Arcadeshop.de developed hardware?

[FrancoB]

Saw this on arcadeshop.de's front page:

E-Limitator: Budget Day


The developmental period lasted about 3 months, but now we (Arcadeshop and our partner Arcadeforge) are on the verge of releasing a revolutionary but inexpensive piece of arcade hardware. Please stay tuned!

Any one have any ideas what it is?

[Akuma]

My guess is on a key encoder

[FrancoB]

You wasn't a million miles away.

[E-Limitator]

Video games which only support 2 or 4 channels are difficult to control with a joystick which has 8 switching positions. The superfluous signals lead to faltering or stucking of the character in games such as Donkey Kong and Pacman. The motion filter E-Limitator analyzes and filters the users intended movement of the game figure. This way the ease of play of 2- and 4-way games with an 8-way joystick improves noticeable. The E-Limitator filters the inputs of up to 2 joysticks. Included are: E-Limitator kit, a RGB LED, 3 resistors.

[Devil Soundwave]

"Limitator" Engrish from Germany!

[polaris]

http://forum.arcadecontrols.com/index.p ... ;topicseen

bit of talk about it here, i still cant work out what its for , but it seems its aimed for people who swap pcbs rather than for mame cabs

[jonny5]

seems a bit of an extravagant workaround considering with something like an ls-32 its just a matter of removing 2 screws and shifting the gate for 2,4 or 8 way play

[Sobriquet]

And not only that, software limitation really doesn't cut the mustard. See how MAME handles it, you can't really get more clever than the stickykeys algo. It's just not the same as a physical restrictor.

[bencao74]

Hi,

I was browsing through some forums until I saw this message. I`m one the developer of E-Limitator.

E-Limitator let you play 4-way games with a 8 way stick. As many of you know this is a little bit problematic. For example Mario stucks going up ladders when playing Donkey Kong. PacMan and Lady Bug are stucking in walls. Those 4-way games makes definitive no fun when played with a 8 way stick. To swap the physical restrictor is no solution for me and the most other MAME or 60-in-1 setups. So I invented a electrical solution. This solution not only filters the diagonals directions. This task could be done with MAME tweaks and other simple stuff, like some smart wiring with a 7400 ICs.

However we worked out some heuristic rules. This rules optimizes the game feeling when playing 4 way games with 8 way sticks.

We've implemented this filtering modes:

- 2 way vertical (just simple as it sounds)
- 2 way horizontal (just simple as it sounds)
- 4 way
- 8 way

This modes will displayed via a led. It`s also possible to connect a 7 segment display.

Take a look in my developer blog http://bencao74.blogspot.com/search/label/E-Limitator



Currently E-Limitator is announced in several forums. Here's is an extract of the first review:

---snip---
But considering that, they work great I was able to get over 100,000 on Donkey Kong on my first run through, so I was very pleased, Played Tron and was able to seamlessly switch between 4-way and 8-way in mid game (now that is cool!) no jaggies for me on the light cycles!
---snip---

Hope you enjoyed my first post in this great forum.

Bencao

http://www.zurborg.info

[Sobriquet]

However we worked out some heuristic rules. This rules optimizes the game feeling when playing 4 way games with 8 way sticks.

Your blog has lots of pretty pictures, but it doesn't describe the algorithm. Feel free to enlighten me.
Given it is simply taking the inputs as input, and manipulating the output based on latching prior values, I can see no way in which this does something that can't be implemented in software on a host running an emulator. MAME has just such an algorithm for 4-way -> 8-way. If you last pushed right, and then push up-right, it will translate this to guessing you now mean to push up. This way it responds by going up rather than simply disabling inputs when diagonals are hit.
The only other thing I thought of is you might be toggling inputs at higher frequency (i.e. up-right results in toggling between up and right on alternate frames), but based on all the games of that era I really can't see this offering any kind of benefit and would likely result in a very twitchy character.
Without knowing more about the target game, or without feeding back knowledge of the game's current state, I don't see how how you can improve upon that.

It's interesting that you're producing a hardware solution for people to use with real PCBs, but it'll always be inferior to a mechanical (or electro-mechanical) restrictor.

[Spectre]

Hope you enjoyed my first post in this great forum.

Welcome Bencao, I enjoyed your first post immensely, due to your rather unique use of english! You just made it into my sig, however, please don't take offense, as your english is a million times better than my <insert your native language here>.

Some stunning pictures on your blog, those are some of the nicest custom sticks I've seen.

[bencao74]

Some stunning pictures on your blog, those are some of the nicest custom sticks I've seen.

Thanks a lot.

Welcome Bencao, I enjoyed your first post immensely, due to your rather unique use of english! You just made it into my sig, however, please don't take offense, as your english is a million times better than my <insert your native language here>.

No problem, hope you get what I was trying to say. Let's say my english needs improvement. ) perhaps you could add a link to E-Limitator (http://www.arcadeshop.de/Encoder-E-Limi ... t_954.html) in your sig )

[bencao74]

Your blog has lots of pretty pictures, but it doesn't describe the algorithm. Feel free to enlighten me.

Thanks. First of all, the E-Limitator is a commercial product. This algorithm is the core of this application. Opening this algorithm will hurt of course arcadeshop.de. Oj has invested lots of money to bring this product on the market. So I hope you understand that I can`t go here into detail.

What I can offer is the first feedback based on tester reviews and interviews. We`ve presented the E-Limitator on some arcade meetings. All users recognized a improvement in playing with 4-way games with E-Limitator. In this meetings we was presenting Donkey Kong, Lady Bug and Pac Man. Even doubtfull users (Wolfsoft) noticed an improvement in game feeling.

Ok, we're still working on this qualitative approach to evalutate our product.

Currently we`re building a quantitative approach, too. Our idea is to measure points in games with E-Limitator and without E-Limitator. We want to set up a double blind test. The person under test don't know if E-Limitator is activated or not. I`m sure that we can publish some good descriptive statistics. We`re open for suggestions!

MAME has just such an algorithm for 4-way -> 8-way. If you last pushed right, and then push up-right, it will translate this to guessing you now mean to push up. This way it responds by going up rather than simply disabling inputs when diagonals are hit.

I wasn't aware of this. Could you point me to this feature?

Bye,

Bencao

[Sobriquet]

Thanks. First of all, the E-Limitator is a commercial product. This algorithm is the core of this application. Opening this algorithm will hurt of course arcadeshop.de. Oj has invested lots of money to bring this product on the market. So I hope you understand that I can`t go here into detail.

Okay, in this case I would think that this is niche enough, and the development of even a direct clone would be prohibitive enough to not make it worthwhile for that not to matter. Especially if the algorithm is as simple as I think it is. 'lots of money'? really? I don't see any harm in self-promotion, I just want to be sure that people see it for what it is.

Ok, we're still working on this qualitative approach to evalutate our product.

Currently we`re building a quantitative approach, too. Our idea is to measure points in games with E-Limitator and without E-Limitator. We want to set up a double blind test. The person under test don't know if E-Limitator is activated or not. I`m sure that we can publish some good descriptive statistics. We`re open for suggestions!

And the person giving the test is also not aware of which is which presumably?

I certainly am not disputing that it'll offer an improvement over 8-way on a 4-way game. Proving this via whatever scientific means, no matter how rigorous is kinda pointless IMHO. Now if you could prove that the solution is statistically as effective as a real 4-way that would be great - of course that will never be double blind.

MAME has just such an algorithm for 4-way -> 8-way. If you last pushed right, and then push up-right, it will translate this to guessing you now mean to push up. This way it responds by going up rather than simply disabling inputs when diagonals are hit.

I wasn't aware of this. Could you point me to this feature?

Not sure if it's still present. It used to be known as 'stickykey' and is either enabled by default now, or totally removed.
It also used to have another heuristic relating to timings such that if e.g. you tried to depress three buttons at once (e.g. spinning lariat in SF2) it would latch them for a frame or two to give the appearance of depressing them simultaneously due to keyboard limitations on the rate of indicating transitions. that one was known as 'steadykey'.

[bencao74]

And the person giving the test is also not aware of which is which presumably?

Jep, this was my idea to have a clear environment.

I certainly am not disputing that it'll offer an improvement over 8-way on a 4-way game. Proving this via whatever scientific means, no matter how rigorous is kinda pointless IMHO.

I think that statements and interviews are enough to show the effect of this hardware, too. Then had this idea about this quantitative approach to be more clearly.

Now if you could prove that the solution is statistically as effective as a real 4-way that would be great - of course that will never be double blind.

Sure we`ll lose this game, but the question is how narrow. And of course this is not our intention of this product. If a 4 way stick can be used for a 4 way game this is by far the best choice. Our product is applicable for Mame Cabs and tables with 8way sticks playing 4 way games, 60-in1 setups...

Not sure if it's still present. It used to be known as 'stickykey' and is either enabled by default now, or totally removed.

Found this in another thread, same discussion.
http://www.jammaplus.co.uk/forum/forum_ ... ess#149390
Citation start

Originally posted by highwayman

i think mame filters the inputs anyway.

Answer Favouredson
If it does it's not implemented very well because if you use an 8way to play 4way games it's still possible to make wrong moves due to diagonals engaging and selecting the wrong direction.

Citation end

[Spectre]

No problem, hope you get what I was trying to say. Let's say my english needs improvement. perhaps you could add a link to E-Limitator in your sig

Not at all, your english is great, just a couple of ways you made use of it in your original post tickled me. Have updated my sig, hope you approve

[cools]

I can understand the use of this for original (or unmodifiable) hardware, but not for a MAME cab - I wouldn't be surprised to see proper 4 way joystick restrictions make it in to the software.

[Sobriquet]

Just to confirm, MAME does have its logic intact, although it's not quite the way I remember it:

Relevant bits from inptport.c:

Code: Select all

				/* lock out opposing directions (left + right or up + down) */
				if ((joystick->current & (JOYDIR_UP_BIT | JOYDIR_DOWN_BIT)) == (JOYDIR_UP_BIT | JOYDIR_DOWN_BIT))
					joystick->current &= ~(JOYDIR_UP_BIT | JOYDIR_DOWN_BIT);
				if ((joystick->current & (JOYDIR_LEFT_BIT | JOYDIR_RIGHT_BIT)) == (JOYDIR_LEFT_BIT | JOYDIR_RIGHT_BIT))
					joystick->current &= ~(JOYDIR_LEFT_BIT | JOYDIR_RIGHT_BIT);

				/* only update 4-way case if joystick has moved */
				if (joystick->current != joystick->previous)
				{
					joystick->current4way = joystick->current;

					/*
                        If joystick is pointing at a diagonal, acknowledge that the player moved
                        the joystick by favoring a direction change.  This minimizes frustration
                        when using a keyboard for input, and maximizes responsiveness.

                        For example, if you are holding "left" then switch to "up" (where both left
                        and up are briefly pressed at the same time), we'll transition immediately
                        to "up."

                        Zero any switches that didn't change from the previous to current state.
                     */
					if ((joystick->current4way & (JOYDIR_UP_BIT | JOYDIR_DOWN_BIT)) &&
						(joystick->current4way & (JOYDIR_LEFT_BIT | JOYDIR_RIGHT_BIT)))
					{
						joystick->current4way ^= joystick->current4way & joystick->previous;
					}

					/*
                        If we are still pointing at a diagonal, we are in an indeterminant state.

                        This could happen if the player moved the joystick from the idle position directly
                        to a diagonal, or from one diagonal directly to an extreme diagonal.

                        The chances of this happening with a keyboard are slim, but we still need to
                        constrain this case.

                        For now, just resolve randomly.
                     */
					if ((joystick->current4way & (JOYDIR_UP_BIT | JOYDIR_DOWN_BIT)) &&
						(joystick->current4way & (JOYDIR_LEFT_BIT | JOYDIR_RIGHT_BIT)))
					{
						if (mame_rand(machine) & 1)
							joystick->current4way &= ~(JOYDIR_LEFT_BIT | JOYDIR_RIGHT_BIT);
						else
							joystick->current4way &= ~(JOYDIR_UP_BIT | JOYDIR_DOWN_BIT);
					}
				}

[Sobriquet]

One final thing (so as not to further hijack the thread). IF you want to see this in action, do the following:

Code: Select all

mame.exe -debug dkong

Then, go to Debug->"New Memory Window" from the debug window, and enter the following address: "7c00".
Hit F5 to start the game running and move the focus to the game window with the mouse.

Now, if you push Down, you'll see that the game sees an '8'. Up and it will see an '4'. Left is '2' and Right is '1'. This is after MAME has processed the inputs and directly what the game sees.
With an 8-way joystick, if you pushed diagonals you'd see the sum of these inputs. e.g. UpLeft is '6'. As the game is flagged as needing a 4-Way joystick pushing UpLeft will take either Up or Left depending on which happened slightly later. If you're really lucky, and your keyboard can indicate both keys changed state in the same frame, then it will randomly pick either Up or Left.
Now try pressing Left + Right, MAME blocks this and returns 0. Try pushing Up + Left + Right - the Left and Right will always cancel leaving the Up.

For 2-way only 2 of the directions will be mapped to the game, but MAME still blocks Left+Right or Up+Down.

[bencao74]

Have updated my sig, hope you approve

I do take a look at my sig.

@Sobriquet

great post. Thanks for this tipp. The algorithm proposed seems powerfull enough to cover this "stucking" problem. I`m wondering why people have still problems with their 8 way stick setups...

[Spectre]

lmao, nice work

It's "sticking". They stick to the walls, I am sticking to the wall, I was stuck on the wall.