Final Fight: Enhanced

[gimbal]

Ditching Streets of Rage for Final Fight because you call Final Fight the "big kahuna"... oof that is going to hurt some people

Imagine if a playable Final Fight Amiga actually manages to appear. I guess we can close up shop here, right? Nothing left to talk about on EAB So far it looks like a port of Final Fight that is very playable. Absolutely love that music too.

[Immortal]

Originally Posted by Brick Nash

Hi folks,

I started learning Assembly language last summer, so I thought I'd post a preview of what I've been doing, as I'm taking a break from it for a while. I was originally going to try and re-do my Amos Streets of Rage demo, but I thought I'd have a go at the Big Kahuna and try Final Fight.

It's only one level and two enemies, but it use 2 buttons and it plays ok.

Nice work, but dont get me wrong, if this was done a few years ago, i have no doubt it would be great, but i think doing this version right now, it could be a waste of time, since recently a developer came up with is own engine, that can produce arcade perfect street brawlers on the Amiga.

I believe that this engine is the future, checkout the number of colours, the number of simultaneous enemy sprites on the screen, parallax scrolling backgrounds, and the speed, its simply mind blowing!


[ Show youtube player ]


[ Show youtube player ]


[ Show youtube player ]

[jotd]

there are no performant "engines". Only people who know their code and reuse it from game to game.

custom code to get maximum performance.

adapting an arcade version is different from doing an original game on the amiga. Arcade doesn't care about amiga constraints.

Doing an amiga original game is done with the amiga in mind at the very start. Easier (but Pixelglass new fighting game looks terrific I'll buy it as soon as possible)

About the linked videos: still waiting for a proof that they're running on a real amiga. So far, no proof. Disregarding.

I can create a demo in SDL on Windows, record it, and post it on YT pretending it's running on amiga...

Brick Nash I say: get it done, at worst you'll learn something, and you'll have some working code for other projects in the end.

[d4rk3lf]

Originally Posted by Immortal

Nice work, but dont get me wrong, if this was done a few years ago, i have no doubt it would be great, but i think doing this version right now, it could be a waste of your time, since recently a developer came up with is own engine, that can produce arcade perfect street brawlers on the Amiga.

I believe that this engine is the future, checkout the number of colours, the number of simultaneous enemy sprites on the screen, parallax scrolling backgrounds, and the speed, its simply mind blowing!

That's really not a fair comparison.
"The engine of the future" by it's own author words, was never tested on a real A1200, but on a Winuae (that is by default faster, then real thing, according to many users), and with additional 64MB FAST RAM, or more!
Then, the author (in the youtube comments) said that he haven't used blitter at all... so that's pretty wasteful on the resources to be called "engine of the future".
With all due respect to the author and engine, of course.

On the other hand, Brick Nash's engine is much less demanding, and the demo he showed us, actually have some (very proper) gameplay, not just walking around sprites.
And his plan to proceed, and further optimize it for OCS machines, and slowest Amiga's, is very, very brave.

[gimbal]

Originally Posted by jotd

there are no performant "engines". Only people who know their code and reuse it from game to game.

Thank you. It really is time to stop it with the engine talk, that was all the rage two decades ago and even back then it was sweepingly inaccurate at best.

[jotd]

Even if engines aren't my cup of tea, Scorpion looks powerful, though. Needs a powerful machine, but the alternative for some devs is : no games done (Considering nice WIPs ongoing with this engine, check the Double Dragon episode for instance: started in C, going nowhere, resumed with Scorpion: making good progress).

Specially because the 8 way scroller is implemented when you start. Not a piece of cake considering what chaderrack wrote about Black Tiger.

But if you target a lowspec amiga, you have to go full specific, C / asm and it's not enough: optimize your code / time the DMA waits, etc... And when you know asm & C, you probably don't want to use Scorpion.

It's difficult to do speed hacks with an engine. Sprite management is not that easy with shared palettes, blitter operations can be intermixed with CPU copies but only sometimes...

Coders weren't idiots in the 1990's (some were brilliant) and had a lot of time as they had full time jobs on that. Difficult to compete against that.

The advantage we have now are: WinUAE & tools, windows cross development tools (which are the reasons I started coding games on the amiga)

[mcgeezer]

Originally Posted by jotd

that looks good. Talented Richard Aplin failed to do a proper conversion in 68000, I don't suggest you try to stick to 68000 if it's too difficult. The game is very demanding.

Make it AGA, add extra colors (like 16 => 32 or 64) and everyone will be happy

100% this.

Do it on stock AGA and I myself will carry you through the gates of Valhalla. You shall ride eternal. Shiny, and chrome!

Seriously though, really well done. Don't attempt it on A500, you'll run out of memory very quickly.

[roondar]

Originally Posted by Brick Nash

This pretty much sums up what I was trying to do. I'm not looking for arcade perfect, but rather just a slightly enhanced version of the original OCS/ECS Amiga game which I played a lot back in the day.

I'd be perfectly happy if I could get it to round about the same speed as Richard Aplin's game - even with the slowdown (which never really bothered me). The Background replacing is such a resource hog though, and I can actually get 5 enemies on screen with no slowdown on a cycle exact 68000 setting with it turned off, but obviously the tiles don't get cleaned.

I even tried duel playfiled mode so Bobs and Backgrounds had their own layers and I didn't have to do any tile replacing, but the reduction to 7 colours for each playfiled just looked too rough.

BTW, I though Richard did a great job on the 1991 game, but obviously he got a raw deal from US Gold in terms of time and resources.

It's good to see you write this, starting with realistic expectations is a very good idea

As for the platform discussion, I'd say that depends on what you want to achieve. If you have no feeling for AGA and can live with the many compromises an OCS/ECS port means (aka fewer colours/speed/far fewer animation frames) then why not? Just don't expect anything even near a 1:1 port and make it for yourself first and others second. It's your project after all

That said, jotd & mcgeezer do have a good point in that doing Final Fight on OCS is not going to be easy (and yes, memory will be a massive constraint - those huge sprites add up quick!). A 1:1 port is impossible and even a cut down one will still need some pretty big compromises to be able to work. The way I see it, apart from the gameplay aspects (which I understand aren't very close on to the Arcade on the Amiga version), the original Amiga version wasn't at all bad, given the vast gulf between the arcade machine and the target system. It's actually quite cleverly done in many ways.

[Girolamo85]

the improvements are impressive, keep going!

[Brick Nash]

Thanks for the feedback folks, it's much appreciated.

I can totally see the advantage of going to AGA, but I just truly feel that it's my bad code which is the main culprit in what's slowing things down on 68000. I mean, it's literally grinding to a halt with one enemy, but fine with none, and it all points to the background tile replacer which is likely over drawing or using too many loops.

It's not even the actual Blitting part which is slowing things down. I can turn that specific part of the replacing routine off and it's the same slow speed, so something in the logging of tile coordinates into the table is taking too long (or something). Unfortunately my inexperience means that I just don't know what to look for to improve things at the moment.

I'm quite happy to post the source code for folk to look at, which would be much appreciated. I'm not looking for anyone to do the work for me, but maybe a more seasoned eye can see something bad that's happening.

Also, please don't think I don't have extreme gratitude for the advice I'm being given by the experienced elders of the forum. I'm absolutely not dismissing the prospect of going to AGA, but I think it would leave me empty if I didn't explore every single avenue of getting it on to the OCS/ECS systems that I loved so much as a boy. Even if it ended in not being good enough, at least I'd know and be satisfied to move on.

I hope that makes sense, and it doesn't just seem like I'm being stubborn or ungrateful?

[jotd]

maybe you should start a github on your project. Then people can look at some parts you find suspicious.

[d4rk3lf]

Originally Posted by Brick Nash

I'm quite happy to post the source code for folk to look at, which would be much appreciated. I'm not looking for anyone to do the work for me, but maybe a more seasoned eye can see something bad that's happening.

May I suggest that you open a separate thread on Final Fight technical talk, on Coders, general subforum, or Coders. Asm / Hardware?
And keep this one for updates and general talk.

I think it would be beneficial to the project.

[Brick Nash]

Originally Posted by jotd

maybe you should start a github on your project. Then people can look at some parts you find suspicious.

Great idea! I'll set one up.

[Brick Nash]

Originally Posted by d4rk3lf

May I suggest that you open a separate thread on Final Fight technical talk, on Coders, general subforum, or Coders. Asm / Hardware?
And keep this one for updates and general talk.

I think it would be beneficial to the project.

Absolutely. I know everyone's not into the technical side of things, so don't want to bloat this thread.

[gimbal]

It'd be a little random to do it specifically for this thread though, it happen all the time that a discussion of a game someone is making dives into the technical side. I don't see a problem, the game is still in the development phase after all. Or perhaps prototyping phase is even more accurate. I would be disappointed to see no discussions of bitplanes and "to blitter or not to blitter" and whatnot.

[mcgeezer]

Bricknash, pop over to the amiga game dev discord and we’ll enrol you in the dev chats. Lots of people there will br happy to help and i’ll certainly take a look at what is going on.

[d4rk3lf]

Originally Posted by gimbal

I don't see a problem

Just to clarify..
I didn't suggested separate thread because technical talk would bother me (not at all, it can become all technical if you ask me).
But I think some coders are maybe not visiting this part of EAB, and also, our chit-chat about game overall, might bother them.

Either way, that discord channel also sounds great.
(although, I personally favorite forums, because solutions might help someone else).

[saimon69]

Originally Posted by mcgeezer

100% this.

Do it on stock AGA and I myself will carry you through the gates of Valhalla. You shall ride eternal. Shiny, and chrome!

Seriously though, really well done. Don't attempt it on A500, you'll run out of memory very quickly.

Richard did a good job on A500 with no fast RAM, though, so if he aim for the 512 expansion should be able to handle most of the stuff - however i would recommend dynamic palettes so that ugly green spots like in the train does not happen that often - plus is one of those game that even at 25FPS look good (blasphemy!!!)

[trydowave]

Originally Posted by Brick Nash

Thanks mate! I was hoping to get it on to a disk image and let folk have a play, but that'll require more learning which I don't have the energy for right now.

About the speed - I'm positive it's only slow on 68000 because I'm coding things wrong or badly and not the system's fault. As for an AGA version, the thing is I never had an AGA machine back in the day so I've no connection to them, and hence no real interest in developing on one. Upping the specs just to make things run faster would feel like a defeat

I could upload the source code, but not sure if anyone's interested in that side of things.

So glad your sticking with OCS/ECS rather than AGA. I agree with you.. thats the whole point; to see if a decent version of FF was possible on the Amiga model that most people had. Im not so interested in an AGA version as these days weve all got FF on hand via mame etc.

Equally, ive never understood running arcade games on a Vampire. I dont see the point.

OCE/ECS all the way

Good luck on the project mate. Its looking excellent. Along with SF2 and Strider this is the game ive desparately wanted to see done right on the Amiga. A proper scrolling beatem up on the Amiga! Apart from Double Dragon 2 i cant think of any good ones.

[mcgeezer]

Originally Posted by saimon69

Richard did a good job on A500 with no fast RAM, though, so if he aim for the 512 expansion should be able to handle most of the stuff - however i would recommend dynamic palettes so that ugly green spots like in the train does not happen that often - plus is one of those game that even at 25FPS look good (blasphemy!!!)

A slow ram expansion board doesn't give you that much to work with, sure your code and tables can sit in the expansion but without serious (and I mean serious) work the gains are minimal, maybe 150kb.

Richard Aplin did an amazing job on the port, I have the utmost respect for him.