Paprium: The Official Thread Mk 2

[SEGA-Jorge]

I really don't see what would keep Sasha from releasing more entries in the series. The engine seems to be in place, aside from the BG's, he'd be essentially laying out the screens and, assuming the scripts are written, adding the text and scripting the paths/interactions. It's certainly not a small task, but I assume the process is pretty efficient now that he's done it once.

Frankly, I'd love to see sprites for the bg's, but the pre-rendered stuff is serviceable.

I wonder if Sasha would be open to letting people make their own interactive novels with the engine.

[Sik]

Wasn't that game made with GINCS?

[Zebbe]

He doesn't agree with Fonzie on some things.

Yes, it was made with GINCS, an open source engine made by some other guy.

[matteus]

Actually on that last quote... I know for a fact Fonzie was always horrible at the job. Saw this first hand some years ago, the guy really has no idea how much effort it takes to make something, and once he got an idea stuck in his mind there's absolutely no way to get it out of him even if you can show there's an obviously better one. You'd think that after Pier Solar the guy would have learned by now.

Indeed he does but that's unfortunately due to him having ocd.

[Barone]

Oh man, that was pretty funny. It looks like a really short interactive novel, minus the good writing... (I'm not trying to be critical here, this is just what I seem to hear from most reviews. Fun distraction, but not exactly wonderfully written.) Still, I can't really scoff at the product. We get so little homebrew on the Genesis that I kinda have to latch onto anything that shows up.

I really wish there was some sort of way to breath life into the homebrew scene for the genesis via some sort of middleware. If someone could just make a really robust Genesis game engine that allowed for rapid prototyping of Genesis software via more current methods, which converted it to stuff the genesis could play...

For example, I am sure that underneath Golden Axe, Alien Storm, and Streets of Rage is a very versatile engine that allows for pretty rapid prototyping of beat-em-ups. Perhaps a thorough hacking of SOR3 or GA3 would allow for a suite of tools that make hacking those roms much easier/faster/efficient.

do that again with the Sega 2D platformer/action games using the Team Shinobi stuff (Revenge of Shinobi, Shinobi III, Shadow Dancer) Etc.
A stripped down Technos or Compile engine for shooters. All this stuff could be like a game genre template... This is so much pie-in-the-sky nonsense though. I really doubt this could work, considering memory allocation, ancillary/misc game mechanics that are not in these games.

For example, I don't think a thorough hacking of Bloodlines would create a completely bad port of Curse of Issyos. In fact, I imagine a rather serviceable port is possible. But the world map would pose a problem.

Same with a rom hack of Ghouls n' Ghosts being turning into a homebrew port of Maldita Castilla.

Point is, Loco Malito games need to be on the Genesis...

It's MUCH easier said than done though.

Some common misconceptions I also had before studying the actual source code of several Mega Drive games:
- "The AAA titles carry some awesome piece of code" -> Not exactly. A LOT of the games I've looked into, including some Sega big shots, have far-from-optimized code. Some games barely make a rational use of the 68000's registers, for an example. Many of them also have a very poor DMA queue implementation (which is crucial for animation, for an example).
- "Japanese developers were the best!!!" -> As tomaitheous had already pointed in the past, you usually find better code in European-developed games than on Japanese-developed games.
- "Let's pick the most advanced games and mod them!"-> Usually, the later iterations of a game series or the hardware-pushing games are the most difficult ones to mod. There's a reason behind all these hacks you see for SOR2 and not SOR3, Sonic 1 and not the later ones or Ristar. The more advanced a game seems to be, the more likely it is to be a bitch to change anything, since the level of customization for their very own assets and design details is insane at times.
- "Let's produce game in batches like the Indie community does for PC."-> Don't forget the Mega Drive's hardware is super limited and most of the game factory tools are unlike to generate optimized code for such an specific hardware architecture.
SGDK is wonderful for what it is and it's already super high level compared to most of the tools you'd have in the past. Now go have a look into how many years and hours of work magicians like Stef and Chilly Willy put into that thing. If you want some higher level than that and still capable of creating official releases-looking games, you'll need to invest a ton of work and expertise on it IMO.

[tryphon]

It's MUCH easier said than done though.

That, 10000 times.

Some common misconceptions I also had before studying the actual source code of several Mega Drive games:
- "The AAA titles carry some awesome piece of code" -> Not exactly. A LOT of the games I've looked into, including some Sega big shots, have far-from-optimized code. Some games barely make a rational use of the 68000's registers, for an example. Many of them also have a very poor DMA queue implementation (which is crucial for animation, for an example).

It may well be true. I saw game codes checking twice the same collision too. I think games were developped fast, and if the code seemed to work, devs hadn't time to optimize / fix things.

- "Japanese developers were the best!!!" -> As tomaitheous had already pointed in the past, you usually find better code in European-developed games than on Japanese-developed games.

I can't say. I had a look in some EA games and it wasn't always crystal-clear But yeah, nothing like Street Fighter II

- "Let's pick the most advanced games and mod them!"-> Usually, the later iterations of a game series or the hardware-pushing games are the most difficult ones to mod. There's a reason behind all these hacks you see for SOR2 and not SOR3, Sonic 1 and not the later ones or Ristar. The more advanced a game seems to be, the more likely it is to be a bitch to change anything, since the level of customization for their very own assets and design details is insane at times.

That's true, even for early games. In fact, game engines are not generic at all. Many things are game-specific and hardcoded, it requires a lot of work to make the code generic and reusable, likely more than developping a custom engine.

I wrote a level editor for Ghouls'n Ghosts some time ago. It's mostly working (you can edit backgrounds, level layouts, ennemy location) but there are bugs from time to time, and some things are painful to implement. Mainly because most special effects in the game are hardcoded (in some routine, there"s something like "if ((level == 1) and (position_x = 42)) then upload some new tiles, and of course it's not always in the same part of the game), it's a pain in the neck to make it generic. AFAIK, there's a great Super Mario World hacking tool (don't remember the name, by Lunar) and there are many hacks inside to make the game more friendly with the custom levels.

Concerning MD, huguesjohnson wrote a very good hacking tool, that (AFAIK) is still waiting for map and event editors, but it should be possible (I wrote a map exporter some years ago, and lorenzo published complete disassemblies of the game).

- "Let's produce game in batches like the Indie community does for PC."-> Don't forget the Mega Drive's hardware is super limited and most of the game factory tools are unlike to generate optimized code for such an specific hardware architecture.
SGDK is wonderful for what it is and it's already super high level compared to most of the tools you'd have in the past. Now go have a look into how many years and hours of work magicians like Stef and Chilly Willy put into that thing. If you want some higher level than that and still capable of creating official releases-looking games, you'll need to invest a ton of work and expertise on it IMO.

I'm not totally sure that ultra-optimized* code is needed for having good quality games (as you said, most of japanese games were so-so in coding). I started a port of Wonderboy 3, The Dragon's trap with SGDK and I'm 99 % sure it can be done in pure C (I posted a video where there are around 10 objects at screen with no slowdown, while no collisions, and I implemented collisions since, I begin to lose a frame here and there from 8 chars onscreen, which doesn't happen in game). And even with Shinobi, I'm more and more confident (it's a little more difficult since sprites are bigger and dynamic, so DMA must be handled carefully).

But yeah, it's a lot of work !

*: and we must distinguish code optimizations, and algorithm optimizations : I think there's much to do on the algorithm side.

[Barone]

*: and we must distinguish code optimizations, and algorithm optimizations : I think there's much to do on the algorithm side.

I agree here.
But what I was trying to say is: the level of optimization in most of the released games is enough to run those games the way they are but not necessarily enough to run an expanded version of them or something else reusing that engine.

A simple example: to include the Super and Hyper Sonic transformations into Sonic 1, Clownacy (which is one of the most skilled hackers of the Sonic Retro Community) had to replace the DMA queue method because the new sprites need to be allocated in the same VRAM space of the regular sprites since there wasn't enough space available in the VRAM to statically load them and the original method couldn't handle this task properly.

About the algorithm vs code optimization: yep, two different things. A third one IMO would be some like "architecture optimization", where I would try to handle things such as disabling the display prior executing DMA transfers (something that a lot of games don't do, surprisingly), splitting DMA transfers in smaller pieces to not halt the Z80 for too long in order to avoid/mitigate hurting the PCM playback quality (very few games took care of it, but Dynamite Headdy is one of the best in such aspect), etc.

[Thief]

^ Didn't know you were that well informed on the technical side of things, Barone.

[SEGA-Jorge]

This thread is awesome, thanks for your in site Barone, super helpful.

I'm always amazed by this community...

[Golpezas]

Point is, Loco Malito games need to be on the Genesis...

L'Abbaye des morts (#AbbayeMD) is about to be released.

[Barone]

L'Abbaye des morts (#AbbayeMD) is about to be released.

That's good news (I guess).

This thread is awesome, thanks for your in site Barone, super helpful.
I'm always amazed by this community...

Thank you for bringing the discussion.
I'm far from knowing as much as some other guys here (like tryphon) and as much as I'd like, but I'm always trying to improve and share what I know.

I think the idea of reusing commercial releases from the past as base for new games is feasible, but one of the main problems is the lack of tools and the scarce amount of MD games with complete and commented source code available.
On one hand, you have a super developed Sonic-specific community, a well developed SOR-specific community with lots of tools, docs, well organized source codes, research group, etc. On the other hand you have SGDK, spritesmind and a dozen of homebrew developers.
There's some void between those two sides. The hacking scene for non-Sonic games and non-SOR games is very, very thin and there isn't really an active community focused on that; the closest we have is M.I.J.E.T and the Russians of elektropage.ru; both far less active now than they were years ago. Our Blast Processing forum is still in embrionary stage in such aspect IMO.

[pikointeractive]

We are getting the rights of Timet the Flyin Squirrel. (me. nutz 2)I can get behind of the idea of making it a platforming engine for the community. Needs to be ported as it was developed in an Amiga; but is easier than hacking a rom.

[Lanzz]

We are getting the rights of Timet the Flyin Squirrel. (me. nutz 2)I can get behind of the idea of making it a platforming engine for the community. Needs to be ported as it was developed in an Amiga; but is easier than hacking a rom.

Wow, that is one of my favorite Amiga game. Do you plan to release the Mega Drive version? The source code of the Mega Drive version was leaked and then compiled long times ago...

[tryphon]

For the record : djcouchycouch wrote some time ago a generic engine with SGDK :

http://gendev.spritesmind.net/forum/...pic.php?t=1373

It seems quite advanced...

[Golpezas]

Needs to be ported as it was developed in an Amiga; but is easier than hacking a rom.

No, the rom already exists. The code where found somewhere and someone where able to compile it.