Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
Great, this 1P run of the game looks even more entertaining than current 2P version. And definitely more polished. Good thing there's 2 branches for this game.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
Using emu.frameadvance is so last week, I suggest "register" methods.
stop_at_lag = false;
function my_fun()
if emu.paused() then
emu.unpause();
emu.speedmode("turbo");
stop_at_lag = true;
end;
end;
function my_auto()
if (stop_at_lag) and (not emu.paused()) and emu.lagged() then
emu.pause();
emu.speedmode("normal");
stop_at_lag = false;
end;
end;
taseditor.registermanual(my_fun);
emu.registerafter(my_auto); -- or taseditor.registerauto(my_auto);
There was a error in FCEUX Lua implementation of emu.unpause(), I've just fixed it, so get latest SVN revision before running this script.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
nanogyth wrote:
AnS wrote:
I think this can be further automated by finding a memory address that can serve as an indicator of sync,
But I'd need that data from Famtasia, right?
Which would be tricky since its source is lost.
Yes, that's kinda non-trivial. You would need to dump the log of Famtasia values (likely using MHS, "Memory Hacking Software", I dunno). Yet the trickiest part would be writing the bot, I guess. And of course it's not obvious how to find actual memory addresses that can be reliable estimate of sync.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
IronSlayer wrote:
I don't think you understood what I was saying. I mentioned tic throws. A tic throw is when you perform an attack (often a jab or short) to create a blockstun state for the opponent and take advantage by throwing them, as they can't recover in time.
Since NES TMNT has no reversals and few animation frames, this necessarily makes for a very braindead game with a degenerate strategy.
Watching feos' videos I haven't noticed such problem. No single degenerate strategy, but lots of unpredictable situations.
IronSlayer wrote:
Fatalities and 2 vs. 2 are completely irrelevant to a fighting game's depth, strategy, and viability as an interesting title. However, the lack of a combo system, the most basic element of ANY FIGHTER, is.
My point is, this is not any fighter. Like many NES games, this is an experimental short with unique gameplay twist, similar to Technos' titles (River City Ransom, etc) that combine beat-em-up with some sports game features.
IronSlayer wrote:
You might as well argue that a platformer without jumping isn't missing anything significant. Yeah, you can find an exception here and there (Bionic Commando), but it's a massive missing piece which needs something really amazing and unique to make up for it.
Sadly, that's not the case for NES TMNT.
Yes it is the case with NES TMNT. That ball mechanic is as outstanding as the grapple mechanic in Bionic Commando.
IronSlayer wrote:
AnS wrote:
SNES version is your average fighting where you only need to memorize all attacks and combos and then just rely on your reflexes to settle attacks with defenses. Maybe I'm not aware, but is there anything innovative or experimental in that game?
Absolutely. Have you watched high-level SNES TMNT play? There's a lot more where this comes from, but here's a completely random example;
What should I see in this video? I see two well-trained players exchanging attacks in similar fashion to hundreds of 2D-fightings on arcades. I didn't see any novelty. It only proves that SNES version is a clone of a clone of a clone of Street Fighter 2. Then what's the point in learning all the attacks and combos of this particular clone? I guess the only point would be because it's based on the Turtles franchise.
IronSlayer wrote:
Huh? Why reference Guilty Gear, which is the most complicated fighting series ever made? That's like saying that applying the principles of chess to Tic-Tac-Toe got me beat at it. Like...what?!
Applying principles of chess to Chapayev game got you beat. Like that.
I mentiond GG on purpose. The series is for nerds only. And for others its tool-assisted speedrun would be as entertaining as NES TMNT, which is: not much. And here goes the conclusion: complexity of a fighting game doesn't matter. Your insults are not justified.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
feos, you're really too passionate about the game.
IronSlayer wrote:
It's missing frames, is based completely around brain-dead tic throws, and has no combo system to speak of.
What's missing frames? Are you talking about lag? Then you're wrong, there's no more lag than in Super Mario Bros.
It is not based on throws, it's just that throws are the fastest TASing tactic. When playing the game in real time, you won't be able to abuse throws that much, because the frame window for buttonpress is rather small.
There's no combo system. Oh yes, and there's no fatalities and 2vs2 mode! And no other features common for modern fightings.
A game doesn't need to encompass everything, it can be fun by having solid set of its own rules.
NES version has more creative game design than SNES version.
SNES version is your average fighting where you only need to memorize all attacks and combos and then just rely on your reflexes to settle attacks with defenses. Maybe I'm not aware, but is there anything innovative or experimental in that game?
Now, NES version has this frantic rivalry for the ball dropped by Splinter from time to time (it's not shown in the TAS). Since the ball can overpower any attack (but also can be blocked and then intercepted), players have to estimate priorities differently than in typical fightings. Some experienced players can work wonders by catching the ball mid-jump and then unleashing it before landing.
You probably tried to play the game using behavioural patterns from SF and GG, and got beaten quickly. The game has more depth than you think.
And the game was kinda legendary here back in 90s among those who couldn't afford Genesis with its MK (as for SNES, I remember only one guy from our town having it at the time). Your insults to the game can be compared to calling SMB1 shitty, because it lacks features from SMW.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
hopper wrote:
I suggest that you assume good intentions unless there is reason to believe that someone posted with bad intentions.
I didn't assume you had bad intentions when suggesting SNES version. I figured you just want to see a TAS of the game you like. So I explained why this isn't any more helpful than suggesting to TAS Mortal Kombat.
Since those games differ so much (not just graphically, but gameplay-wise), TASing experience with NES version won't help author to TAS SNES version. For him its a random game from some foreign console.
Actually I'm sure pure speedrun of SNES version won't be much more entertaining, the only solution here is playaround superplay.
hopper wrote:
Due to the limitations of the NES, the NES version of this game is quite ugly, while the SNES version has much better graphics and beautiful backgrounds.
OK, better graphics. What's more? Sound isn't any better, it's just different (chiptune vs. MOD music). Gameplay? NES version is the most balanced and dynamic fighter on its platform, and SNES version doesn't hold up to UMK3, KI and SF2.
three different fighting games based on the Teenage Mutant Ninja Turtles franchise
See, these aren't ports in any way. All 3 games have different fighters, different gameplay modes and different mechanics of fighting.
I assume all 3 games should be TASed to check their entertainment potential.
feos wrote:
We have a comunity of online players competing in NES TMNT, recording insane videos of what is in fact possible in this game
Then please go ahead and record a playaround showcasing all the funny tricks. Even if IronSlayer will vote No on that playaround just for the sake of his hatred for NES version, other people should find the playaround entertaining.
As for this speedrun, I have to vote No because of extremely repetitive scenarios.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
Для этой игры имеет значение число джойстиков. Ты делал проект в режиме 2P, а FM2 экспортировал в 1P. Надо либо экспортировать в 2P, либо переделать сам проект в 1P (для этого нужно сделать File->New project и там выбрать 1P и поставить галку "Copy current input", а потом подправить место рассинхронизации).
Вообще, какой-то скучный мувик, одни броски и суперудар Лео, не заметил смешных ситуаций (которых в этой игре можно спровоцировать довольно много).
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
Всё правильно, эта игра обновляется раз в 4 кадра, получается 15FPS. Придётся приноровиться к работе в такой обстановке. При игре с джойстика же как-то получается!
Судя по скриншоту, ты используешь фанатский перевод (T-Eng), что не разрешается по правилам TASVideos. Используй оригинал (J-версию). Ты, кстати, вводил в поиск по форуму слово "kakutou" хотя бы? Уже были ТАСы этой игры. Имеет смысл делать с ней не спидран, а развлекательный ролик.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
Lomax wrote:
Форум мертв чуть менее чем полностью) народу вообще нет)
А чего болтать-то? Тасить надо! :D Вот ты мог бы в первом Чип и Дейле разобрать и документировать всю физику коллизий, чтобы потом работать с субпикселями осознанно, а не на уровне народных примет, почерпнутых из наблюдений за RAM.
P.S. Заначка из шариков на первом боссе просто убила. Это ж надо такую батарею заготовить...
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
This isn't known bug, because most people use Memory Watch instead of RAM Watch.
I fixed these 2 loading bugs in r2498. Now try finding a .wch that can crash FCEUX!
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
CtrlAltDestroy wrote:
For me, the most irritating part of TASing is not being able to figure out one or more aspects of why the game acts the way it does.
For instance, not being able to find a specific value in RAM. Perhaps the value is a 2-byte integer or a 4-bit nibble, and doing a simple search doesn't help very much.
Or worse... not being able to find a specific rule in RAM. Such as: why does an enemy turn right sometimes and left other times, even though his data structure doesn't change? Do I have to step through the AI assembly? Where the hell is that? GAAH!
That's the problem of reverse-engineering complex systems. Actually it can be fun on its own! To me it's more interesting than simply squeezing frames by brute force (replaying one segment several times). So while it may be hardest part of TASing, it's definitely not the most irritating.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
О, а ты где попрощался? Тоже что ли из-за этой первоапрельской шутки про технофашизм? Блин, какие все чувствительные, невозможно расслабиться ни на минуту, кругом требуют политкорректность.
Технической стороной сайта заведует Nach, остальные в этом разбираются гораздо меньше, так что лучше напиши ему в личку. Только хостить сайт на сервере какого-то русского они вряд ли согласятся. :) Зато, может быть, он сможет тебе рассказать, почему сайт так тормозит (мне и самому интересно было бы узнать, но я в этом плохо разбираюсь, так что даже не спрашивал).
Насчёт второй ссылки... в общем, это тоже к Nach'у. :D
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
There's old topic discussing RC Pro AM TASing.
http://tasvideos.org/forum/viewtopic.php?t=2779
RC Pro AM 2 supports fourscore, allowing 4P TAS, which greatly extends entertainment possibilities. So I think this game can produce good TAS. Maybe it isn't the best choice for a newbie to pick, but it's not a bad game choice, definitely worth investigating.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
feos, if you're talking about openness, you have to share something too! ;)
goofydylan8, do you use Branches to compare alternative strategies, or you still cater to old habits?
Also, this XLS looks cool, but it only compares old run with new. You should use same thoroughness when comparing your own strategies (like: whether to enter boss room at frame 10000 or 10002), but instead of making Excel tables you can simply branch your movie here and there, using Bookmarks.
BTW, I hope you guys all use OldPPU.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
In (U) ROM players have separate timers on their Clinger Wingers.
For Player 1 the address is $0466, for Player 2 it's $0467 in RAM.
Timer starts counting from 0 when respective toad takes his Clinger Winger.
When any timer hits 160, respective player gains the ability to control his vehicle, and the race begins.
What they didn't test: the green toad (Rash, 1P) gets to his Clinger Winger 27 frames earlier than the yellow toad (Zitz, 2P), because 1P appears closer to his Clinger Winger.
So after the race begins, 2P has to wait 27 frames before he gains contsols. Naturally he gets squashed. Fun fact: it he had to wait 2 frames less (25 instead of 27), he could get away successfully, because his Clinger Winger is located further from starting point.
In (E) and (J) ROM they fixed the bug by leaving only one timer (at $0466). Also this timer starts counting from 0 right at the beginning of the level, and the race begins when the timer hits 251.
Personally I think the first idea in (U) ROM was cool, it's like vehicle warmup of sorts. And in (E)/(J) versions it's just a performance show with scripted events.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
fiskerN wrote:
Feos, посмотрел твой обгон на 54 кадра.
Ну здорово, ваще молодец. Не ожидал что так можно взять ящик. Шедеврально!
Ты, наверное, не очень эффективно используешь Тасэдитор ещё. Вот в этом месяце точно выложу Руководство на русском. А ты пока что больше экспериментируй и меньше экономь ререкорды. Ещё и Дилан обещает через неделю выложить обгон, так что не заскучаешь.
Emulator Coder, Experienced Forum User, Published Author, Experienced player
(724)
Joined: 2/23/2006
Posts: 682
It's really a matter of semantics.
If you want to define lag similar to old console games then modern games don't have lag. Because good games don't sync input/physics with rendering anymore, they run them either in separate threads or simply using separate delta timings (for example simulation can use fixed timestamp, while rendering can have variable FPS).
http://en.wikipedia.org/wiki/Delta_timehttp://gafferongames.com/game-physics/fix-your-timestep/
But if you define lag as a time difference between expected and received result, then probably anything lags from time to time. But nowadays you can't measure lag in frames, you should measure it in milliseconds.