-Update-
3 levels done: http://nvdata.pilif.ch/Vectorman_Zer0.gmv
EDIT: 4th level done :), just click the old link
EDIT: ya, 5th level done
EDIT: 6th done, next is Rock 'n' Roller, yay Baby!
EDIT: yo, done with 8
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
Could someone help me find the memory address for vertical velocity? I've tried my best and don't come up with any useful data. (WIP is close to 17 seconds ahead of my submission, due to nitsuja's tactics and a much, MUCH improved 1st level boss fight).
Can anyone help me out?
Please don't cross-post.
0xe36e as a signed short (2-byte value) seems to work for me. Positive means down, negative means up.
Edit: If you've got wiki editing privileges, try starting a Vectorman tricks page. If you don't, you can ask Bisqwit.
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
Thanks for that, the actual address didn't work, but knowing that up is negative and down is positive helped me find one that worked (I had figured it would be the other way around). For some reason Vectorman has alternating memory addresses, the HP for the bosses will change depending on which movie I'm playing, but other things will stay the same, although it may just be Gens being weird. And I'll probably get around to a wiki page for the game sometime soon, thanks for the idea.
I tried to post this info the other day, but kept getting "invalid session. please resubmit the form."
X Pixel(left) 00FFE358 2-byte unsigned
Y Pixel(top) 00FFE35A 2-byte unsigned
X Pixel(right) 00FFE35C 2-byte unsigned
Y Pixel(1-bottom) 00FFE35E 2-byte unsigned
X Pixel(center) 00FFE364 2-byte unsigned
X subpixel 00FFE366 1-byte unsigned
Y Pixel(center) 00FFE368 2-byte unsigned
Y subpixel 00FFE36A 1-byte unsigned
X speed 00FFE36C 2-byte signed
Y speed 00FFE36E 2-byte signed
As for the boss HP. The boss will load into a different part of RAM depending on how many objects are in memory when the boss loads.
How fleeting are all human passions compared with the massive continuity of ducks.
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
Thanks for those upthorn, and that news about the boss HP makes sense. But since I'm new to RAM viewing, I'm a little lost in how to use the pixel addresses. Are they useful for hit detection and such? Any other techniques that you find useful?
EDIT: Here's how fast this game is meant to be played
http://dehacked.2y.net/microstorage.php/info/4597/Vectorman%20%28W%29%20%5B%21%5D.gmv
Note: At frame 6200 it looks like I miss a double boost, but I tried all manner of boosting and downwards shooting and I couldn't make that platform, it's just a tiny bit too far away.
Thanks for those upthorn, and that news about the boss HP makes sense. But since I'm new to RAM viewing, I'm a little lost in how to use the pixel addresses. Are they useful for hit detection and such? Any other techniques that you find useful?
The pixel addresses can be used to figure out the hit detection some, but the primary usefulness is for comparing times.
IE: If the end of a level is at the right, having a value of 1800 would generally be better than having a value of 1795 for the same frame.
If you watch this address on your run and a prior run (like nitsuja's test) at the same time, you can make sure you never fall behind. (Which you do, by the way. This WIP finishes level 1 6 frames behind Nitsuja's test. Much better than before, though)
N. Harmonik wrote:
After you're through with this, do you plan on doing Vectorman 2?
There's already an accepted vectorman 2 submission, in fact, it's been "publication underway" for several months due to encoding issues.
How fleeting are all human passions compared with the massive continuity of ducks.
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
upthorn wrote:
This WIP finishes level 1 6 frames behind Nitsuja's test.
At first I thought this said 16 frames and I got all freaked, but 6 is closer to the 4 that I had it measured as. I use the first frame where the screen turns black as my end of level state, and when I finished nitsuja's run out to do that, I got 4. A couple of those probably came from dropped frames in the boss fight (which this game likes to do every once in a while), and a slightly different boss fight. But now since I'm through level 4, with over 30 seconds saved over my submission, I don't feel like going back. Maybe for revision 3 ;)
N. Harmonik wrote:
After you're through with this, do you plan on doing Vectorman 2?
I'm actually doing them concurrently, and was thinking about keeping it a secret, but I guess the news is out now. I figured I'd give mike89's run time to finally get published since he's been on the queue for so long, but if I submit it before it hits the front page, then I'm sure he'll have plenty of time to try to beat it. However I'm trying to make these as frame perfect as possible, with the possible exception of those 4 or 6 frames in level 1 :-\
Emulator Coder, Site Developer, Site Owner, Expert player
(3571)
Joined: 11/3/2004
Posts: 4754
Location: Tennessee
The vectorman 2 wip has been on queue for so long because it is a problematic game to encode. I suspect vectorman 1 will also be in a similar situation. Just a heads up that these may be in queue for awhile despite great interest in the games by at least 1 publisher (me).
And also, I am happy to hear that a new vectorman 2 run is on the way too :)
Joined: 8/8/2005
Posts: 296
Location: NSW, Australia
mmbossman wrote:
I'm actually doing them concurrently, and was thinking about keeping it a secret, but I guess the news is out now.
Aww man... Looks like I can give up hope of my V2 being published now.
(PS. the main problem with my TAS is the final boss - don't make the same mistake I did! Only shots on the brain hurt!)
(PPS. I never actually used any of the RAM watching tools so there's probably time in the run that I can't discern.)
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
mike89 wrote:
mmbossman wrote:
I'm actually doing them concurrently, and was thinking about keeping it a secret, but I guess the news is out now.
Aww man... Looks like I can give up hope of my V2 being published now.
(PS. the main problem with my TAS is the final boss - don't make the same mistake I did! Only shots on the brain hurt!)
(PPS. I never actually used any of the RAM watching tools so there's probably time in the run that I can't discern.)
No worries quite yet, school is still taking it's toll on me and I'm taking my time with both of these runs, so maybe, if all the quirks get worked out in the mean time, it'll get published. I wanted to give it as long as possible just out of respect, since you did do a good first run, and I know what it feels like to work really hard on a run to have it beat. As for the final boss, I was curious about that, and had suspected the same thing, but hadn't proved it since I haven't bothered to find the RAM addresses for the bosses yet. However I am done with the first two scenes, and am about 2 seconds up due to an improved route and some of the techniques I found out from Nitsuja. Anyway, I hope your movie does manage to get published, since it's been in limbo for longer than I've been a member here.
EDIT: One interesting tidbit that I found on the first level through RAM watching is that you have a constant speed of 192 throughout the whole thing, with (seemingly) random switches to a speed of 224, which happen more often on the outer two thirds of the screen, ie not right down the center where the photons are. Only problem is when I tried to maximize the time spent with my speed at 224, I actually ended up with a final frame slower than yours. Very weird.
EDIT 2: I'm halfway through level 6 and I'm already up to 40 seconds saved over my previous run (two better boss fights and a route change in level 6 account for about 25 of it, increased speed is about 15), so I'm shooting for 11:30 now for a total end time.
Joined: 8/8/2005
Posts: 296
Location: NSW, Australia
I'm currently re-learning Vectorman's levels in preparation for a live speedrun at MAGfest (along with a few other games), and have of course been using both the speedrun and TAS as reference. In so doing, I believe the TAS has a number of (small) areas for improvement:
- At the end of level 1 (Terraport), once you hit the ground the camera scrolls back to the centre before the boss appears. Ideally one would land in the middle and there would be no scrolling, but I think there was a bit of sideways scrolling in this one (perhaps matching the six frames mentioned earlier?)
- I strongly suspect that the Absolute Zero boss (the polar bear) can be done faster; I would prefer to land the final hit with the fourth boost-jump.
- I think some of the large enemies in level 6 (Bamboo Mill) can be handled better, I'm pretty sure you can effect whether it will shoot at you by shooting it at the right time.
- Also, the boss of the same level would lag substantially less if you stood away from it, I think?
- Consider an entirely new strategy for level 13 (or any level with a satellite, but 13 is the most likely): destroy the satellite to go to the bonus stage and then quickly fail out of it. This will probably be slower, but at least worth considering
- Consider an alternative strategy for level 15: near the start of the level is a jet powerup, which can be taken quite some way through the level as demonstrated in the real time speedrun
If I were to make the improvements myself, it'd probably be in the S3&K cut-and-paste style rather than a full redo, as I've got enough projects on my plate as it is. Anyone else is free to take this information and apply it to a new run, of course. :)
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
We talked about a few of these things via PM last night, but I just wanted to document it here as well
mike89 wrote:
- At the end of level 1 (Terraport), once you hit the ground the camera scrolls back to the centre before the boss appears. Ideally one would land in the middle and there would be no scrolling, but I think there was a bit of sideways scrolling in this one (perhaps matching the six frames mentioned earlier?)
Excellent find, and there may be more bosses like this, I cannot remember.
mike89 wrote:
- I strongly suspect that the Absolute Zero boss (the polar bear) can be done faster; I would prefer to land the final hit with the fourth boost-jump.
I don't think so, boosting damage is not done on a per-frame basis, the maximum damage that can be done per boost is 16 IIRC. Some time may be able to be saved by taking less damage before the boss to be used to get better positioning for boosting though
mike89 wrote:
- I think some of the large enemies in level 6 (Bamboo Mill) can be handled better, I'm pretty sure you can effect whether it will shoot at you by shooting it at the right time.
Could be, I don't remember the specifics of those guys too much, but I seem to want to say that it was impossible to boost through them without a large damage knockback. Could be very wrong though.
mike89 wrote:
- Also, the boss of the same level would lag substantially less if you stood away from it, I think?
Most likely. His invulnerability period is pretty long, so it would probably be feasible.
mike89 wrote:
- Consider an entirely new strategy for level 13 (or any level with a satellite, but 13 is the most likely): destroy the satellite to go to the bonus stage and then quickly fail out of it. This will probably be slower, but at least worth considering
Huh, never thought of that, and my memory is too foggy to give a good answer. Worth looking into though.
mike89 wrote:
- Consider an alternative strategy for level 15: near the start of the level is a jet powerup, which can be taken quite some way through the level as demonstrated in the real time speedrun
Yeah, that clip you showed me would likely be a good strategy change. Route planning has never been my strong point, particularly when I was new to TASing.
mike89 wrote:
If I were to make the improvements myself, it'd probably be in the S3&K cut-and-paste style rather than a full redo
:(
mike89 wrote:
Anyone else is free to take this information and apply it to a new run, of course. :)
If you decide against doing these improvements, or if hexing doesn't work out for you, I would certainly take the time to work on this again. I actually had a note to myself on my personal page about looking into it again to see what else I could find. Keep us posted!
I actually started a run of this game two days ago. I had thought about trying out a while ago but never got to it, but then I noticed the post in this thread and figured I could try out the first thing listed, it worked so I decided to do a full run. I Just got to level 4, 252 frames faster than mmbossman's movie.
35 frames where saved in level 1, 10-11 or so on the actual level (most of which where lag reduction, sometimes at the cost of a smaller amount of ingame time), another 10-12 came from landing in the best spot for the boss and the rest was some movement tweaking during the bossfight to get more shots after it flied away. 18 frames where saved on level 2, which is purely lag reduction (23 less lag frames at the cost of 5 ingame frames). The rest comes from level 3, where the main bulk of the improvement came from handling the platforms with those annoying turrets on them differently.
http://dehacked.2y.net/microstorage.php/info/1697327789/Cpadolf%20-%20Vectorman.gmv
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
Wow, very nice Cpadolf! It's been a while since I watched my old movie, but yours felt a lot smoother, particularly in day 3. I'm looking forward to seeing the rest of this completed.
Joined: 8/8/2005
Posts: 296
Location: NSW, Australia
Very nice Cpadolf! Suffice to say I wasn't expecting this :)
Interesting that that is the best way to handle the turrets, I might try that as a real-time run strategy.
Looking forward to the rest!
Completed another 2 levels, and am now about 530 frames ahead. The polar bear could be killed with 4 boosts and 1 shot. It has 48 HP and you can do 11 damage per boost if you start from the ground, which is why it had 4 HP left after mmbossmans 4th boost, but the first time you boost through him you have time to 1st do a normal jump and start boosting just before you hit the ground which if timed right will do 14 damage. After that you just to 3 normal boosts and shoot him once where it will cause the least lag. The bird was probably improved even more, and I managed to kill it with just 2 boosts and 1 shot, doing 21 damage with the first boost and 26 with the second.
Level 5 wasn't improved by too much, I think it was around a second or so. I noticed that in this level you didn't need to do the diagonal shot after landing to keep your speed for the next jump, which was really weird, and saved some time both from reduced lag and making earlier jumps (you still need to do it if you don't jump immediately when landing, though I don't think the speed loss is as great as in other levels anyway).
http://dehacked.2y.net/microstorage.php/info/597809072/Cpadolf%20-%20Vectorman.gmv
I basically dropped out on TASing for a while due to, well, life. And so I didn't work on this for a really long time. But I actually picked it up again recently, and I'm currently pretty close to the end, so I expect to be done pretty soon.
Joined: 10/18/2017
Posts: 169
Location: San Francisco
Here is the video that had a bad ending:
Link to video
Planned runs:
Marble Madness (Arcade, Genesis J, GG/SMS)
Proposed:
Ecco the Dolphin (Genesis, GG/SMS, CD: regular, camera freeze)
Ecco: The Tides of Time (Genesis, GG/SMS, CD; normal mode)
Mario Kart DS (all cups, all missions)
In progress: Grand Poo World 3 (all exits hard mode)
