Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
well, it's still tricky (and I agree, at this point #2 becomes less appealing). #3 involves making a slight detour to reduce the time (and sometimes, a long detour.)... up until now, EVERY trick discovered has been a skill based thing, not a "stop-the-clock" type deal. Doing the trick, as fast as I pull it off still looks like a detour", and most of the awesomeness in this game is from going at break-neck speed and not stopping for anything.
So I'm still deciding between #1 and #3.
Joined: 8/16/2006
Posts: 24
Location: New Brunswick
I'd go for real time... I'm actually disappointed my file has been tainted with the 13.98 abusing the text blocks. While this trick does offer a lot in terms of getting lower times, I think the complete lack of actual technique speaks for itself.
I am unsure if I'm even going to use this to improve times on on other levels and might force myself to get 13.98 or better on 1-02... personally, I think its pretty cheap and overall takes from the impressiveness of a run. Of course I'm not talking about TAS' but I think it applies to them in the same manner.
Realtime is noticable to the viewer, while in-game time only flashes by really quickly at the end. The site has set a precedent that a realtime run won't be replaced by an IGT run too, so don't worry about this trick obsoleting the run.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
Here is what it looks like, using full abuse on an easily "on the way" clancer. I made this optimal by using my memory watching to see what the farthest X pixel position I could have before the text box disappeared. As you can see, I still made it fast... but there's 2 more clancers I can do it with earlier in the stage, and I have to slow down in order to activate the texts off of them too. I'm in full agreement with Veysey... it's cheap and removes impressiveness (despite me being able to nail this particular clancer in an impressive way.) Lengthening the movie just to reduce the in-game time seems against what makes the Mischief makers TAS look so good.
The only problem with labeling the run as "aims for fastest real time " is thus:
-Many stages are quite laggy.
-In order to do some really impressive things (like showing off), sometimes I lag up the game unintentionally.
-Not only do I suck at lag reduction, but making a lag perfect run is just not going to happen for this game.
-Stages are split into 52 separate levels, and defined by the time I got on it. Aiming for the shortest amount of time (without using the text trick) for each of these stages makes sense, as that is what the game is about.
-In-game time based make the day of events faster (jump event.)
-All other reasons mentioned on the first page of this thread.
Here is what it looks like
-Not only do I suck at lag reduction, but making a lag perfect run is just not going to happen for this game.
Thank you for uploading
Just because its hard to reduce lag doesn't mean you shouldn't try. This does sound like the hardest element in the run, but its also the hardest element to do in SMB3 and Genisito pretty much nails it, I don't see how it would detract from the fun if you destroy enemies at the earliest opertunity in order to reduce lag or other possible methods.
If I remember right, attacking in this game requires a bit a thinking, as our hero Marina does a bit of a toss action to beat stuff up. When she throws something she pauses for just a brief second...so yes. Lag reduction in this game = difficult.
I am just a silhouette, a silhouette of a memory of a solitary night .. nothing more.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
Not only that, but sometimes I have to do lots of stuff really fast in really short amounts of time. This causes lag.
Try not to think of lag being created in ways you expect like in SMB3. It usually wastes time for me to destroy an enemy, since I have to grab something. Mainly the reason I'll interact with enemies is if I have to or if I want to do a clancer boost.
Here's where I have experienced lag and why:
1-03: After I grab and shake the last white clanball, lag appears if I do a certain something. There are many ways to get onto the ground after I shake the clanball; some cause lag, some don't. I chose the method which gave me the lowest in-game time, since at the time that was my goal; this method also lagged it a little bit.
When I completed 1-05, I was freakin' fast (over 1.5 seconds faster than nitsuja I believe) I simply make the game go "wtf how can he done already" and the song sequence with all the tiles is massively lagged, simply because I beat the stage extra fast. A few other stages will have the last sequence be lagged because it's completed so fast, like snowstorm maze for example.
In 1-08 (a playaround stage) I go absolutely nuts in moving Marina around and throwing things. Once or twice, the sheer amount of moves I'm doing in a short amount of time causes lag.
Now, 3 of the stages I've encountered so far- 2-03, 2-08, and many of the day of events, like dash and hurdle- naturally run at 30 FPS, there's nothing I can do about it.
So, in the past I never cared about reducing lag since it didn't affect in-game time. Now, I probably still won't care about it much either (and hopefully that guy can make a plugin that removes that unnatural lag). I'm continuing the TAS (albeit hexing in some improvements to stages as I find them, rather than redoing whole stages all over again) as if the text trick wasn't discovered at all.
I'd say both in-game time and real time should be stated as goals, but make one the 1st priority goal and stick with it. As the options pertain to this decision...
In-game first, real second: Aim for fastest in-game time and go for fastest real time when it doesn't interfere with the first goal. This would mean heavily abusing the text box trick to reduce stage times.
Real first, in-game second: Use the text-box trick, or any other trick, to get a low stage time but only when it doesn't cost you real time frames.
I don't really have a preference, but I don't like fuzzy goals and I hope you'll be consistent one way or the other. I'd hope for lag reduction either way (as real time should be a significant goal, if not the primary goal), although I understand that may be asking a lot.
And hey, I'm probably going to love the whole thing regardless.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
I hate fuzzy goals too. Veysey and I are 100% dead set on not using the trick at all. It will be mentioned, and a video posted of the trick in the submission, but I am going to avoid using it except for in 4-04, because that gives a speed boost for Teran besides stopping the clock.
Bad news: I'm back in world 1 again.
Good news: I shaved even more time on the first 5 stages. Many of the next 16ish stages should be directly hexed in, and if all goes well be back to Migen's Brawl in no time.
Also, I've yet to hear back from 'le dude who's making a new plugin, but one of his goals is to reduce the lag that is found in a few of the stages.
Sorry for not getting back to you earlier comicalflop.
I've been very busy with other projects and university, and so I've had very little time to allocate to my N64 graphics projects, yet alone to the things we discussed. I'm very sorry about this. I'm hoping to get back into Rice Video development soon to do some work on the OpenGL renderer. This might be a area in which Mischief Makers might be a focus, but I'm hoping to focus on the more pressing issues first.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
Hey, it's alright, I'm back in world 1 again anyways, so the timetable got extended even more. Do what you gotta do, pressing issues always come before less mandatory ones. One question though: do you know if the massive lag in 2-03, 2-08, and dash events are caused from the ROM or the plugin?
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
Of course I know that, having watched veysey's/reik's/TheCape's vids for months. I know for a fact that the only stage on console that has unnatural lag is the hurdle event in the day of. I'm curious if it's the ROM because some ROMs have been known to be created (dumped) incorrectly... for instance, DK64 (U) doesn't run, but DK64 (J) does. Also, not a single plugin that I've tested doesn't have those laggy stages in 2-03, 2-08, and events.
Also, not a single plugin that I've tested doesn't have those laggy stages in 2-03, 2-08, and events.
Well, Mischief Makers is a very sprite intensive game, and sprites in this game soak up video memory (especially when using scaling filters like HQ4X). You could try using GlideHQ with S3TC texture compression to help somewhat.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
Update!
I just finished the first stage in world 2, and I just passed the "1 second faster than old WIP" mark. Some of the following stages will be hexed, some I want to test more.
Question: Could an encoder fiddle with all sorts of computer settings and such to make the final AVI not have those stages as laggy, then switch to normal settings for the rest of the stages?
Oh, and Halamantariel is awesome. Big shout out to him, who has done so much hard work with all of the hexes for all of the hundreds of .m64's we send back and forth.
Question: Could an encoder fiddle with all sorts of computer settings and such to make the final AVI not have those stages as laggy, then switch to normal settings for the rest of the stages?
One could. I doubt it would be accepted as publishable though.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
World 2 is officially done! After 4 months, I can finally move on to world 3 without ever feeling the need to go back and redo anything in world 1 or 2 again. Thank goodness, I am sick to death of those 21 stages.
Here's a little teaser to whet appetites:
Migen Brawl Old Version 1:33.88"
This is the first version of Migen Brawl I did last year (about 3 and a half seconds faster than Nitsuja's WIP), in the movie where I had gotten as far as Missile Surf. My current movie will be faster both in time and in how Marina moves. Enjoy!
Mudlord, any progress?
That was a nice teaser. Thanks, comicalflop.
I guess it's pretty obvious and likely corrected in the new version, but the dashing back and forth around 0:57 looked pretty laggy.
There must be a huge amount of luck manipulation involved in these boss fights. Is that what the lava dives are about?
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
That's exactly what the lava dives are about. They manipulate luck like nothing I've ever seen before. If I don't like a particular sequence Migen does, I take a lava bath, and within a few minutes of varying what frame I hit the lava I can find the exact sequence that I want.
Also, I don't give a flying banana about lag, since a lot of the lag in the game is unavoidable, and the run is in-game time based anyways. I don't think I did that same laggy dashing thing, but I moved about so fast in my current version that I may have lagged things up a bit.
This fight drives me insane because:
-you're right, it has a huge amount of manipulation. I'd be so easily gaining and and then losing frames all the time compared to other versions I had done
-It's a pretty long fight, ergo incredibly tedious
-Each punch, I vary what frame I actually counter the punch, and then check every single punch following until his catchable punch comes up, and I test every single possibility until I have the fewest amount of waited frames aka his catchable punch comes at the absolute earliest moment.
-I have to look incredibly hard to find time savers. The counter-intuitive and weirdest things save time, and I have to redo constantly each time I find a new time saver.
-Due to randomness, sometimes a previous version will be faster than what I currently have, so I have to research to find the cause and start from scratch.
-On top of everything, I have to keep Marina constantly moving in entertaining ways. It was very hard to do massive manipulation and keep Marina moving like she's on espresso at the same time, all the time.
Joined: 8/16/2006
Posts: 24
Location: New Brunswick
Boss fights usually aren't bad ... there is actually very little luck manipulation in the other boss fights... though Lunar and this are the two worst levels for it... this one especially so.... I absolutely hate trying to get a good time on it =/
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
There's actually some luck manipulation in the other bosses.
-Wormin' obviously has manipulation. Even the smallest movement can make the exploding worm go left instead of right. A heavy source of desynchs when Hex Editing.
-Western World: I have a OHKO that I have been unable to ever reproduce with different factors. It's in my current movie. Basically everything from how fast you're going, to how high you go, etc. affects the positions of the clancers that you throw the gun-clancer into.
-Just like Migen, going into the lava will affect Flambee's 3rd sequence. Desynched once when hex editing.
-Chilly Dog does have some manipulation, not a lot though, but enough for me to make a difference.
-Toadly Raw is random.
-7 Clancer Kids is random.
-Inner Struggle is likely incredibly random. Varying what frame you enter probably greatly affects what words will appear, I think it's possible to make death not show up at all.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
That's odd. Do you often have problems watching published .AVI's from the site? Because basically what I did was a single pass with h264 codec, same as what most of the published vids use.
edit: aha, I have completely figured out the randomness of the ball event. n Neither varying entering the stage, nor what frame you press Z to skip the cutscene affects it, but what you do when inside the stage entirely determines the outcome. I've gotten 10's, 8's, 9's, etc. all with a savestate after I skipped the cutscene. Well, this is going to suck major time, as if my theory is correct then I will have to redo the entire 60 seconds over and over again until I get the highest amount. yipee!
Joined: 8/16/2006
Posts: 24
Location: New Brunswick
Comicalflop wrote:
There's actually some luck manipulation in the other bosses.
-Wormin' obviously has manipulation. Even the smallest movement can make the exploding worm go left instead of right. A heavy source of desynchs when Hex Editing.
-Western World: I have a OHKO that I have been unable to ever reproduce with different factors. It's in my current movie. Basically everything from how fast you're going, to how high you go, etc. affects the positions of the clancers that you throw the gun-clancer into.
-Just like Migen, going into the lava will affect Flambee's 3rd sequence. Desynched once when hex editing.
-Chilly Dog does have some manipulation, not a lot though, but enough for me to make a difference.
-Toadly Raw is random.
-7 Clancer Kids is random.
-Inner Struggle is likely incredibly random. Varying what frame you enter probably greatly affects what words will appear, I think it's possible to make death not show up at all.
Eh? Wormin' left instead of right? Not sure what you mean, but I can consistently enter the stage and grab celes, then the worm? Doesn't seem too random...
Not sure about Western but i can consistently make him pop quickly?
3rd sequence in Flamebee is random as far as i could tell as well.
Chilly Dog can consistently be done in the same pattern in real time?
Toadly Raw as well, seems to follow the same pattern every time in real time
7 Clancer Kids - oddly enough, I've noticed that only the giant clancer beam thing is random... the pots actually seem to follow a set pattern in dropping the rocks... or maybe they have 2-3 set patterns... but either way, I've noticed a set pattern in particular that shows up a lot (I've been playing with this recently)
Inner Struggle is random but the words always come up in relevance to Marinas position...
When I say random, I mean there is no realistic way to (in real time) control the battle flow. Bosses like Wormin, Chilly Dog and Toadly Raw can be completely controlled right from the getgo stopping the boss in their tracks. In this game there seems to be certain ... techniques... you can incorporate to ensure an S rank/best time with little regard to random attack patterns and such. And even though Migen is random - by using the 'avoid fireballs' trick - you can consistently get an S rank... my point being that Treasure obviously tried to make it so that if you know what you are doing, it's possible to get an S rank with no worries about something 'random' screwing you over mid battle. So pretty much - Get the flow of battle going and you are set.
Of course if you let the bosses go outside of the sequence, they go seemingly berserk/random - but I don't think thats the issue at hand here and shouldn't be a concern.
