Greatest HuCard of all time - Poll

[awack]

Some of those frames of animation you count aren't specifically a 'pixeled' frames (they use palette and/or multiple sprites overlaying via script or such).

Absolutely, in 2700/2900 total, i include most of that stuff, not all but almost all.I subtract 300 of those to get a total of 2400/2700, if i subtract 400 frames, thats removing 1 in every 6.8 frame.. thats to many in my estimation.

In contrast, the totals i got for Dracxx(1005/1080) SCIV(650/680) all include color cycling, mirroring, and sprite shifting.

Heres an example of Bloodlines





We also have to remember that not every frame is equal, some of the largest sprites are completely missing or downsized.


Rondo


Dracxx



Not only that, but check this out, looking at the shot below, you will notice that the Dracxx  dracula is just shifting sprites from left to right to simulate animation...right above his elbow, his arm shifts to the right a few pixels and then back, which is why you see his arm farther away from his body and then closer....now look at Rondos Dracula, from his upper body to his lower body to his face, those are all redrawn .

Still, i counted all of these as unique.


So, in the end, when you compare these games in detail, you realize that rondo is far, far more impressive than even the 2700/2900 versus 1005/1080 frames would suggest.

It might have come out in March '94, but the card was supposed to come out in fall '93 (design wise, it was final and ready, etc). It was delayed because of production reasons

I read somewhere there was a fire at the plant that manufactured the memory and thats what set it back.

[spenoza]

For all the mistakes NEC and Hudson made with the PCE hardware, I'm sure they were watching the market and aware that delays were likely. I'm not convinced the delay was last-minute and totally unexpected. Still, given the price up the upgrade AND the hardware requirements behind it, not to mention the quickly rising star of the SFC, there were tough decisions to be made. Nintendo did not make things easy on NEC and Hudson.

Also, Super SF2 used a new engine on brand new arcade hardware, not just a simple tweak. It was the debut of the CPS2 hardware and so the development would have been based on two options: totally reprogram the new SF2 for the PCE or do a massive retro-fit on the old engine.

I wonder if Super SF2 would really have been a solid boon to the Arcade Card and PCE in general. The SFC was already doing so well in the market, and access to Square titles and the great stuff Nintendo was putting out meant that the SFC Super SF2 would have been more than adequate even if an ACD release was better in some regard.

 It might have come out in March '94, but the card was supposed to come out in fall '93 (design wise, it was final and ready, etc). It was delayed because of production reasons (IIRC, relative to cost.. the ram ICs). The Arcade card started out mid-to-late '92, was upgraded and working/finalized mid '93 (the original arcade card has missing a lot of key features and required the CPU to be in slow mode (1.79mhz) to access the data ports). Art of Fighting was finished by fall '93 and ready for release. I can't say for other titles, as I don't have access to the source code for that project. But it wouldn't be unthinkable for other titles as well. That puts a different time frame to all of this. AoF came out in late '92, and it would've been released almost a year to date for ACD (though they had the direct source code for the Neo Geo version. I can see it in the comments and parts of the translation to 6280 code). SF2 game series were mostly upgrades to existing engines. Much easier to tweak and modify and existing engine that to build a game from the ground up. The turn around/dev time to produce the additional games in the series wouldn't have taken as long, since they already had SF2 CE developed and released. More profitable from that standpoint, and most profitable to produce a disc than a hucard with a custom mapper and lots of rom. But then again, your target audience is smaller. But, the more A list games you have for the new format - the more popular it becomes and the base increases (every potential Duo unit and user, was a potential buyer for the upgrade). They also should have played the benefit bi-compatible card more often of the ACD (SCD titles the loaded much less when using the ACD).

[fragmare]

Yeah, I see what you are saying. I was looking at it from even if each sprite was 1meg each, the system with SCD card + console ram is 2.5meg.

The Super CD-ROM system used 2.0Mbit (256KB) of RAM, not 2.5Mbit.  There was 64KB built into the unit and the Super System Card added another 192KB.  In a Duo, the 64KB and 192KB were both simply just built onto the system board.

It really doesn't matter how you divide up the way you believe the HuCard allocated the data.  It doesn't change the fact that each full character animation set used around 1Mbit (128KB) at a MINIMUM.  I'm looking at this from a PCE pixel artist perspective... if someone told me to try to fit two SF2 characters into 256KB of RAM and still have room for code, backgrounds, and voices, I'd tell them they basically have three options: 1.) Compress the graphical data and suffer slowdown, 2.) Drop some animation frames and suffer choppy animation, or 3.) f*ck right off.  There's just no way all that data is getting crammed into 2Mbit of RAM.   

[SignOfZeta]

There isn't a huge difference between CPS1 and CPS2 (see the CP Changer port of Street Fighter Zero to see what the CPS1 could do).  I don't see why the existing SFII' couldn't just be ported to the Arcade Card and have the Super Turbo assets added to it. It's just five more characters/BGs and one super move each.

As for it being worth doing financially...again, that's hard to say. I'm sure it wouldn't have been worth it for Capcom, but it might have been worth it for Hudson or NEC to do it.

The absence of any decent port of SSFIIX (only 3DO and Marty ports existed) until years after it's arcade release is puzzling. Sure it would have sold in as large of numbers as SFII'Turbo, but neither to %95 of games in general, and it wouldn't have been very expensive to develop.

[Arkhan]

There's just no way all that data is getting crammed into 2Mbit of RAM.   

3 FRAMES! STANDING, HALF OF PUNCH THROWN, ALL OF PUNCH THROWN, ETC.

WE CAN FIT IT ON A 5.25" FLOPPY.  C64 STYLE.

That would suck so bad.

[Black Tiger]

I would assume that any port of SSFIIT would have been based off of SSFII for SNES.

[Bonknuts]

Yeah, I see what you are saying. I was looking at it from even if each sprite was 1meg each, the system with SCD card + console ram is 2.5meg.  There are 12 characters.  20meg card - 12meg = 8meg for the rest of the game.  There are 14 levels(with bonus rounds), battle over screen with lots of talking (counting 10 to 1) that loads separately, 12 character endings that load separately, the opening screen animations, the setting screens and character selection screens, and so forth.  You dived that 8 meg by all that other stuff.  Are you loading .5 meg of stuff for each of just the 1/14 of the levels?  If so 14 x .5 is 7meg +12meg for characters that is 19meg.  That leaves 1 meg for all the other endings, openings, selection, etc.  And still leaves room for it to be theoretically doable.

 It's 2.5megabit *IF* you treated ADPCM ram as slow ram (which is fine, just not for code. But SF2 still needs 'samples', so it's almost useless to talk about ADPCM ram as holding anything else but audio. I.e. my statement that you can't play ADPCM samples and read other parts of it for non-audio use at the same time). Excluding ADPCM, it's 0.5 megabit for CD 2.0 and additional 1.5 megabit for SCD 3.0 upgrade, for a total of 2.0 megabit SCD 3.0 CDRAM.

[Vecanti]

Well, I think there should have been a SFII laseractive version anyway.  Floor and sprites are SCD and the backgrounds laserdisc.  Kickass!

[SignOfZeta]

I would assume that any port of SSFIIT would have been based off of SSFII for SNES.

I can't imagine why they would go that route when they already had the bulk of it done for the SFII' HuCard. Fundamentally it's almost the exact same friggn game. Add some new characters and backgrounds, different endings, supers, make Ken's shouryuken burn people...thats about it.

[spenoza]

I would assume that any port of SSFIIT would have been based off of SSFII for SNES.

I can't imagine why they would go that route when they already had the bulk of it done for the SFII' HuCard. Fundamentally it's almost the exact same friggn game. Add some new characters and backgrounds, different endings, supers, make Ken's shouryuken burn people...thats about it.

SSF2 is actually a bit more different than simply that. I'm pretty sure there were some fundamental engine changes, even if it doesn't appear that way to the casual player.

[SignOfZeta]

I'd like you to explain what you are talking about because I'm not a casual player of Street Fighter and I don't see much of anything fundamentally different. They changed the sounds a bit, added assets, it counts combos and reversals. Does the whole game need to be rebuilt to do that? I'd be surprised if anything but the first of the CPS versions were ground-up constructions.

EDIT: It occurred to be that the biggest difference between builds of SFII on CPS as compared to CPS2 were probably just related to the way the boards were made. As far as I can tell the whole point of switching to CPS2 at all was to combat piracy. There were some other minor advantages, but they weren't that big of a deal. Kind of like the PSPGo. Of course, this tactic worked for Capcom, but didn't for Sony.

[spenoza]

I'd like you to explain what you are talking about because I'm not a casual player of Street Fighter and I don't see much of anything fundamentally different. They changed the sounds a bit, added assets, it counts combos and reversals. Does the whole game need to be rebuilt to do that? I'd be surprised if anything but the first of the CPS versions were ground-up constructions.

Well, I don't think either of us has enough insider knowledge to truly assert one way or the other. I'm pretty sure, but I don't really have anything to hold up, which seems to be the same for you. So I guess we're just going to have to agree to disagree pending any actual evidence either way.

[fragmare]

Okay, I did a little more number crunching and I can say that unequivocally, absolutely, and without question there's NO way to fit all the SFII' CE data into 2Mbit (256KB) of RAM, even if you used the 64KB of ADPCM RAM for more data.  Take a look at the sprite rip of Chun Li below from the Genesis SFII:SCE.



I count 96 unique character frames.  So (96 frames x ~2KB per frame) x 2 characters = ~384KB.  Also consider this may or may not be a complete sprite rip, so if there are additional frames not in the sprite sheet, that adds more to the total.

Then I looked at the backgrounds from SFII' CE and they seem to use anywhere from 800-1000 unique 8x8 tiles.  Each 8x8 tile uses 32 bytes of RAM/ROM, so that means each background image uses roughly 24KB - 32KB of space.

So we're already up to ~416KB... Add in another ~8KB for misc. background sprites, another ~16KB for music/sound effects code, and sampled voices, and a possible ~32KB for game code and we arrive at ~472KB... WAYYYY more than would fit in the Super CD-ROM systems RAM.

[Black Tiger]

I would assume that any port of SSFIIT would have been based off of SSFII for SNES.

I can't imagine why they would go that route when they already had the bulk of it done for the SFII' HuCard. Fundamentally it's almost the exact same friggn game. Add some new characters and backgrounds, different endings, supers, make Ken's shouryuken burn people...thats about it.

That's what I mean, all thr 16-bit console ports are built off of the SNES World Warrior engine. SFII' PCE obviously used new assets found in the other CE/Turbo ports. If NEC did do a port of SSFII or SSFIIT, it would probably look like the Genesis and SNES versions of SSFII, instead of a new unique port based off of the arcade version.

[Bonknuts]

I just looked at the SF2 hucard rom in TMOD2. E. Honda is 161k and Blanka is 143k. Assuming you had those two chars for a level, that's ~304k just for the frames. That doesn't include the code (game logic) or animation tables and other LUTs. And thoses frame sets are *packed* and optimized for size. No wasted space there (although uncompressed).