nesdev.org

So you're saying the SNES hardware is significantly more complex than the NES' hardware enough to take way more time to make software for?

I disagree. To make X for NES and make an identical X for SNES is roughly the same amount of time. The hardware isn't harder to use. The assembly language isn't harder to learn.

People are expecting more out of a SNES game. Mode 7. Larger, more colorful sprites. More levels. Multiple background layers.

If you don't use them, you will get "this looks like a NES game".

dougeff wrote: ↑Tue Nov 03, 2020 9:35 am People are expecting more out of a SNES game. Mode 7. Larger, more colorful sprites. More levels. Multiple background layers.

If you don't use them, you will get "this looks like a NES game".

Yeah.

They're expecting better graphics which themselves take more time to deliver since you have to add shading every time you draw something, they're expecting more assets to give the background some depth, which takes more time as there's more content to make, and they're expecting homebrew developers to figure out the weirdness of mode 7's affine mathematics.

Honestly, if I was making a SNES game. I wouldn't care if it looks vaguely NES-like, because even then, I'd still be doing things that NES can't do, like having more sprites per line, having 100 or so sprites on screen(however colourful or big they are). I'd willingly sacrifice background colour for multiple background layers since I've seen Shantae on GBC and it looks way more colourful than an NES game, though it could use a lot more parallax layers(LOL SNES has 2BPP GB/C character data format).

Actually, yeah, my SNES games would probably look more like GBC in terms of background colour usage than NES due to having 8 palettes per layer instead of 4, and being able to colour by 8x8 boundaries instead of 16x16.

psycopathicteen wrote: ↑Tue Nov 03, 2020 9:29 am Alisha's Adventure would've taken much less time if I had the foresight of every possible thing I wanted to add to my game.

Which is why some devs prototype on a PC and then demake the game later.

Since I don't know the SNES hardware enough to make a game yet, I can't say if developing for the NES or SNES is easier by a skilled developer that knows the hardware, but Doug is probably right that they are about equal. The SNES doesn't have the same tight ROM and RAM restrictions that the NES has (without the right mappers), which I imagine would be an advantage over the NES. For the assembly you can even treat the 65816 like a 6502 and it will still work.

As for learning the SNES hardware, you have a lot more to read and take in before you can even display anything on the TV. Just the required initialization code on the SNES is much longer than the same thing on NES, Game Boy, Mark3/MS or PC-Engine, which newbies first has to tackle. Then you have the unique audio hardware to figure out.
I believe this hardware hurdle is probably stopping a lot of newbie homebrewers.

SNES is too limited for general purpose gamesystem but not limited enough to make it intresting challenge for developers/artists. The best NES games still have that "retro-feel" by todays standards no matter what you do. But SNES games can compete with modern 2D games. So the SNES homebrew is in the same playfield with ALL 2D games for every console.

Related topics: Homebrew complexity; Why no SNES homebrew scene?

Now that I think about it, the topic sounds bit weird. Well, If I'm the Judge, you have every right to expect any random SNES ROM to be the most mindblowing thing in the World... :)

At least for my own SNES development, I don't care that much what the average retro gamer will think since they aren't the intended audience.

When comparing the complexity of programming the NES and SNES you should take into account the complexity the NES has (attribute table, bank switching, 30x32 tilemaps, needing to properly time writes for effects, needing to work around lack of CHR RAM DMA, etc.) that the SNES removes. Having more stuff means you've got tools to do things easier. The 65816 also simplifies a lot of code significantly compared to similar code I've written for the 6502.

nin-kuuku wrote: ↑Tue Nov 03, 2020 12:19 pm SNES is too limited for general purpose gamesystem but not limited enough to make it intresting challenge for developers/artists. The best NES games still have that "retro-feel" by todays standards no matter what you do. But SNES games can compete with modern 2D games. So the SNES homebrew is in the same playfield with ALL 2D games for every console.

How can SNES games compete with modern 2D games if the SNES is clearly not HD by any means? The SNES also doesn't support widescreen, since its 8x8 minimum tile boundary isn't good for anamorphic widescreen, and if you use letterboxing, you lose resolution, making it harder to compete with modern 2D Switch games that don't deliberately try to look retro.

The 256x224 resolution used by the overwhelming majority of SNES games is small enough that it's easy to see each pixel separately even with low quality video cables.

The SNES can't even compete with 2D games on later non-HD consoles because 512x448 requires 4 times as much VRAM as 256x224 for every discreet background element, and you only have 64kB of VRAM.

If the SNES is too limited for 'general purpose gamesystem', then it just can't compete because those very same limits set it back compared to modern 2D game consoles.

Idk if that thing about peoples expectations into SNES homebrew is really true. There's so few projects around, there isn't really much to compare against. I don't think people would compare something to the really good commercial SNES games (truth is lots of them were shovelware and really not very good) unless you try to sell it to them (and then, that expectation would be justified)

My personal expectation / goal for SNES homebrew, from a dev perspective, is that it does something that "works well" on SNES, e.g. using the HW to make some neat effects, or otherwise pulling something of that you would have not expected to be possible on the platform or in another way just fits. Otherwise it would be pointless and you could as well choose another platform (but that's just me - I can see how opinions might differ there).

So viewed from that perspective, you're right, expectations are higher than for NES, but within reason, I think. It's not like it is rocket science and impossibly difficult to make a neat HDMA effect. Maybe another thing that stops people from even getting into SNES dev might be all the talk about how hard it is - I don't find coding for SNES as difficult to get into as people often claim it is. It's really fun, people should try it out, you get hooked really fast.

I think it should be advertised more what is so unique about the SNES and what makes coding for it so great. Everything about the system is really strange - the CPU, the graphics hardware, the sound hardware - it is a mess, in fact - but that's actually a good thing. It sets it apart from all the other platforms and makes for a fun experience for people who are otherwise sick of always doing the same stuff again and again. I really suggest to every programmer to try and make an SNES project at least once. It is the most fun coding stuff I did in ten years or so.

However when I started out and looked stuff up (guides, tutorials, other project etc) there were some pitfalls that I could have avoided just if I had known from the start which ones were the ones I could get the most value out of.

Maybe someone should compile a "definite" newbie guide, not necessarily a full tutorial but rather a compilation of where are the "good" resources, where to begin, what options for tools you have with pros and cons, good forum threads, good inspiration / examples for how to use the HW to full effect etc. - a kind of "meta" guide would be great I think.

But you're also right in that a (complete) well written and coherent step-by-step tutorial is somehow strangely missing.

Also someone said somewhere that as people scale down the scope of their project to something that can actually get finished, they figure out that the same thing can be achieved on the NES as well and just settle for that (but that's not my opinion / experience).

tepples wrote: ↑Tue Nov 03, 2020 11:22 am

psycopathicteen wrote: ↑Tue Nov 03, 2020 9:29 am Alisha's Adventure would've taken much less time if I had the foresight of every possible thing I wanted to add to my game.

Which is why some devs prototype on a PC and then demake the game later.

IMO, that's not the right approach, because the limitations can be a tool that works for you for in guiding the design process and that's the whole point in developing for a retro platform. Design is an iterative process. When you hit a wall, you start over. If you do the whole concept in a space where you have total freedom, the wall will not be there and so the whole advantage you have is gone.

Nikku4211 wrote: ↑Tue Nov 03, 2020 3:58 pm If the SNES is too limited for 'general purpose gamesystem', then it just can't compete because those very same limits set it back compared to modern 2D game consoles.

I don't think someone just casually interested would get the difference.

NovaSquirrel wrote: ↑Tue Nov 03, 2020 3:55 pmAt least for my own SNES development, I don't care that much what the average retro gamer will think since they aren't the intended audience.

Exactly my opinion. Just the fact that people just can't download it from steam is a huge limiting factor. Todays average gamers attention span is a few seconds at best. There just isn't a big audience for these types of projects, a homebrew game is something you should make mainly for yourself. And that is the good thing I think, something that is made with an audience / sales in mind is in my opinion likely to turn out mediocre at best.

Nikku4211 wrote: ↑Tue Nov 03, 2020 3:58 pm

nin-kuuku wrote: ↑Tue Nov 03, 2020 12:19 pm SNES is too limited for general purpose gamesystem but not limited enough to make it intresting challenge for developers/artists. The best NES games still have that "retro-feel" by todays standards no matter what you do. But SNES games can compete with modern 2D games. So the SNES homebrew is in the same playfield with ALL 2D games for every console.

How can SNES games compete with modern 2D games if the SNES is clearly not HD by any means? The SNES also doesn't support widescreen, since its 8x8 minimum tile boundary isn't good for anamorphic widescreen, and if you use letterboxing, you lose resolution, making it harder to compete with modern 2D Switch games that don't deliberately try to look retro.

The 256x224 resolution used by the overwhelming majority of SNES games is small enough that it's easy to see each pixel separately even with low quality video cables.

The SNES can't even compete with 2D games on later non-HD consoles because 512x448 requires 4 times as much VRAM as 256x224 for every discreet background element, and you only have 64kB of VRAM.

If the SNES is too limited for 'general purpose gamesystem', then it just can't compete because those very same limits set it back compared to modern 2D game consoles.

HD, wide-screen, high resolution, bazillions of memory... Fancy stuff which none of makes the games any better on their own. The SNES hardware has the capabilities of making games comparable to modern 2D games, graphically and mechanically. "Comparable" doesn't mean that the games must look exactly like the modern games, only that they don't have that retro look. Yoshi Island is a good example (although you might argue it's not purely SNES hardware since it's a GSU-2 game).

none wrote: ↑Tue Nov 03, 2020 5:20 pm But you're also right in that a (complete) well written and coherent step-by-step tutorial is somehow strangely missing.

Yup, and the reason to that is probably that the SNES homebrew scene is, compared to the NES homebrew scene, smaller and have less tools, guides, forums etc as I mentioned earlier in this thread.

Not knowing where to start is probably the single biggest hurdle for any newbie, especially if you haven't programmed before, or you've only done high level programming. If there was a Nerdy Nights for SNES that helps you make Pong, newbies would just be raining into the homebrew scene. No but really, I think it would make a big difference, and it's definitively one of the goals to achieve if one aims to make the SNES homebrew scene become more lively.

Pokun wrote: ↑Tue Nov 03, 2020 5:22 pm HD, wide-screen, high resolution, bazillions of memory... Fancy stuff which none of makes the games any better on their own. The SNES hardware has the capabilities of making games comparable to modern 2D games, graphically and mechanically. "Comparable" doesn't mean that the games must look exactly like the modern games, only that they don't have that retro look. Yoshi Island is a good example (although you might argue it's not purely SNES hardware since it's a GSU-2 game).

SNES games kind of do have the retro look in my opinion. As long as you can still see the pixels, and as long as there is no anti-aliasing at all, it looks retro.
And even if they don't look retro, they don't look modern either. It's not a dichotomy, it's a spectrum, and the SNES falls in its middle.

Even if the average non-retro gamer isn't actively looking for these things, they will definitely notice them. Sure, the games can still be fun regardless of what they look like(which makes whether the games are visually comparable to modern games or not irrelevant), but the low resolutions can make it harder for them to get into the games.

Most of Yoshi's Island's aesthetic technically has nothing to do with the GSU-2. Yes, it does look timeless, but even then, your average gamer probably doesn't just want 'timeless', they want something that looks 100% 'modern' so that they can feel like they can fit in with other people when playing this game.

From what I know about programming gaming consoles, GBA is probably the easiest - more tools to aid in asset creation, devkitPro abstracts away all the start-up code you'd have to write (though I made a simple text demo all in Arm from scratch) and it's basically a much more streamlined SNES with most features available in some equivalent form (except the priority-per-tile deal - you can only set priority for the whole BG) plus affine transformations available for sprites. It might be a good alternative for someone who wants to do SNES homebrew without 65816, unless the lower resolution and sample quality are deal breakers.

Well what you call retro is a matter of opinion. I guess the SNES do have the typical 16-bit look, and it's definitely no Neo Geo which kind of symbolizes the peak of 2D game graphics.