around a bit because he was using it against the Death Heim version which isn't the same. In any case, I saved quite a bit of time in this stage mostly since I didn't have to spend a bunch of time taking out monsters just before the boss. Now a total of 193 frames saved so far. I also stumbled upon an unfortunately useless glitch: http://www.mediafire.com/?vusqv00y0n85u2q
Active player, Experienced Forum User (474)
Joined: 6/9/2010
Posts: 12
Location: Japan
zidanax wrote:
I would like to ask if you also made the following TAS:http://www.nicovideo.jp/watch/sm3944699.
I know the people who made it.However, it doesn't know the where to make contact:( This is a minor speed trick,though you might already have known. http://dehacked.2y.net/microstorage.php/info/222896519/actraiser%20%28J%29_early%20landing.smv
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
I didn't know about that trick, so thank you. Unfortunately I only see one place within the first 4 stages where it could be useful--on the approach to the miniboss in Filmoa Act 1. Unfortunately this eventually introduces extra lag, thus nullifying the 2 frame gain that would otherwise be had. For many jumps, waiting for God to crouch is okay because you also move forward while waiting for God to crouch, and moving forward is your overall goal at that point anyway. For places where what matters is jumping from lower to higher platforms as quickly as possible without regard to X position(like in Filmoa Act 2 just before the miniboss) the height difference has been too great, so that attempting an attack late enough for this trick instead results in God doing a crouching slash. It seems that the height difference between the two platforms must be small. A height difference of 32 is small enough, but a height difference of 48 (as in Filmoa Act 2) is too great. Of course, if you guys see another place in the first 4 stages where the trick could be useful, let me know.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
OK, Casandra Act 1 completed: http://www.mediafire.com/?k1c46xebbtv62mm There was much struggling with the first section and the enormous lag generated by those flaming clouds, or whatever they are. I was getting three clouds towards the end, thus generating an enormous amount of lag. I thus decided I had to somehow get rid of one of those clouds. I did some experimenting with various methods, but ended up using Hotarubi's idea of jumping into one of the clouds to make it run away. There is still some lag, but it's not quite as bad. I went ahead and got the apple at the top of one of the towers like Hotarubi did. It's a fairly big detour, but the big drop afterward and the tons of time saved not having to refill so many life bars after beating the boss make it a timesaver in the end. I found a place to use that quick land trick while ascending one of the towers, thus saving me a few frames. Overall, I saved 84 frames in this stage over my old run.
Experienced Forum User, Player (203)
Joined: 2/12/2006
Posts: 373
Location: Oregon
Casandra Act 2 completed: http://www.mediafire.com/?hei15i5tsptiiqt 81 frames saved, mostly from a revised boss strategy where I make the boss go to the right instead of the left. EDIT: Changed the SMV url.
1 2
18 19 20 21