Coder shows possible way in which some elements of Street Fighter 2 might be possible

[Amiga1992]

Originally Posted by Steril707

Never played the Amiga version back in the day, had an SNES at that time already.
But I am mighty impressed by the Amiga version, I can tell you.

I'll add this much and hardcore MK people always told me I was crazy for it: I prefer playing MK2 on Amiga with 2 buttons than with 5 buttons on Megadrive or anywhere else.
The super complicated moves become a lot easier and you can block like in any other fighting game (" block" button? gtfo). Made the game a lot better for me.

The one thing that is bad in the Amiga version, and not sure if it is the same in the Megadrive version, is that there's some problem with the AI when you play it at certain difficulty levels. Like the extreme hard level was super easy because you could fool and lock AI enemies into a silly loop.

[jotd]

I completed both MK amiga (bought the original) and MK arcade (with 1 credit). I remember buying this game in easter 1994 and played it to death on my 2MB A500.

What happens a lot in amiga conversions is that the soundtrack (Allister Brimble on MK IIRC) was better than the arcade. Just because it was possible.

Sure it was impressive how it was converted. The PC version was much closer to the arcade but hey they had 4MB and 256 colors...

[redblade]

Originally Posted by mcgeezer

Here you go, laying this one to rest with proof it can be done on Amiga AGA stock at 50 frames.

[ Show youtube player ]

Roughly 120 colours with 9 palette swaps on Y Axis only.
2 128x128 hardware sprites at 16 colours
Multi-parallax scrolling with animation
Chipmem use:
128kb for mirror table
100kb hardware sprite buffers
160kb foreground and background screens
120kb sprite tiles constructing 64 128x128x16 colour sprite frames
50kb program code (using my framework)

Resolution is 256x224, using SNES assets


around 150 scanlines (probably can be improved.

Think I'll have a few beers now.

Congrats, amazing work, the only problem was that I couldn't read the source .

Is CD32.asm code for reading the CD32 control pad?

[mcgeezer]

Originally Posted by redblade

Congrats, amazing work, the only problem was that I couldn't read the source .

Is CD32.asm code for reading the CD32 control pad?

Thanks. CD32.asm is for writing to nvram, nothing joypad specific in there.

[Reido]

That's really impressive McGeezer!

[d4rk3lf]

Originally Posted by Akira

I'll add this much and hardcore MK people always told me I was crazy for it: I prefer playing MK2 on Amiga with 2 buttons than with 5 buttons on Megadrive or anywhere else.

I keep repeating that.
Name me one move I can't pull out the same speed with Amiga 1 button, as opposed to arcade with 5 buttons.
The point of buttons is make game play easier, not complicate it (imho), and what they did with MK1&2, and how they transferred gameplay is pure art.

Well, to be honest, if they ported MK 3, then it would not be possible, with all combos involved into gameplay.

Have you tried Mortal Kombat 9?
It perfectly transfers graphics, feel, design, atmosphere from MK 1,2 (and party 3). All the famous stages too. And the "story mode" is strictly tight to first 3 MK's.
If you didn't - Can't recommend it enough.
I am not too impressed with last 2 versions of Mortal Kombat though (10 and 11), but 9 is a masterpiece.

Originally Posted by Akira

The one thing that is bad in the Amiga version, and not sure if it is the same in the Megadrive version, is that there's some problem with the AI when you play it at certain difficulty levels. Like the extreme hard level was super easy because you could fool and lock AI enemies into a silly loop.

I think I've read in Richard Costello interview, that there's no difference on AI between arcade and Amiga (not sure about genesis or snes), simply because he didn't touched that part of the code, but basically only copy-pasted it.

But MK2 is infamous for his AI, because it cheats a lot on harder levels (just automatically react to whatever your moves are (and often, you are unable to counter react)), so it causes a lot's of frustration.
Yeah, there's some workarounds on this... but still...

Also, AI work that way, that even if you choose "very easy", it grows up on the ladder. So after 2-3 fights it switches to "easy", then to "medium", and so on.
That's why it's not easy for most folks, to finish it, even on "very easy" level.
Personally, I got trauma from Baraka, back in the day

Originally Posted by jotd

I completed both MK amiga (bought the original) and MK arcade (with 1 credit)

That's impressive.
It gets very hard to me when I get to the point that I must fight 2 fighters.
Although, I haven't played MK1 that much at all, because MK2 was always my favorite.
------------------

Oh, and sorry everybody for OT.

@mcgeezer
Amazing work.
Thank you for finding time to do it.

[vulture]

Yup, pressing a button to block is just so counter-intuitive, the Amiga method is much better.

[Tigerskunk]

@d4rk3lf:
I preferred the AI in MK1 over MK2 any day.

MK2 AI resorted to cheap throws every time you went near your enemy after a few levels. Horrible, really. No real person would play like this.

[mcgeezer]

I put an ADF in the zone of the little Street Fighter demo I put together for those wanting to take a look.

Interestingly I checked the available chip ram remaining and it's a whopping 1.2mb making the game or a derivative of it easily viable as a port of the SNES version.

Big job though, which I estimate at about 8-12 months work, I'm not really a fan of the game so it's something I would pass on.

The good thing for me to come out of this is that I was able to improve my dev framework by including some new routines around making dynamic copper lists which has been on my to-do list for a while.

[DanScott]

Nice work.

In my mind, there was never any question of if an A1200 (or CD32) could do a port of SF2... it was just how it would be approached.

Your approach is different from the one in the video originally posted, although both approaches do have their merits.

With 64 bit 16 colour sprites, and 2 x 16 colour playfields, both methods could produce approximately the same result, although I do feel the "sprites for backgrounds" method *may* give more flexibility for the character playfield (ie. fireballs / hit effects etc.. appearing in front of the characters)

[mcgeezer]

Originally Posted by DanScott

Nice work.

In my mind, there was never any question of if an A1200 (or CD32) could do a port of SF2... it was just how it would be approached.

Your approach is different from the one in the video originally posted, although both approaches do have their merits.

With 64 bit 16 colour sprites, and 2 x 16 colour playfields, both methods could produce approximately the same result, although I do feel the "sprites for backgrounds" method *may* give more flexibility for the character playfield (ie. fireballs / hit effects etc.. appearing in front of the characters)

I'll probably have to disagree with you Dan which was partly the reason I wanted to effectively debunk the video. Using the sprites for background will work until it comes to doing the player sprites with the blitter/cpu at which point things become complicated - hence why plotting the player sprites wasn't exactly covered in the video.

[Retro1234]

the problem is over the years so! many of these threads etc have been made - you can do this and that etc but it's all theory no-one has ever given hard proof except master484 it's all been hot air, until now thanks to mcgeezer.

[Tsak]

Originally Posted by mcgeezer

I'll probably have to disagree with you Dan which was partly the reason I wanted to effectively debunk the video. Using the sprites for background will work until it comes to doing the player sprites with the blitter/cpu at which point things become complicated - hence why plotting the player sprites wasn't exactly covered in the video.

Nice work

However would you like to elaborate further in regards to which part from that video is debunked? You say doing the player sprites with the blitter/cpu things become complicated. What do you mean? No enough blitter time left? Issues with ordering? Something else?

Other than that I should note that Dan might have a point with the projectiles and effects ordering. How would you approach this issue?

Another question is in regards to frames, character tiles and ram. Can you still have the characters saved and stored split into 16px wide tiles/parts and reconstruct them as 64 wide sprites or to use sprites you need to save each frame/part only as a 64 wide one? I was under the impression that when you choose to use f.e. 64 wide sprites you can only have the parts saved as 64 wide as well, so f.e. even if a character frame is much less (say 32px wide) you are still forced to save it as 64 wide non the less.

*edit: by 'saved' I mostly mean the chip ram it needs to occupy (which is the main issues here), not the disk/storage space.

[jotd]

so it's established that the video is a fake now?

@d4rk3elf: I completed MK I only once with one credit, on a real arcade machine. And many times (AFAIR) on the amiga. When you don't lose any credits, the game difficulty becomes harder and harder, but with properly timed kicks it's possible (now I'm not able to do that again)

And MK2, yeah I would have loved to play it properly but I always thought it was insanely hard. Each time I play I get my ass kicked.. although I saw it with my very eyes a long time ago in the arcade: local boy mastering the game with Shang Tsung over and over, beating every character with mostly jump kicks / unlocking secret fighters / morphing in a different character everytime to perform a different fatality/babality... impressive.

[Tsak]

No, it's not a fake. But it does focus (as McGeezer mentioned) more on backgrounds reconstruction than characters.

[Retro1234]

what part is real and what part is PC animation.
thanks

[roondar]

Didn't the authors of the video already confirm that all but the final segment where they used the extreme horizontal Copper re-colouring was real?

Here's the relevant part of the text they provide on YouTube in the description:

Originally Posted by YT

- Some people have shown concern regarding performance, this demo is running on stock Amiga 1200 currently at 60fps. (we still have obstacles to overcome). The reason for this is that the whole stage background is drawn once, since movement is handled by the copper no redraws or cleanup routines need to be present, except for the crowd (which do not update every frame)... Since the character are on a separate playfield the don't need to cleanup/redraw the background when characters are moving (which is common when blitting objects on the same playfield). Character one is drawn first on screen, and due to this doesn't require a blitter mask, and no cleanups are required... Only character two requires this. This is how we achieved 60fps.

Originally Posted by YT answer to my question on Copper colour changing

I have been talking to neoman about the palette swaps, it's still not implemented in the "real amiga engine". In that version we still uses the 16 color boat from the beginning of the video. while the last part of the video I am showcasing the graphics through a webgl engine mirroring what we have done on the Amiga. (we're doing it like this since it is much faster to iterate ideas in the webgl engine, before committing to full blown assembly implementation)

Edit: I forgot to add the bit where they talked about the Copper colour changes, added that now.

[Retro1234]

ok thanks so the seconds at 26:18 ?

edit:::26:18

so the boat and the characters are large bobs? really?

[roondar]

I don't have the time stamp, sorry. But it's the part where the boat on the left has a crowd in many more colours than the first part they show.

[Retro1234]

the boat is moving - it's huge bob?

edit: sorry no that's not what was said I thought earlier in the video he said it was a huge bob
I only skimmed the video, just looked so fake to me sorry if this is not the case, probably best he uploads it for proof

Thanks.